Citation
Public bicycle sharing as a population-scale health intervention for active transportation in Denver, Colorado

Material Information

Title:
Public bicycle sharing as a population-scale health intervention for active transportation in Denver, Colorado
Creator:
Duvall, Andrew L
Publication Date:
Language:
English
Physical Description:
1 electronic file : ;

Subjects

Subjects / Keywords:
Bicycle sharing programs ( lcsh )
Cycling -- Health aspects ( lcsh )
Bicycle sharing programs ( fast )
Cycling -- Health aspects ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Inadequate physical activity associated with acute car dependence is linked to increased risk of obesity and related chronic health conditions, including heart disease, diabetes, hypertension, depression and cancer. More than half of U.S. adults do not meet the minimum recommended levels of physical activity. This research is a mixed methods quantitative/qualitative study of the impact of a public bicycle sharing system as a health intervention to induce active transportation behavior. The primary quantitative outcome was a net increase in active transportation through shared bike use. Qualitative outcomes included investigation of the effects of the intervention on the behaviors of individual users, sources of motivation to participate and continue use, indicators and pathways of diffusion of the intervention, and broader impacts of the intervention. Denver B-cycle annual members logged an average 60.3 minutes of weekly checkout time, of which an estimated 35.5% to 50.0% replaced car trips. Annual members differed significantly from the general population, being more likely to be male, non-Hispanic, Caucasian, aged 25 and 44, more educated, with higher income, higher self-reported health status, and tended to be of normal weight. Multivariate models of influencing factors found two key variables associated with the number of checkouts among annual members. Commuting via shared bikes and the ability to replace car use increased the odds of higher checkouts. Proximity of residence to Denver B-cycle stations was not a significant predictor. Women were as likely as men to commute via shared bikes. The use of Denver B-cycle increased net active transportation among annual members. Users discovered their own meaningful ways in which the use of shared bicycle best served their needs. Some participants best able to integrate Denver B-cycle into their lifestyles reported weight loss, increased fitness, reduced car dependence, and economic benefits. The central implication of this research is that public bicycle sharing can be effectively applied to increase active transportation behavior. The findings are consequential for two trends: shifts away from car use toward use of shared bikes, and increases in overall active transportation.
Bibliography:
Includes bibliographical references.
Statement of Responsibility:
by Andrew L. Duvall.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
857717299 ( OCLC )
ocn857717299

Auraria Membership

Aggregations:
Auraria Library
University of Colorado Denver

Downloads

This item has the following downloads:


Full Text
PUBLIC BICYCLE SHARING AS A POPULATION-SCALE HEALTH
INTERVENTION FOR ACTIVE TRANSPORTATION IN DENVER, COLORADO
by
Andrew L. Duvall
B.A., University of Wyoming, Laramie, 1994
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado Denver in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Health and Behavioral Sciences
2012


This thesis for the Doctor of Philosophy degree by
Andrew L. Duvall
has been approved for the
Health and Behavioral Sciences
by
Debbi Main, Chair and Advisor
John Brett
Wesley Marshall
Eric France
Date: April 6, 2012
11


Duvall, Andrew L. (Ph.D., Health and Behavioral Sciences)
Public Bicycle Sharing as a Population-Scale Health Intervention for Active
Transportation in Denver, Colorado
Thesis directed by Professor Debbi Main
ABSTRACT
Inadequate physical activity associated with acute car dependence is linked to
increased risk of obesity and related chronic health conditions, including heart disease,
diabetes, hypertension, depression and cancer. More than half of U.S. adults do not meet
the minimum recommended levels of physical activity.
This research is a mixed methods quantitative/qualitative study of the impact of a
public bicycle sharing system as a health intervention to induce active transportation
behavior. The primary quantitative outcome was a net increase in active transportation
through shared bike use. Qualitative outcomes included investigation of the effects of the
intervention on the behaviors of individual users, sources of motivation to participate and
continue use, indicators and pathways of diffusion of the intervention, and broader
impacts of the intervention.
Denver B-cycle annual members logged an average 60.3 minutes of weekly
checkout time, of which an estimated 35.5% to 50.0% replaced car trips. Annual
members differed significantly from the general population, being more likely to be male,
non-Hispanic, Caucasian, aged 25 and 44, more educated, with higher income, higher
self-reported health status, and tended to be of normal weight.
Multivariate models of influencing factors found two key variables associated
with the number of checkouts among annual members. Commuting via shared bikes and
iii


the ability to replace car use increased the odds of higher checkouts. Proximity of
residence to Denver B-cycle stations was not a significant predictor. Women were as
likely as men to commute via shared bikes.
The use of Denver B-cycle increased net active transportation among annual
members. Users discovered their own meaningful ways in which the use of shared
bicycle best served their needs. Some participants best able to integrate Denver B-cycle
into their lifestyles reported weight loss, increased fitness, reduced car dependence, and
economic benefits.
The central implication of this research is that public bicycle sharing can be
effectively applied to increase active transportation behavior. The findings are
consequential for two trends: shifts away from car use toward use of shared bikes, and
increases in overall active transportation.
The form and content of this abstract are approved. I recommend its publication.
Approved: Debbi Main
IV


DEDICATION
I dedicate this work to my wife Julie, my daughters Stella and Piper,
and to my family and friends.
Their support was the wind at my back, and their love helped my wheels to spin.
v


ACKNOWLEDGMENTS
The author is greatly appreciative of the U.S. National Science Foundation
Integrative Graduate Education and Research Traineeship (IGERT) program, IGERT
Award No. DGE-0654378.
The Association of Schools of Public Health / Centers for Disease Control and
Prevention Environmental Health Community Design scholarship program, and the
Community Benefit Initiatives Committee of the Kaiser Permanente Colorado Institute
for Health Research provided generous continued support.
Scholarship and camaraderie fostered by the faculty, staff and students of the
University of Colorado Denver Department of Health and Behavioral Sciences, and the
Center for Sustainable Infrastructure Systems at the University of Colorado Denver,
directed by Dr. Anu Ramaswami, has been of substantial and ongoing value.
The author wishes to express deep gratitude for the opportunities and support
afforded by dissertation committee chair Dr. Debbi Main, and committee members Dr.
John Brett, Dr. Wesley Marshall and Dr. Eric France. These talented and committed
educators contributed insight, encouragement, direction and advice, which have greatly
enhanced the quality of this work.
This research would not have been possible without a great many people. The
author wishes to thank those who directly affected the advancement of this work.
Organizations'. The City and County of Denver, Denver Bike Sharing, BikeDenver, the
Denver Mayors Bicycle Advisory Committee, GreenPrint Denver,
Bikes Belong, and B-cycle.
Individuals'. Parry Bumap, Steve Sander, Emily Snyder, Piep van Heuven,
Tracy Halasinski, Dick Gannon, Karen Good, Nick Bohnenkamp, Cindy Bosco,
Ken Gart, Corina Lindley, Nate Kvamme, Tim Blumenthal, Bob Burns,
Governor John Hickenlooper, Mayor Guillermo Vidal, and the volunteers, employees and
board of Denver Bike Sharing.
Special thanks goes to everyone who rides or aspires to ride a bicycle, for recreation, for
transportation, for the betterment of themselves and their community, and especially for
enjoyment.
vi


TABLE OF CONTENTS
CHAPTER
1. INTRODUCTION.............................................................1
Background...........................................................1
Significance Of The Study............................................2
Research Question and Specific Aims..................................7
Overview of Methods..................................................8
2. LITERATURE REVIEW.......................................................10
Development Of The Current Built Environment........................10
The Emergence Of Car Dependence..................................10
Decoupling Transportation From Physical Activity.................11
The State Of Obesity And Why It Matters..........................12
Obesity Risk Through The Urban Form..............................14
Beyond The Car: Public Bike Sharing..............................16
Guiding And Applied Theories For This Research......................19
Informing the Study.................................................21
Intervention Through The Built Environment.......................22
Intervention And The General Populations.........................26
The effects of environmental context.............................32
Public Bicycle Sharing And Active Transportation.................37
Public Bicycle Sharing Participation.............................41
Motivation To Use Public Bicycle Sharing.........................44
Summary Of Literature Review........................................45
3. METHODS OF DATA COLLECTION AND EVALUATION...............................48
Study Design........................................................48
vii


Brief Glossary Of Terms Specific To This Research..........................49
Study Setting..............................................................49
Data Sources...............................................................51
Data Variables.............................................................53
Geographic Data Overview...............................................56
Denver B-Cycle System Usage Dataset Overview...........................56
Denver B-Cycle User Survey Dataset Overview............................57
Select Participants........................................................61
Human Subjects Review......................................................61
Methods of Data Collection and Evaluation, by Specific Aim.................61
4. RESULTS........................................................................79
Introduction...............................................................79
Aim 1: Examine the physical characteristics and human activity in the built
environment in which Denver B-cycle stations are situated..................79
Aim 1.1: Identify features that distinguish low and high performing sites..79
Characteristics Of The Built Environment At Select Stations............79
Market Street Station..................................................81
16th and Little Raven..................................................84
REI....................................................................87
19th and Pearl Street..................................................91
Union Station..........................................................93
Denver Health..........................................................96
Five Points............................................................99
15th and Tremont......................................................102
25th and Lawrence.....................................................105
Pepsi Center..........................................................108
viii


Analysis of Characteristics of the Built Environment at Select Stations.Ill
Aim 1.2: Investigate the integration of Denver B-cycle within the existing urban
transportation system........................................................115
Station Network Density of Denver B-cycle................................116
Proximity To Transit.....................................................117
Proximity To Bicycle Facilities..........................................119
Summary Of Denver B-Cycle Within The Urban Transportation System... 121
Aim 2: Describe the characteristics of Denver B-cycle users..................123
Description Of System-Wide Patterns Of Use...............................123
Checkout Activity........................................................123
Membership Characteristics...............................................128
Aim 2.1: Determine demographic characteristics of Denver B-cycle users......130
Demographic data summary.................................................130
Evaluating Gender, Ethnicity and Race....................................134
Evaluating Age Group, Educational Attainment and Household Income.... 135
Health-Related Survey Data...............................................137
Summary: Denver B-cycle Users Versus the General Population..............146
Aim 2.2: Investigate factors influencing Denver B-cycle use.................148
Multivariate Modeling Of Number Of Checkouts By Denver B-Cycle Annual
Members..................................................................149
Commute Via Denver B-Cycle...............................................153
Summary Of Factors Influencing Use.......................................156
Aim 2.3: Determine impacts of Denver B-cycle on shifts toward active
transportation...............................................................158
Summary Of The Impact Of Denver B-Cycle On Active Transportation.... 160
Aim 2.4: Identify active transportation benefits for Denver B-cycle annual
members......................................................................161
IX


Physical Activity Benefits............................................161
Summary Of Active Transportation Benefits For Denver B-Cycle Annual
Members...............................................................167
Aim 3: Provide an in-depth description of the impacts of Denver B-cycle as a
public health intervention................................................168
Motivation for Active Transportation Behavior.........................168
Health outcomes.......................................................169
Economic Outcomes.....................................................174
Factors For Incorporating Behavior Change Into Lifestyle..............176
Diffusion of Use within Social Networks...............................181
Social change.........................................................187
5. DISCUSSION AND LIMITATIONS....................................................195
Discussion................................................................195
Interpreting The Findings.............................................195
Denver B-Cycle As A Lifestyle Intervention............................196
Stealth Intervention................................................200
Population Reached Through The Intervention...........................204
Broader Impacts on Bicycling Activity In Denver.......................206
Long-Term Outcomes....................................................210
Ongoing Intervention..................................................211
Informing Application: Policy And Infrastructure......................214
Theoretical Implications..............................................218
Opportunities For Future Research.....................................224
Study Limitations.........................................................228
Conclusions...............................................................230
x


REFERENCES................................................................232
APPENDIX..................................................................244
A. Denver B-cycle Select Station Observation Sheet...............245
B. Denver B-cycle User Survey Instrument...........................246
C. Semi-structured Interview Question Guide for Denver B-cycle Select
Participants.......................................................256
D. Code Definitions................................................259
xi


LIST OF TABLES
Table
II. l Obesity: Chronic Health Conditions and Risks.................................14
III. 1 Data Sources Associated with Study Aims And Hypotheses......................51
111.2 Data Variables Used in this Research, Arranged by Source.....................53
111.3 Denver B-cycle Select Stations...............................................62
111.4 Denver B-cycle In-Service Area ZIP Codes.....................................70
111.5 Characteristics of Select Participants.......................................75
IV. 1 Characteristics of High and Low-Use Denver B-cycle Select Stations..........Ill
IV.2 Means, Standard Deviations and Intercorrelations for Total Checkouts and Predictor
Variables (\ 32)...................................................................120
IV.3 Simultaneous Multiple Regression Summary for Transit Stops, Bicycle Facilities,
and Network Density Predicting Total Checkouts (N=32)..............................121
IV.4 2010 Denver B-cycle Checkouts Per Station in Central Downtown.................126
IV.5 Demography and Characteristics of Denver B-cycle Users........................132
IV.6 Gender, Ethnicity and Race of Denver B-cycle Annual Members and Denver
County Population (U.S. Census 2010)...............................................134
IV.7 Age Group, Educational Attainment and Household Income of Denver B-cycle
Annual Members and Denver County Population (2010 U.S. Census American
Community Survey 1-Year Estimates).................................................136
IV.8 Health Characteristics of Denver B-cycle Annual Members Versus the Denver
County Adult Population............................................................138
IV.9 Calculated BMI Figures from Self-Reported Height and Weight of Denver B-cycle
Annual Members.....................................................................140
IV. 10 Percent within Weight Class Groups as Determined by BMI, Denver B-cycle
Annual Members Versus Denver County Population.....................................140
IV. 11 Active Transportation Behavior of Denver B-cycle Users......................142
xii


IV. 12 Ordered Logistic Regression Predicting Number of Denver B-cycle Checkouts by
Annual Members.................................................................150
IV. 13 Multiple Logistic Regression Predicting Commute Via Denver B-cycle By Annual
Members.........................................................................154
IV. 13 Year-end Average Totals and Estimates of Denver B-cycle Use Per Annual
Member..........................................................................162
IV. 14 Average Frequencies of Use Per Denver B-cycle Annual Member Per Week of
Use.............................................................................162
IV. 15 Average Net Year-end Indicators of Active Transportation Behavior Per Annual
Member Attributo Denver B-cycle Use.............................................164
IV. 16 Average Indicators of Active Transportation Behavior Per Annual Member Per
Week of Use Attributo Denver B-cycle Use.......................................164
V. 1 Denver Commuter Population and Estimated Numbers of Bicycle Commuters, 2009
versus 2010....................................................................208
V.2 Estimated Effects of Denver B-cycle Commuters on the Increase in Bicycle
Commuting in Denver, 2009 to 2010...............................................208
D.l Code Definitions
259


LIST OF FIGURES
Figure
II. 1 U.S. Cities Larger than 400,000, Ranked by Bicycle Commuting..........34
III. 1 Map of Denver B-cycle Stations in the Central Downtown Area, within the City and
County of Denver (2010).....................................................50
III. 2 Denver B-cycle station Groupings by Geographic Service Areas.........67
IV. 1 Denver B-cycle Select Station Locations...............................80
IV.2 Behavior Mapping: Market Street Station................................83
IV.3 Behavior Mapping: 16th and Little Raven................................86
IV.4 Behavior Mapping: REI...................................................90
IV.5 Behavior Mapping: 19th and Pearl Street................................93
IV.6 Behavior Mapping: Union Station.........................................95
IV.7 Behavior Mapping: Denver Health.........................................97
IV.8 Behavior Mapping: Five Points..........................................101
IV.9 Behavior Mapping: 15th and Tremont Street..............................104
IV. 10 Behavior Mapping: 25th and Lawrence Street...........................107
IV. 11 Behavior Mapping: Pepsi Center.......................................110
IV. 12 Network Distances to Denver B-cycle Stations.........................116
IV. 13 Location of Denver B-cycle Stations Relative to Transit Stops and Stations in
Central Downtown............................................................118
IV. 14 Denver On and Off-Street Bicycle Facilities..........................119
IV.15 Cumulative Denver B-cycle Checkouts in Central Downtown, 2010.........124
IV. 16 Number of Denver B-cycle Checkouts Per Week from Stations in Central
Downtown....................................................................125
IV. 17 Weekly Average Checkouts Per Station by Neighborhood Service Areas...127
xiv


IV. 18 Activated Denver B-cycle Memberships by Type, Per Week.
128
IV. 19 Geocoded Addresses of Denver B-cycle Annual Members in Central Downtown
Denver..................................................................129
IV.20 Self-Assessed Bicycling Ability of Denver B-cycle Users...........143
IV.21 Bicycle Ownership among Denver B-cycle Users......................144
IV.22 Car and Bicycle Commuting Mode Share: Denver B-cycle Users, Denver County,
Colorado and the United States (U.S. Census Bureau 2011)................145
IV.23 Annual Member Responses for Frequency of Denver B-cycle Trips that Replace
Car Trips...............................................................159
IV.24 Percentage of Recommended Weekly Physical Activity Met by Average Weekly
Total Denver B-cycle Minutes of Checkout: 90-Minute Recommendation Versus 150-
Minute Recommendations................................................163
IV.25 Percentage of Recommended Weekly Physical Activity Met by Average Weekly
Minutes of Denver B-cycle Checkout Time Replacing Car Trips: 90-Minute
Recommendation Versus 150-Minute Recommendations........................165
IV.26 Average Weekly Number of Checkouts Per Active Denver B-cycle Annual
Member..................................................................166
IV.27 Helmet Use Among Denver B-cycle Users While on Private Bikes Versus While
on Denver B-cycle Bikes.................................................191
IV. 1 Denver Bicycle Commuter Mode Share, 2005 to 2010..................207
xv


CHAPTER I
INTRODUCTION
Background
Acute dependence on automobiles for transportation contributes to lack of
physical activity; a condition identified with increased health risks. A well-supported
body of literature has identified links between high reliance on cars for transportation and
increased risk of obesity and obesity related chronic conditions, which are associated
with inadequate levels of physical activity (Handy, Boarnet et al. 2002; Frank, Andresen
et al. 2004; Ewing, Brownson et al. 2006; Jones, Rutt et al. 2006; Lopez-Zetina, Lee et al.
2006; Jacobson, King et al. 2011). The intensity of car use and air pollution also affects
community livability and quality of life due to environmental degradation, pollution, and
reduced access to land, rendering areas inhospitable to outdoor activities (MacKerron and
Mourato 2009).
As the health risks and impacts attributable to high car dependence become
increasingly known, strategies to mitigate and reduce the quantity of car use have
become increasingly important. In recent years, some cities have introduced public
bicycle sharing as an infrastructural tool to reduce car dependence while increasing
opportunities for physical activity. Public bicycle sharing is a form of public
transportation in which a fleet of bicycles within a network of automated stations is made
available to users for short trips, for a minimal fee.
Public bicycle sharing projects have been implemented in cities throughout the
world, primarily in countries in Europe and Asia already supportive of bicycling for
1


transportation. However, in the past few years, public bicycle sharing systems have
begun to appear more broadly, including several car-dependent cities in the United States
(Shaheen and Guzman 2011). Still in its infancy, literature specific to public bicycle
sharing is quite limited presently, with relatively little known about its associated health,
social, behavioral, environmental and infrastructural impacts. The findings of this
dissertation research are intended to inform public health policy through examination of
the operation of a public bicycle sharing system and its effects on the user population and
the city in which it operates.
This dissertation research focuses on Denver B-cycle, a public bicycle sharing
system in Denver, Colorado, which opened to the public on April 22, 2010. Denver B-
cycle was the first large-scale public bicycle sharing system in the United States,
primarily designed to target the behavior of people who live and/or work in downtown
Denver. During its initial season, spanning from April to December 2010, Denver B-
Cycle was comprised of 500 bicycles available from 50 stations distributed
predominantly in the central downtown business district of the city, with two outlying
smaller groupings of stations in the Cherry Creek and Denver University neighborhoods.
This dissertation research focuses on the initial 2010 Denver B-cycle season of
operations.
Significance Of The Study
Inadequate physical activity is a modifiable risk factor contributing to the current
problem of obesity in the United States. The costs associated with treating obesity-related
diseases are expected to rise dramatically, with total health-care costs attributable to
2


obesity and overweight doubling every decade, accounting for 16-18% of total US
health-care costs by 2030 (Wang, Beydoun et al. 2008). Health care expenditures for
obese people are 42% more than for people of healthy weight, with obesity-related
medical costs accounting for $147 billion in U.S. annual medical costs (Levi, Vinter et al.
2010).
Unfortunately, more than half of U.S. adults do not meet the minimum
recommended levels of physical activity (Centers for Disease Control and Prevention
2005), even though an increase in regular physical activity can significantly reduce their
relative risk for related chronic illnesses (Warburton, Nicol et al. 2006; Haskell, Lee et al.
2007). Even modest improvements to physical activity can reduce morbidity and
mortality (Wen, Wai et al. 2011), with the greatest benefit to individuals who shift from
no activity to low levels of physical activity (Woodcock, Franco et al. 2010).
Reduction of preventable chronic conditions and associated costs is of paramount
importance if the health care system is to sustain any ability to meet projected needs. One
study predicts that by 2015, 75% of adults in the U.S. will be overweight or obese (Wang
and Beydoun 2007). Faced with these staggering health and economic figures, broadly
applied and effective interventions to increase physical activity and reduce incidence of
obesity in populations are critically needed to avoid future catastrophic health and
economic outcomes. Obesity affects everyone, as a majority of the population is either
presently obese or at elevated risk of obesity, and the entirety of the population bears the
associated costs.
A key barrier to reducing the health and economic costs of obesity is the high
dependence on automobiles for transportation, a norm strongly established in much of the
3


United States. In 2009, 40% of all trips in the U.S. were shorter than two miles, though
cars were the dominant mode of choice for these short distances (Alliance for Biking &
Walking 2012). During the rapid spread of urban areas in the post World War II period,
car use has risen steadily, concurrent with a decrease in bicycling and walking trips
(Ogden and Carroll 2010). Over the same period of time, per capita vehicle miles traveled
has also climbed, also attributed as a factor influencing risk of obesity (Jacobson, King et
al. 2011). Time spent in a car is time that cannot be spent being physically active. One
study of the effects of car dependence on obesity found a 6% increase in obesity risk for
each hour spent in a car daily (Frank, Andresen et al. 2004). As a result, replacing
behaviors of car dependence with physically active transportation has become a more
recent focus of health promotion efforts.
The impact of the physical form of the built environment on behavior has been
widely studied, perhaps most notably in how the types and form of infrastructure
encourages car use at the expense of physically active transportation (Handy, Boamet et
al. 2002; Frank, Saelens et al. 2007; Pendola and Gen 2007). Socioeconomic factors of
the built environment, such as access to opportunities for more healthy choices are also
important, as populations in lower resource communities are at increased risk of obesity
(Babey, Hastert et al. 2009; Sallis, Saelens et al. 2009; Beech, Fitzgibbon et al. 2011).
Although research points to how factors of the built environment contribute to behaviors
associated with risk of obesity, it also suggests the potential for modifying features of the
built environment to induce health-promoting behaviors.
Elements of the built environment have been intentionally designed to target
physical activity behaviors at a limited scale in community settings, such as the
4


development of walking routes in a community (Krieger, Rabkin et al. 2009). Behavioral
health interventions applied through infrastructure to affect active transportation at a
large population scale such as bike sharing are not common, and have not been well
studied. The implementation of public bicycle sharing is an opportunity to evaluate the
viability of a population scale active transportation intervention. A public bicycle sharing
system by design introduces a readily accessible alternative to car use for short trips,
while simultaneously introducing an opportunity for low to moderate levels of physical
activity to users.
Behavioral health interventions for physical activity have largely been conducted
with the assumption that information access will lead to motivation and action. Such
minimal intervention approaches have not proven successful (Keyserling, Hodge et al.
2008). Some smaller scale interventions use enhanced approaches, including individual
level strategies involving high-intensity tailored coaching and/or peer counseling to
realize impact (Dutton, Provost et al. 2008). Such enhanced approaches, however, are too
logistically challenging and costly as a population level solution.
As an intervention, public bicycle sharing is a broad-based approach to reach a
large population. The designed intent is for use of the system to fit the lifestyles of users,
who develop their own motivational reasoning to become participants. The users
themselves are important operational elements of the intervention, self-selecting for
participation, self-moderating continued activity, and playing roles as recruiters and
transmitters of information to engage additional participants.
The design and outlay of the public bicycle sharing system is intended to facilitate
development of individual motivation by meeting basic transportation needs. Through
5


supportive placement of public bicycle sharing stations near places of residence and
desirable destinations, a wide variety of urban trips for which a car might have previously
been used can be replaced with shared bicycle trips. Additionally, with access to public
bicycle sharing system bicycles, transit use, which has been identified to encourage active
transportation (Besser and Dannenberg 2005), might be made more attractive and
accessible to people who do not currently use transit due to inconvenience. Having the
option for quicker trip completion than walking without the need to use a car positions
public bicycle sharing as an attractive transportation mode. Locations too distant from
transit stops to access by walking, but within the public bicycle sharing system service
area are prime for shared bike/transit trips.
In summary, this research begins to fill substantial gaps in current knowledge
regarding population-scale interventions to increase physical activity. As public bicycle
sharing system is deployed in more cities, this information is of mounting importance to
provide public health professionals and policy makers with information for evidence-
based health policy and community design decisions.
Obesity is in part a consequence of high car dependence, and as such is a
modifiable risk. Given that obesity and associated illnesses are a looming threat to an
increasing majority of the population, and that costs to treat obesity-related illnesses are
projected to rise dramatically in the coming decades (Wang, Beydoun et al. 2008), it is a
population-scale problem that may best be addressed by a population-scale intervention.
By using infrastructure to induce changes in behavior, a large number of people can be
simultaneously targeted. Introducing shared bicycles for transportation has the potential
both to increase active transportation and replace car trips, joint goals of mediating health
6


risks associated with obesity.
Research Question and Specific Aims
The primary purpose of this dissertation research is to address the following
question:
Does public bicycle sharing serve as a population-scale intervention to catalyze
active transportation within a population?
The following specific Aims and hypotheses are designed to address this
question:
Aim 1: Examine the physical characteristics and human activity in the built environment
in which Denver B-cycle stations are situated.
Aim 1,1: Identify features that distinguish low and high performing sites.
Aim 1,2: Investigate the integration of Denver B-cycle within the existing urban
transportation system.
Hypothesis 1.2: Denver B-cycle stations having greater integration with transportation
infrastructure, as indicated by proximity to public transportation stops, on and off-street
bicycle facilities, and other Denver B-cycle stations in the network will experience
greater numbers of checkouts.
Aim 2: Describe the characteristics of Denver B-cycle users.
Aim 2,1: Determine demographic characteristics of Denver B-cycle users.
Hypothesis 2.1: Denver B-cycle users will be more likely to be male, Caucasian, more
highly educated, and have a higher household income than the general population.
Aim 2.2: Investigate factors influencing Denver B-cycle use.
7


Hypothesis 2.2: The number of Denver B-cycle checkouts by annual members will be
related to lifestyle factors indicating ability to replace car use with shared bike use.
Aim 2.3: Determine impacts of Denver B-cycle on shifts toward active transportation.
Hypothesis 2.3: Denver B-cycle annual members will shift mode choice away from car
use and toward active transportation via shared bicycles.
Aim 2,4: Identify active transportation benefits for Denver B-cycle annual members.
Hypothesis 2.4: Denver B-cycle annual members will exhibit a net increase in quantity of
active transportation.
Aim 3: Provide an in-depth description of the impacts of Denver B-cycle as a public
health intervention.
Overview of Methods
This dissertation research is a mixed methods quantitative/qualitative study of the
impact of a public bicycle sharing system as a health intervention for active
transportation behavior. The timeframe for the study was the first season of operation of
Denver B-cycle, in Denver, Colorado, from April to December 2010. The study uses
geographic and spatial data to examine how Denver B-cycle is integrated into the existing
infrastructural and social components of the city system, and to provide contextual
understanding of how the system is used. This investigation also describes use of the
system during its initial year, determines the socio-demographic makeup of users in order
to distinguish participating and non-participating groups, and explores factors influencing
use of the system. The study detects changes to active transportation behavior of annual
members, identifies net active transportation increase due to use of the system, and
8


describes the impacts of the system on individuals.
Quantitative findings derived from exploring the system usage of Denver B-cycle
are viewed through an evaluative lens informed by qualitative observations and in-depth
interviews of selected users of the system. The combination of quantitative findings and
more nuanced qualitative inquiry provides for a richer and more contextually sensitive
understanding of the effects of Denver B-cycle on the city.
9


CHAPTER II
LITERATURE REVIEW
Development Of The Current Built Environment
The Emergence Of Car Dependence
The prevailing shape of the urban form in the United States did not emerge on its
own; much has been designed to support private motorized transportation. Many of the
elements of the built environment incorporate features that, by intention, are deferential
to the primacy of the car. These physical features have induced the evolution of social
preferences, observable in behaviors of the population. As evidence, an accounting of
travel behavior in 2009 found that 40% of trips in the U.S. are less than two miles, and
27% of trips are less than one mile in length, though cars are the mode of choice for 87%
of trips under two miles and 62% of trips under one mile (Alliance for Biking & Walking
2012).
The present acute dependence on cars for transportation is the product of
deliberate efforts to change society. At the end of World War II, in his book entitled,
When Democracy Builds, visionary Frank Lloyd Wright voiced the aspirations of the
times when he imagined a society in which the city was decentralized, expansive, and
supported by ubiquitous personal car transportation (Wright 1945). An ideological
dichotomy between progression and regression formed around the mode of personal
mobility. Prevailing ideals framed the road to the future as broadly paved, populated with
speeding cars, and nearly without limit. Crowded cities with narrow streets full of
10


pedestrians were emblematic of the past. Ever since, this course of thinking has had vast
effects on the scale of urban development.
The ambitions of Wrights era have materialized in our present; today it is
completely accepted that nearly any daily activity involves or requires the use of a car. In
our society, expansive and omnipresent roads, along with cheap, plentiful and convenient
car parking are not only expected, but are enshrined in the policy of building codes,
municipal and regional plans.
Decoupling Transportation From Physical Activity
The city as a sprawling metropolis to support a culture of car dependence is a
contemporary phenomenon. Prior to the twentieth century and since the dawn of
humanity, most primary travel modes required physical activity. Physically active
transportation is any mode in which an individual expends his or her own energy for
locomotion, with walking and bicycling the most prevalent forms. Active transportation
may be a stand-alone activity or incorporated as part of a multi-modal trip, such as is the
case when an individual walks to a bus stop or rides a bicycle to a train station. Active
transportation fulfills two objectives at once; serving basic transportation needs while
integrating healthy physical activity into daily life.
Transportation modes have affected urban development throughout the course of
time. Cities developed in accordance with the Marchetti Constant, which posits the span
of a city is limited to no more than what can be traversed in one hour, historically by foot
(Marchetti 1994). The Marchetti Constant has held up to retrospective examination of the
evolution of cities throughout the world. However, in very recent times, the advent of the
fossil fuel age dramatically increased the distance of a one-hour travel budget, leading to
11


striking changes to many of the parameters of cities. First came the train, then the car,
each greatly expanding the breadth of cities (Newman and Kenworthy 2006).
The city is a complex adaptive system of interconnected actions and reactions,
between and among physical and social components. Changes in prevalent travel modes
and subsequent changes to urban geography resulted in substantial impact on perceptions
of comfort, utility, safety and appropriateness of modes. An individual consciously or
unconsciously considers these items when weighing active transportation against the
range of possible modes for any given trip. With the ubiquity of cars, a feedback loop of
rapid change developed between people choosing cars and a built environment to support
car use.
Ordinary travel needs were, in effect, decoupled from expenditure of physical
activity. Over time, this decoupling has led to unfavorable health outcomes. As preferred
travel mode choice shifted to cars, the daily amount of time spent being physically active
dropped (Frank, Andresen et al. 2004; Lopez-Zetina, Lee et al. 2006). This, in turn,
increased the relative risk of obesity; a factor linked with serious health concerns. As
active transportation activity dropped, obesity rates among adults skyrocketed, from 13%
in 1960 to 35% in 2009 (Alliance for Biking & Walking 2012).
The State Of Obesity And Why It Matters
The shadow of obesity looms large over the Unites States. Many of the factors
influencing the risk of obesity are associated with lifestyle, not only including levels of
physical activity and active transportation, but also a dramatically different nutritional
landscape than existed in previous generations (Glanz 2009). The lifestyles in which
many Americans live contribute to greatly increased risks. To put the situation in
12


perspective, by 2015, an estimated 75% of U.S. adults may be overweight or obese
(Wang and Beydoun 2007). A spectrum of elevated health risks is associated with
obesity, imposing considerable impact, not only on those individuals who are directly
affected, but also on society. Obesity and related conditions contribute disproportionately
to the challenge of maintaining a functioning system of healthcare.
As healthcare costs continue to increase and incidence of obesity within the
population escalates, reduction of costs associated with this preventable condition is of
paramount importance for economic sustainability. While ever more people become
obese, the expense of treating obesity-related diseases are expected to rise dramatically.
Total healthcare costs attributable to obesity are predicted to double every decade, to
account for 16 to 18% of total US health-care costs by 2030 (Wang, Beydoun et al.
2008).
A key reason for the high expense of obesity is the many associated chronic
health conditions, highlighted in Table II. 1. Several of these conditions are difficult or
expensive to treat, and contribute to elevated morbidity and mortality. Some of the most
serious and common obesity associated illnesses include cardiovascular disease, cancer,
diabetes, and osteoporosis (Warburton, Nicol et al. 2006). Faced with staggering health
and economic figures, effective interventions to reduce risk of obesity among large
portions of the population are essential.
13


Table II.l Obesity: Chronic Health Conditions and Risks.
Premature death High blood cholesterol
Type 2 diabetes Complications of pregnancy
Heart disease Menstrual irregularities
Stroke Hirsutism (presence of excess body
Hypertension and facial hair)
Gallbladder disease Stress incontinence (urine leakage
Osteoarthritis (degeneration of caused by weak pelvic-floor
cartilage and bone in joints) muscles)
Sleep apnea Increased surgical risk
Asthma Psychological disorders such as
Breathing problems depression
Cancer (endometrial, colon, kidney, Psychological difficulties due to
gallbladder, and postmenopausal breast cancer) social stigmatization
Adapted from: The Surgeon Generals Call To Action To Prevent and Decrease Overweight and
Obesity 2001 (Office of Disease Prevention and Health Promotion, Centers for Disease Control
and Prevention et al. 2001)
Obesity and related risks are closely associated with inadequate physical activity,
a modifiable risk factor. An increase in regular physical activity can significantly reduce
the relative risk for obesity-related chronic illnesses (Warburton, Nicol et al. 2006;
Haskell, Lee et al. 2007). The recommended quantity of daily moderate physical activity
for adults is 30 minutes a day, five days a week (Haskell, Lee et al. 2007), but as little as
15 minutes a day or 90 minutes a week has been found to be of benefit, even for high-risk
individuals (Wen, Wai et al. 2011). Unfortunately, more than half of U.S. adults do not
meet even the minimum recommended levels of physical activity (Centers for Disease
Control and Prevention 2005).
Obesity Risk Through The Urban Form
As beneficial as it may be, opportunities to participate in physical activity can be
elusive. Related to the consequences of car dependence, as previously discussed, the very
14


places in which people live often act as a barrier. Numerous studies have established
strong evidence between the physical makeup of the urban form and individual physical
activity (Berrigan and Troiano 2002; Handy, Boamet et al. 2002; Sallis, Frank et al.
2004).
An ecological study using data from the Behavioral Risk Factor Surveillance
System (BRFSS) found that the urban form affects engagement in physical activity and
subsequently health outcomes (Ewing, Schmid et al. 2003). Infrastructural elements that
reduce the viability of non-motorized modes affect the quantity and accessibility of active
transportation (Frank, Andresen et al. 2004; Frank, Saelens et al. 2007). Over the past 20
or more years, such studies have become relatively plentiful, building a solid evidence
base linking the built environment to levels of physical activity.
In application of the Marchetti Constant, the distance a car can traverse in an hour
is much greater than that of a pedestrian, so cities have spread into the surrounding
landscape. With burgeoning size, the scale of cities and the form of the built environment
encourage automobile use at the expense of physically active transportation (Handy,
Boamet et al. 2002). Car use for most types and distances of trips has become the
standard, shaping social norms regarding transportation mode choice (Cervero and
Radisch 1996). Feedback loops between the built environment and behavioral action
have, over time, created cultural expectations for car use over active modes of
transportation. These expectations are manifest in physical surroundings as high volume
streets, acres of parking lots, absent or vestigial sidewalks, and large distances from
residential areas to centers of employment.
In sum, historically plentiful opportunities for active transportation have all but
15


disappeared in much of the modern extended urban setting. Communities that are highly
car dependent exhibit limited supportive elements for active transportation behavior.
Many metropolitan areas in the U.S. that matured with the advent of car culture are now
typified by sprawl, accompanied by the byproduct of undesirable health outcomes among
citizens.
Beyond The Car: Public Bike Sharing
For good or bad, high car dependence and associated impacts are part of modern
life. However, elements in a complex adaptive system do not remain static. New
solutions are constantly generated to address new problems. In the past few decades, a
number of cities throughout the world have begun to meet the challenge posed by obesity
by promoting active transportation behaviors. Among a range of efforts in these cities,
several have implemented public bicycle sharing systems.
The introduction of shared bicycles into an urban environment is not a new
concept. Progenitors of public bicycle sharing appeared in postwar Europe (DeMaio
2003). Early bicycle sharing attempts were comprised of ad hoc assemblages of old or
disused bicycles, often loosely organized by idealists or college students. No strategies
for maintenance were incorporated; the bicycles were simply released into the urban
landscape to be used freely by anyone. Unfortunately, most efforts of this type quickly
fell victim to theft and vandalism.
Beginning the late 20th century, primarily in a few Western European cities, more
innovative systems evolved (Shaheen, Guzman et al. 2010). Experimentation and
technological advancement (DeMaio and Gifford 2004) led to mechanisms to induce
responsible behavior among users, greatly improving the viability of public bicycle
16


sharing, and strengthening its reliability as a practical form of transportation. Through
development, public bicycle sharing systems morphed from localized curiosities into
items of interest to leaders of large cities. As of 2011, at least 165 cities around the world
have implemented bike sharing (Shaheen and Guzman 2011). In the U.S., several cities,
including Denver, Minneapolis, Washington, Boston, New York, San Francisco,
Philadelphia, Chicago, and Portland have installed or expressed interest in public bicycle
sharing systems.
Although public bicycle sharing systems have received some high-profile
attention in the popular media, until very recently, little academic research has been
committed specifically to the subject. Previous to 2010, few peer-reviewed articles
mention public bicycle sharing either in relation to connection with transit (Pucher and
Buehler 2009), or as a potential form of transportation in the unspecified future (DeMaio
2003; DeMaio and Gifford 2004). Nevertheless, the situation is changing with rising
interest, with literature related to the health and social impacts of public bicycle sharing
growing.
Articles on bike sharing outside the realms of transportation or planning are rare.
The most substantial bike sharing publication to date that features a health component
examined the public bicycle sharing system in Barcelona, known as Bicing, from an
epidemiologic perspective (Rojas-Rueda, Nazelle et al. 2011). The study focused on all-
cause population mortality rates between car drivers and Bicing users, and found that for
Bicing users benefits outweighed risks, as determined by engagement in physical activity,
exposure to air pollution, and traffic incidents. Although the findings of the Barcelona
study begin to fill in some of the very large gaps of understanding associated with public
17


bicycle sharing systems, the authors acknowledged some serious methodological
limitations, largely due to lack of availability of data. Considerable data gaps in the study
include a demographic accounting of Bicing users and evidence-based estimates of car
use replaced by shared bicycle use. The Barcelona study is important as the first to take a
substantive look at health effects of public bicycle sharing, but it only touched the surface
of understanding the impact of public bicycle sharing on the health of users.
Another recent publication explored transportation behaviors among users of a
public bicycle sharing system in Hangzhou, China (Shaheen, Zhang et al. 2011). The
authors of the study conducted intercept surveys and found that many users are car
owners, suggesting the potential of bike sharing to attract modal share away from cars.
The Hangzhou study, among the first to have a substantial behavioral component, found
that users of the system exhibited a higher rate of car ownership than non-users,
suggesting that bike sharing appeals to car owners. This finding also suggests selection
effects, in that those with more money may be more likely to own cars as well as use
shared bikes. Although this study begins to flesh out an understanding of use, the authors
acknowledge a number of limitations, primarily that findings are not readily transferrable
to other locations or populations.
The health impacts of public bicycle sharing systems, and how such systems can
begin to diminish societal path dependence on cars for transportation are topics that, at
present, remain underrepresented in peer-reviewed literature. In spite of the data
limitation acknowledged in both the Barcelona and Hangzhou studies, the data generation
possibilities of public bicycle sharing systems are intrinsically greater than that of
traditional studies of bicycle use. Because the operation of bike sharing systems is
18


conducted through computer databases that record and track individual users as well as
the performance of the system at large, the data generated are much richer in detail than
that afforded by typical methods such as bike counts or surveys alone. Current bike
sharing systems now have the capacity to contribute new knowledge about active
transportation behavior among urban populations.
Guiding And Applied Theories For This Research
Denver B-cycle is an infrastructural element introduced into a city full of other
physical and social elements, all part of a complex adaptive system. In examination of
how individual agency results in outcomes, such as how Denver B-cycle is used, a
framework of supporting theory is helpful. This research uses systems theory for
guidance at a conceptual level, suited for examination of multiple influences on
population behaviors (Kay 2008). At the level of individual behavior, social cognitive
theory (SCT) offers insight in how observed behavior and social interaction between
individuals and within small groups affects behavior (Bandura 1989), and is used to
inform research methods and evaluation. Diffusion of Innovations (DI) theory (Rogers
2003) fits as a connective theory between the broader conceptual and individual levels, to
assist understanding of how Denver B-cycle use transfers from person to person within
the population, affecting systemic changes over time.
The holistic nature of systems theory affords a useful perspective in
understanding the complexity of interactions between multiple physical and social
systems and components. This is particularly useful in understanding the effects of
agency on resilience (Bohle, Etzold et al. 2009), in this case, the function of Denver B-
19


cycle over the course of the initial operating season. The ecological framework, which
shares many conceptual elements with systems theory has been suggested for evaluation
of infrastructural influences on active transportation behavior (Ogilvie, Bull et al. 2011).
However, systems theory has methodological limitations that make measurement and
testing specific constructs difficult at the individual level.
One study used SCT to investigate individual perceptions and social interactions
as factors influencing physical activity in university students. The self-regulation skills of
a group of 350 young adults attending a personal health class were assessed using a
survey instrument, and their self-reported physical activity was tracked over four weeks.
The study found that a combination of factors, including outcome expectancy, exercise
role identity, positive exercise experience, and social support among peers affect an
individuals self-regulation skills and therefore self-efficacy in maintaining physical
activity (Petosa, Suminski et al. 2003). These findings suggest that physical activity
behavior is in part determined by internalized perceptions and social interactions to which
individuals are exposed, either consciously or unconsciously.
Internal and external experiences shape the ability of individuals to learn and
apply skills for self-regulation, a key factor in achieving behavioral self-efficacy. In the
case of use of Denver B-cycle, individuals may weigh many issues when considering
whether to regularly use shared bikes for transportation. As in the case of the university
student study, internalized perceptions of bicycling, as well as the actions of an
individuals peer group are factors affecting active transportation behavior (Rovniak,
Anderson et al. 2002; Robertson-Wilson, Leatherdale et al. 2008).
DI theory provides a lens to explore how decisions of individuals propagate as
20


behaviors adopted within and among groups, as knowledge of or engagement in an
activity makes its way through a population. DI theory expands on many of the constructs
of SCT by identifying traits of innovators and early adopters; individuals who serve as
popular opinion leaders and who exert asymmetric influence on other individuals within
groups (Rogers 2003). In the context of Denver B-cycle, the diffusion within the
population of the concepts of shared bicycles in general, and how the system functions
more specifically, are integral to the implementation of the intervention.
Evaluating the effects of multiple social and physical factors on physical activity
and/or active transportation behaviors is facilitated with a conjunction of the
aforementioned theories. Methodologies incorporating combined theories as ecological
approaches have been used to inform development and evaluation of studies of active
transportation behavior (Pikora, Giles-Corti et al. 2003; Sallis, Cervero et al. 2006;
Shannon, Giles-Corti et al. 2006), to more fully appreciate influences on behavior at
multiple levels.
Informing the Study
This review identifies literature relevant to bike sharing as an intervention and its
potential impacts, and informs the study questions, aims and hypotheses. Relevant current
findings are presented, and gaps in knowledge of public bicycle sharing as a behavioral
health intervention are identified. Much is unknown about how public bicycle sharing as
is supported through existing elements of the built environment, the patterns of use and
influences on participation among users of bike sharing, and the effects bike sharing has
on the active transportation behavior and engagement in physical activity of users. The
21


remainder of this chapter presents themes that are germane to the pursuit of the primary
research question:
Does public bicycle sharing serve as a population-scale intervention to catalyze
active transportation behavior within a population?
Several threads of inquiry support this research. Each thread is supported by a
review of relevant literature, in which exposed gaps in knowledge are identified.
Intervention Through The Built Environment
Studies have established a strong foundation supporting the idea that the form of
the built environment influences transportation behaviors, which in turn affect health
outcomes (Ewing and Cervero 2001; Handy, Boamet et al. 2002; Cervero and Duncan
2003; Frank, Andresen et al. 2004; Beech, Fitzgibbon et al. 2011). The existence or
absence of urban form elements supportive of active transportation affects travel mode
choice, and therefore the potential for inclusion of physical activity into regular habits.
These effects are measurable in locales ranging from suburban (Ewing, Brownson et al.
2006) to urbanized areas (Pendola and Gen 2007). The fact that the majority of trips,
whether urban and suburban, are conducted by car underscores the mode preference built
into the present system, in which a hierarchy of mode choices establishes car
transportation at its peak.
A concentration on how elements of the built environment result in undesirable
effects is a commonality among publications in this arena. What is missing, or at least
understated in many studies of the built environment on physical activity, is that the
findings also indirectly suggest that the built environment itself can be harnessed to
induce desirable behaviors. A few interventions, however, have been designed around
22


how infrastructural changes can affect behavior.
One such study examined how improved public park facilities affected the local
population (Cohen, Golinelli et al. 2009). The Cohen et al. study found that
improvements to park infrastructure did not lead to increased physical activity. However,
as a physical activity intervention, a weakness in the design was overreliance on an
assumption that the targeted population would make the time to plan for and engage in
leisure-oriented physical activity at the park, or to otherwise become self-motivated
through the presence of park facilities. Although the study found the effects on increased
physical activity to be marginal, the authors note that marketing or programmatic
assistance to engage the community in park activities may have resulted in a more
positive outcome.
As was the case in the Cohen et al. study, physical elements often receive the
lions share of attention when considering effects of the built environment on physical
activity. However, social components, such as time availability, perceived safety and
other factors may also be important but are less well studied. James Sallis is a proponent
of examining the ecological context of behavioral influences on active transportation
(Sallis, Cervero et al. 2006). This approach posits that many factors at many levels exert
influence on behavioral action, and that the complexity of situational context should be
taken into account. The importance of contextualizing the risk factors and fundamental
causes of health outcomes (Link and Phelan 1995) is essential in the development of
appropriate interventions that will be received within communities and taken up by
populations at risk. Link and Phelan posit that examination of basic social conditions is
integral to understanding how social factors affect risk. It is preferential that interventions
23


designed to target a broad population dont unwittingly exacerbate rather than narrow
health disparities. Integrating multilevel approaches into the design of an intervention to
affect behavior at numerous leverage points within physical and social context can help
to mitigate the potential for undesirable widening of health disparities.
Multilevel approaches, incorporating intentionally designed elements of the built
environment, and supported by social interaction can be used to target specific behavioral
actions. One such study of the High Point public housing community in Seattle
incorporated community-based participatory research to develop a multilevel intervention
to increase physical activity, primarily for recreation and utilitarian purposes (Krieger,
Rabkin et al. 2009). Members of the community identified barriers to participation in
physical activity, and then devised methods for remediation. Participants identified
needed infrastructural improvements, such as restricted car parking near intersections to
increase visibility and improve the safety of pedestrians, and the development of
walking-oriented social group activities, effectively modifying social norms to include
more physical activity.
The High Point study revealed that an intervention involving members of the
community could help to create social support and to instill a meaningful sense of
purpose that fit the lifestyles of participants. In the development of interventions that
target populations, understanding the context in which the intervention is to occur is
essential. However, a shortcoming of the High Point study was that it was largely focused
on recreational physical activity through walking groups, presumably comprised largely
of those who had the leisure time availability to participate. The study did not improve
understanding of how active transportation habits of residents may have been affected.
24


Changes to the built environment are commonly implemented, ostensibly with
goals to increase physical activity. Municipalities have long sought to encourage physical
activity through development of parks, trails, on-street bike lanes, walking paths and
other features encouraging of physical activity, such as skate parks, and basketball or
tennis courts. Many studies have identified the presence of these and similar features as
being beneficial to health of a residential population (Ewing, Brownson et al. 2006;
Pendola and Gen 2007; Sallis, Saelens et al. 2009). However, direct behavioral effects of
specific infrastructure or amenities are seldom measured or studied.
Considering parks, trails, and other facilities to encourage physical activity,
infrastructure interventions designed to affect behavioral health through physical activity
are not uncommon. However, the Cohen et al. study and the High Point study are set
apart as examples of infrastructure interventions that include assessments of health
impacts specific to defined projects. These studies have begun to sketch a picture of how
interventions can be applied through infrastructure with the intent of affecting specific
health outcomes. Still, the Cohen et al. and High Point studies focus on small, defined
and localized populations and are designed to affect physical activity through leisure
activities. Moreover, unlike a public bicycle sharing system, they are not designed for
broad integration across a gamut of existing urban structures.
As discussed previously, obesity is a population-scale problem that may benefit
from a population-scale intervention. However, for that to happen, it is important that an
intervention be of a sufficient size and type to actually affect a population. As
interventions conducted through infrastructure with measured health outcomes are not
common, there are many important questions that remain. For example, do existing
25


physical and social elements of the built environment support an intervention delivered
through infrastructure? The present research will address this question.
On the social side of the continuum of influential factors, little is known about
how Americans in a car dependent society will react to public bicycle sharing. In the
High Point study, intervention efforts were designed to actively engage a defined
community and tailored to a specific audience. Such is not the case with a public bicycle
sharing system. Although the presence of pedestrians and human activity near potential
station sites was a factor when the Denver B-cycle operators considered station
placement, whether people would engage the system was unknown. Some potential users
are likely to quickly catch on to the concept; others may not.
A public bicycle sharing system is an element of physical infrastructure. Stations
and bikes are placed according to the plans of the administrators, but how well they
integrate with the existing urban landscape is unknown until operation commences. Some
literature has addressed various aspects of the use of privately owned bicycles in
conjunction with public transit (Rietveld 2000; Pucher and Dijkstra 2003; Martens 2004;
Givoni and Rietveld 2007; Martens 2007). Other research has focused on how urban
environments might be designed to support bicycle trip choice (Dill and Carr 2003;
Sallis, Frank et al. 2004; Dill 2009), but none directly addresses the interrelation of bike
sharing with existing transportation systems.
Intervention And The General Populations
The populations that have been subjects of study in physical activity interventions
vary demographically. However, most studies to date have targeted children or narrowly
defined adult populations. Denver B-cycle has the potential to target broader populations;
26


particularly young and middle aged adults.
Studies of groups delineated by age are common, with school-aged children
frequently targeted. Investigation of active transportation behavior among children and
teenagers has revealed connections between active commuting and risk of obesity. For
example, a longitudinal cohort study of kindergarten and first grade children in Quebec
found that active transportation to school was predictive of lower body mass index (BMI)
scores, an indicator of obesity (Pabayo, Gauvin et al. 2010). The Quebec study also found
that sustaining active transportation reduced the risk of obesity for children in the early
years of grade school.
Another study used a social-ecological approach to examine the influences on
active commuting to school among high school students in Ontario, Canada. In the
Ontario study, several variables, notably parental encouragement, smoking status, amount
of sedentary time, physical activity level and perceived weight status were significant
indicators of active transportation behavior (Robertson-Wilson, Leatherdale et al. 2008).
An interesting finding from the Ontario study was that the percentage of students who
commuted to school via active modes declined in grades 11 and 12, coinciding with the
age at which students became eligible to obtain a drivers license.
It is logical that children are often the focus of physical activity interventions
because of the importance of establishing healthy habits at a young age. At the other end
of the spectrum, aging and elderly populations have also commonly been subjects of
focus. As an example, a randomized controlled trial examined the impact of two different
variants of a physical activity intervention among population of Dutch adults aged 50 and
over. One group received tailored letters with physical activity advice. Another group
27


received enhanced information, encapsulating the same information given to the other
group, but were additionally provided with access to an online buddy forum with more
tailored information, and maps of bicycling routes in their community highlighting
walking and bicycling possibilities (Stralen, Vries et al. 2009). The Dutch study found
that the enhanced intervention strategy significantly improved bicycling behavior over
the basic information strategy. The increase in individual awareness of nearby bicycling
supportive facilities contributed to improved efficacy of the intervention.
Preventive interventions for groups at either end of the age continuum are
important. As with the study of active transportation behavior of children, a focus on
older populations is also logical. The percentage of the total population made up of aging
adults is on the rise, and treatment of preventable conditions among this group are an area
that may be targeted for health care cost reduction. It should be noted that Denver B-
cycle may be able to recruit users from the older population, but children under age 18
are not eligible to use the system.
The majority of Denver B-cycle users are likely to fall in age between young
adults and the elderly. Physical activity interventions for the portion of the population in
this range have also been studied. Many interventions target specific demographic
groups. For example, a physical activity intervention that focused on middle-aged urban
women in Germany found that a group that was taught self-regulation techniques was
substantially more active than a group that was only provided with health information
(Stadler, Oettingen et al. 2009). Groups targeted by race or ethnicity are also typical of
intervention efforts. As an example, a physical activity intervention designed to be
culturally and linguistically adaptive to Latina women found that contextual tailoring of
28


the intervention strategy was more effective in increasing physical activity than health
information alone (Pekmezi, Neighbors et al. 2009).
Some groups that have been the focus of physical activity interventions are
demographically diverse, but share other similarities within the group. Groups defined by
diagnosed health conditions, such as diabetes status are not uncommon. For example,
participants of a physical activity intervention in Florida who had all been diagnosed with
type II diabetes received tailored classes and information based on their current activity
levels and motivational readiness (Dutton, Provost et al. 2008). The results of the Florida
study showed no significant difference in physical activity or progression along the
stages of change outcomes as compared to usual care.
Other physical activity interventions that comprise more heterogeneous groups
make assumptions of participants, which, in effect create homogeneity of a different sort.
One such study examined the effectiveness of an email/internet based diet and physical
activity intervention among employees based at a large health care administrative
complex (Sternfeld, Block et al. 2009). The employer-based intervention found that the
web delivered intervention, which centered on a personalized web page where each
participant could track progress toward goals, increased self-regulation and adherence to
behavior over basic information alone. Although the employer-based study population
was mixed, all participants shared the same employer and work location.
By design, most of the aforementioned interventions include only people who
have made conscious decisions to participate, and are aware of the goal of increasing
physical activity. This is not an inherent weakness of such interventions, but overt
objectives toward physical activity might not be sufficient to appeal to members of the
29


general population at risk, or may even turn off a portion of the target population.
Although the majority of the studies of physical activity interventions are thoughtfully
designed and conducted, and are informative in reference to specific populations,
conclusions obtained through examination of more homogenous, often motivated groups
are not readily applicable to broader groups of people. Therefore, findings from these
studies may not necessarily apply to the present research.
Among studies of physical activity interventions, the least represented population
is that of mixed young to middle-aged working adults of the general population.
Coincidentally, this population is comprised of many who are presently healthy and who
might most benefit from a preventive health intervention, before drifting toward
increasing risk of obesity. Young to middle-aged adults in the U.S. are less active than
counterpart populations in other industrialized countries (Bassett, Wyatt et al. 2010). The
goal of Denver B-cycle is to target a range of people to prevent or help slow down rising
rates of overweight and obesity.
If successfully implemented, the impact of such an intervention will likely expand
beyond the initial scope, because the group of young to middle-aged adults encapsulates
many decision-makers, ranging in jurisdiction from the family level to the public policy
level. Affecting change in the physical activity behaviors of the general adult population
is essential to mediating the widening threat of obesity and related conditions.
A potentially important result of Denver B-cycles reach to younger and middle-
aged adults is that they are often embedded within strong familial and/or social systems,
where their beliefs and experiences can influence others. In fact, some intervention
studies have examined how family and social context can be used to induce change in
30


active transportation. Fitness and health can be linked to the familial environment,
including activity preferences (Gruber and Haldeman 2009). Children who grow up
observing and taking part in active transportation behaviors may continue to do so as an
adult. Family resources are also a factor for active transportation. Children of families
with lower household income have been found to be more likely to actively commute to
school (Babey, Hastert et al. 2009). Much of the focus of these and similar studies centers
on only a single trip type: commuting to school. Introducing active transportation to
school commuting habits is certainly important, but by concentrating primarily on school
commuting, opportunities for many other family or social utilitarian trips in which active
transportation can be encouraged may be missed.
As the largest segment of the working population, most young and middle-aged
adults commute as part of employment. A physical activity intervention to reach this
population would logically encourage active transportation as part or all of a commute, or
for incidental trips during the day. Public bicycle sharing is well suited for this purpose.
By integrating a public bicycle sharing into the transportation infrastructure of a city, a
convenient alternative to car use for short trips is offered to a population, simultaneously
introducing an opportunity for moderate physical activity among users.
The popularity of existing public bicycle sharing systems in many cities in the
world leaves little doubt that people do participate, but to date no literature exists on the
demographic makeup of users. Authors of the Barcelona study acknowledged using
central estimates of the population, as opposed to specific demographic data of users,
noting this item as an objective for future research (Rojas-Rueda, Nazelle et al. 2011).
The authors of the Hangzhou study found that many bike sharing users were working age
31


adults using the bikes for part or all of a commute (Shaheen, Zhang et al. 2011). Yet, the
cultural history of China in relation to utilitarian bicycling differs so greatly from the
equivalent population group in the United States so as to make findings less
generalizable.
Although the demographic makeup of public bicycle sharing system users has not
been studied, some is known about the demography of bicycle commuters and others who
use bicycles for utilitarian transportation, chiefly that the majority are male (Garrard,
Rose et al. 2007; Dill 2009), tend to be Caucasian, and have higher household income
than the general population (Alliance for Biking & Walking 2012). Beyond gender, race
and household income, other known demographic characteristics of bicycle users are
scarce. Denver B-cycle requires a credit card for participation, posing a barrier for those
at the lower end of the socio-economic strata. As commuting by bicycle is an activity in
which only 2.2% of the Denver population engage (U.S. Census Bureau 2011), this alone
is enough to set bicycle commuters apart from the general population, and underscores
the likelihood that users of bike sharing will also exhibit differences.
The gap in demographic knowledge of who uses a public bicycle sharing system
in a major U.S. city is large. This research will provide early estimates of the populations
that use and do not use Denver B-cycle in its first year of operation. Although Denver B-
cycle will reach a broader population of young and middle-aged adults than other
physical activity interventions, it is likely that users will not reflect the general population
in important ways.
The effects of environmental context
32


Throughout much of the world, bicycles are very common vehicles for everyday
transportation. In rapidly developing countries such as China and India, bicycles have
historically been and remain an important vehicle, even as car ownership increases
(Kenworthy and Hu 2002). In many countries, especially evident in Asia and Western
Europe, commuters ride bicycles as a solitary mode, or as an initial or final leg of a
journey in conjunction with public transportation (Martens 2004; Martens 2007).
Bicycles are even important vehicles for commerce, being used to carry passengers and
cargo (Pucher, Peng et al. 2007). In much of the world, in both developed and developing
nations, bicycles are firmly established within transportation systems as a vital
component.
The situation for transportation bicycle use in the Unites States is very different.
In the U.S., the percentage of commuters who choose to ride bicycles, referred to as the
bicycle commuter mode share, remains consistently low. As shown in Figure II. 1, the
2010 American Community Survey found that just 0.5% of U.S. commuters rode a
bicycle to work, a figure that has remained mostly unchanged nationally for years (U.S.
Census Bureau 2011).
33


United States Omaha Baltimore Memphis El Paso Fort Worth Kansas City Charlotte Jacksonville Dallas San Antonio Oklahoma City Detroit Louisville Nashville Las Vegas Indianapolis Houston New York Raleigh Colorado Springs Mesa Fresno San Jose Phoenix Milwaukee Columbus Virginia Beach 1 1 1 1
Atlanta Los Angeles

Austin San Diego

Sacramento Long Beach Chicago Albuquerque



Boston Philadelphia

Denver
Tucson Washington

San Francisco
Seattle
Poitland
o.c % l.C )% 2.C )% 3.C )% 4.C >% 5.C )% 6.C )% 7.C )%
Source: U.S. Census 2010 American Community Survey (U.S. Census Bureau 2011)
Figure II. 1 U.S. Cities Larger than 400,000, Ranked by Bicycle Commuting.
Exceptions to the low norm exist. Portland, Oregon achieved twelve times the
national average with 6.0% bicycle commuter mode share. By comparison, Denver
34


ranked sixth in the same category with 2.2% bicycle commuter mode share, more than
four times the national average (U.S. Census Bureau 2011).
As respectable as a 2.2% bicycle commuter mode share in Denver is compared to
other U.S. cities, it pales in comparison with Western European and Asian cities where
bicycles are used for 20% to 50% of all trips (Pucher and Dijkstra 2003; Pucher, Peng et
al. 2007). Installation of a public bicycle sharing system in Denver is not likely to
institute sweeping, immediate changes, but it will probably have some affect on bicycle
use in the city.
While it is impossible to consider all possible contributors, several key factors
may affect the active transportation behavior of users of Denver B-cycle. The perceived
safety of bicycling is a factor that influences active transportation behavior (Dill 2009).
Perceptions of safety are related to the type and availability of bicycle supportive
infrastructure and the form and characteristics of the built environment. As evidence,
higher levels of bicycle supportive infrastructure have been found to be positively and
significantly correlated with elevated rates of bicycle commuting (Dill and Carr 2003).
Bicyclists prefer to use infrastructure and facilities that are supportive of bicycling, even
to the point that increased travel time is considered to be worth the trade off to ride in a
marked bike lane as opposed to an unmarked street (Tilahun, Levinson et al. 2007).
Socio-economic factors, such as fuel prices, have an influence on transportation
mode choice. A study of recent dramatic escalations in fuel prices found that rises in fuel
price corresponded to a small, yet statistically significant increase in transit ridership
(Lane 2009). Although the fuel price study was focused on public transportation, the
observed changes in behavior may well translate to higher use of other modes, including
35


bicycles. The demand elasticity of fuel prices does not appear to carry large, long-term
effects, with research showing that for every 10% increase in gasoline prices there is a
corresponding 1.2% increase in transit use (Litman 2004; Currie and Phung 2008).
However, a prolonged and significant change in fuel prices could yield systemic effects.
As a type of public transportation of its own, and as a feeder mode designed to support
access to light rail and bus transit, Denver B-cycle is an option for commuters seeking
relief from fuel prices
The influence of weather on bicycling habits is less simple to understand when
comparing areas with similar climates, such as the U.S. and Europe. Many countries with
colder climates than much of the Unites States sometimes exhibit higher bicycle
commuting mode shares. For instance, Canadas Yukon Territory, at the same latitude as
Alaska, has more than two times the bicycle commuter mode share of California (Pucher
and Buehler 2006). Although Copenhagen is subject to cold and inhospitable weather
conditions, it is known for swarms of bicyclists, in contrast to U.S. cities at comparable
latitudes. However, no population is immune to the cold when it comes to outdoor
activities, and weather extremes have been found to reduce the propensity to ride a
bicycle (Richardson 2000). Adverse weather, however, has been found to affect the
quantity of recreation bicycle trips more than utilitarian bicycling trips (Koetse and
Rietveld 2008). This finding may signal that people who depend on Denver B-cycle for
transportation may continue to use the system even after recreational riders have largely
abandoned their use during cold weather.
The preceding key factors are only a few that may influence Denver B-cycle use.
When taken in context of multiple and unpredictable emergent properties of a complex
36


adaptive system, it is challenging to determine what will influence Denver B-cycle use in
its first season. However, given the paucity of research of bike sharing systems in the
U.S., the present research will help address a few important gaps.
Key gaps include understanding the rate at which use of shared bicycles will
occur, which geographic sections of the service area will experience the most use, and
whether there will be discernable differences in the way annual members use the system
as compared to short-term users. The present research will use quantitative and
qualitative methods to address these key gaps in knowledge.
Public Bicycle Sharing And Active Transportation
A societal shift from a long-established pattern of integrated active transportation,
to a high dependence on car transportation associated with urban sprawl has had
unintended consequences. Evidence has found urban sprawl to be associated with
elevated risk of obesity (Ewing, Brownson et al. 2006). Areas dominated by a single,
prevailing land use variant, such as vast residential subdivisions without integrated retail
or commercial elements, have also been found to have an adverse impact on obesity risk
(Cervero 1995; Cervero and Radisch 1996). When employment destinations are far
removed from areas of residence, car-dependent commutes are common. In the past, land
use that comingled residential and commercial areas enabled walking or bicycling
commutes, thus integrating beneficial physical activity into daily life. The present state of
commuting is dominated by car use, which contributes to sedentary time, decreases
integrated physical activity, and increases health risks. In satisfying commute needs, a
healthy behavior has been directly replaced by an unhealthy behavior.
The rising prevalence of obesity parallels the dominance of car culture. Between
37


1960 and 2008, active transportation dropped, while obesity rates steadily rose (Alliance
for Biking & Walking 2012). The effect appears localized largely to obesity status, rather
than overweight status, as the percentage of overweight population remained fairly stable
during the same period (Ogden and Carroll 2010). Links between car use and risk of
obesity have been identified, with an important study finding a 6% increase in obesity
risk for each hour spent in a car daily (Frank, Andresen et al. 2004).
The quantity of car vehicle miles traveled (VMT) has expanded in tandem with
obesity rates, a phenomenon not limited to the United States, but also observed in other
countries with elevated car dependence (Jacobson, King et al. 2011). The Jacobson et al.
study found a six-year lag time between increased VMT and increased obesity rates,
displaying evidence of the cumulative risk effect posited in the Frank, Andresen et al.
study of 2006. This is an environmental change in lifestyle, and like that of an increase in
caloric availability in the modern diets as compared to the past (Glanz 2009), contributes
to cultural and contectual influences on the overall risk of obesity.
Health risks attributable to the influence of car dependence extend beyond the
physical occupation of a car, and are embedded within the physical and social fabric.
Generations of people have been enculturated with social norms embracing the use of
cars to fulfill nearly any transportation need. As evidence, areas designed or developed
with integral car dependence have been found to exhibit higher risks of obesity within the
local population (Berrigan and Troiano 2002; Ewing, Schmid et al. 2003; Lopez-Zetina,
Lee et al. 2006). Such findings reveal, in part, the complexities that have developed
surrounding societal car dependence, and suggest reducing that dependence as a primary
intervention objective.
38


Ongoing research has identified agendas for physical activity and active
transportation that coincide (Sallis, Frank et al. 2004). In the disciplines of planning and
public health, mixed land use is a strategy to improve conditions for walking and
bicycling transportation by providing common destinations close to residences. The
concepts of Transit Oriented Development and New Urbanism incorporate mixed land
use among residential, commercial and office spaces to ensure likely destinations are
within range of non-motorized modes, using the form of the built environment as a
positive behavioral determinant (Handy, Boarnet et al. 2002). The incorporation of active
transportation strategies within community design holds promise both for planners and
engineers seeking pragmatic solutions to move people from place to place, as well as
policy makers and public health professionals who desire to improve the health status of a
population.
A transition from motorized to non-motorized modes results in health benefits to
individuals and to communities, in terms of physical activity, air quality and economics
(Rabl and Nazelle 2012), although active transportation does increase risk of injury from
collisions as compared to motor vehicle users (Reynolds, Winters et al. 2010). Bicyclists
are indeed more physically at risk and exposed to exhaust pollutants as compared to car
drivers, but when health benefits of being physically active are weighed against health
risks, the results indicate, on average, a strong net benefit for participation in active
transportation (Hartog, Boogaard et al. 2010).
The Barcelona study assumed a very high rate of car trips replaced by shared
bicycle trips. The key assumption by the researchers was that 90% of users were new
bicyclists who shifted their use kilometer-for-kilometer from car trips (Rojas-Rueda,
39


Nazelle et al. 2011). While it is possible that this may be the case, the figure seems overly
optimistic on two major counts: in assuming a vastly altered and solidly maintained
behavior of new bicyclists almost completely eschewing car use, and that an enormous
majority of shared bicycle trips are replacing car trips.
The Hangzhou study made no assumptions about the replacement of car use by
shared bikes, finding that a shift from car trips to shared bikes occurred, but also that
shifts from other modes including transit, walking and biking to shared biking also
occurred (Shaheen, Zhang et al. 2011). Although the authors note that the historically
high bicycle transportation rates in China have been declining, car ownership remains
much lower than in the U.S. Although the findings of this study are interesting and useful
within the context of the study environment, the cultural and infrastructural differences
between China and the U.S. do not make direct comparison possible.
It remains important to understand the modal shift capabilities of public bicycle
sharing. However, in order to evaluate the outcome of a public bicycle sharing as a health
intervention to increase active transportation, identifying any net change in active
transportation behavior among users is essential. In the present research, the focus is on
how Denver B-cycle has changed active transportation behavior, that is, the extent to
which shared bike trips are actually replacing car trips.
Some of the trips made by shared bike are likely to directly replace car trips,
resulting in a net increase in active transportation behavior. Annual members are more
likely to be habitual users than short-term users for a number of reasons, in part based on
ability to replace car trips. Therefore, it is hypothesized that use among annual members
will show a net increase in active transportation behavior.
40


Public Bicycle Sharing Participation
An increased risk of obesity affects a large portion of the population. Any
intervention designed to reach the greatest number of people must by necessity be
inexpensive per capita in implementation. Evaluation of physical activity interventions
suggest that behaviors develop in complex social and ecological contexts, and that to be
cost effective, interventions should be designed to take advantage of specific contexts
(Roux, Pratt et al. 2008), in which participants discover their own reasons to engage.
Therefore, individuals in the targeted population for Denver B-cycle will be largely
responsible for developing their own motivation to initiate and continue participation in
the intervention.
Denver B-cycle differs from many previous physical activity interventions in
several ways. Participants may or may not be aware of any health benefit to using the
system. Upon subscribing to Denver B-cycle, users are not provided health information
or asked to be involved in a health study. Encouragement for motivation to participate is
not delivered through coaching, counseling, or individually tailored goals for weight loss
or physical activity. Instead, Denver B-cycle is simply framed as an inexpensive and
convenient alternative to a car for short urban trips.
The framing strategy sets Denver B-cycle as a lifestyle intervention. Users of the
system, whether consciously or unconsciously aware of the objectives of the intervention,
procure their own motivation to self-select to initiate action and to self-regulate
continuance of active transportation behavior through use of Denver B-cycle. As opposed
to dependence on external health information or counseling to lead to action, individual
participants devise their own contextually meaningful reasoning for using Denver B-
41


cycle, finding a place for it within their lives.
Lifestyle interventions are designed to fit within the daily routine of targeted
populations, often using physical activity to serve purposes other than that of overt
pursuit of fitness (Dunn 2009). Lifestyle interventions for physical activity can
encompass everything from using the stairs instead of an elevator to commuting by
bicycle. A study of the Active-For Life Initiative, a lifestyle physical activity
intervention, revealed that physical activity behaviors incorporated into lifestyle are more
easily maintained (Wilcox, Dowda et al. 2009). The study also suggests that lifestyle-
based interventions are a promising conduit through which to deliver evidence-based
interventions at a community level, an objective that is logistically or economically
elusive with individually tailored physical activity interventions.
A key motivator for initial and continued use of Denver B-cycle is the perception
of benefit from using the system. A cost-benefit study of shifting from car use to active
transportation revealed benefits outweigh costs from an economic standpoint (Rabl and
Nazelle 2012). Costs of car ownership versus bicycle ownership, car parking versus
bicycle parking, and perceptions of time budgeting play into decision-making. Active
transportation can be more convenient than car transportation, and perceived convenience
contributes to adherence of behaviors (Lewis, Marcus et al. 2002; Dunn 2009).
Leveraging favorable aspects of active transportation against unfavorable aspects of car
transportation serves to attract participants and reinforce benefits to keep people
motivated.
Motivation for initial and continued use is central to the design of Denver B-
42


cycle. The system has been designed to facilitate motivation through increased
convenience, accessibility and awareness. The majority of Denver B-cycle stations are at
or near popular destinations in central downtown, and in areas that have high
concentrations of residents near downtown, and at major transit stops. In addition,
Denver B-cycle is promoted through media, primarily through news outlets who covered
the launch and operation of the system. Bicycle advocacy and recreational groups in the
community also quickly adopted Denver B-cycle into existing events and activities,
indirectly promoting the system. Through a combination of visible physical presence and
media attention, people who spend time in or near central downtown Denver are likely
have been exposed to Denver B-cycle.
Denver B-cycle also provides tools to support the self-regulation of users. The
system website features personalized user pages to help users monitor their progress and
track totals. The tools include estimates of miles traveled, calories burned and dollars
saved through use of Denver B-cycle, per trip and aggregated. Personalized web pages of
the sort used by Denver B-cycle have been linked to self-regulation skills and adherence
to changed behavior (Stemfeld, Block et al. 2009). Users must have access to a computer
connected to the internet in order to gain the full range of self-monitoring tools as a
benefit of using the system.
It is likely that use of Denver B-cycle will fit the lifestyle of some users, though
specifics of that fit are unknown. Identifying some of the reasoning for behavioral
motivation and action of users can aid the understanding of pathways and barriers to use.
Knowing how people use the system can assist evaluation of the effectiveness with which
the intervention has been implemented, and provide clues for future application and
43


improvement to reach a larger share of the population. Gaps in these areas will be
addressed using a quantitative/qualitative mixed methods approach.
Motivation To Use Public Bicycle Sharing
As a health intervention to induce active transportation behavior, Denver B-cycle
relies on SCT constructs of modeled behavior and social interaction to expose members
of the population to the intervention, and to spread use from individual to individual. As
an encapsulating theory, DI informs the examination of how the use of Denver B-cycle
penetrates the population through word of mouth, shared experiences of participants, and
the observation of participant behavior by the pool of those who are not yet participants.
Within the population, individuals classifiable in diffusion theory as innovators
are at the vanguard of adoption of new behaviors, followed closely by early adopters,
who take on a new behavior as a result of the influence of innovators (Rogers 2003). In
its initial year of operation, many Denver B-cycle users are likely to fit the profile of
innovators or early adopters; risk takers who are willing to try new things. Although the
innovators and early adopters of a new behavior reside at the narrow front end of the s-
shaped adoption curve of behavioral uptake (Henrich 2001), the reasoning for initial
motivation and adoption of a new behavior by an innovator forms the kernel of reasoning
for transfer of that new behavior to others.
Identifying what compels early users to convince their peers to try something new
is important in understanding how the intervention is diffused. Commonalities among the
beliefs and opinions of innovators and early adopters of Denver B-cycle may signal the
advent of modified social norms more accepting of bicycle transportation. In a society
heavily dependent on cars for transportation, any pathway to broad or meaningful change
44


will be paved by changes to social norms (Lucas 2009).
Diffusion of behaviors that contribute to modification of social norms is key to
growing and sustaining the intervention. Discovery of if and how use of Denver B-cycle
is diffused within the population can begin to provide a glimpse into the momentum of
the intervention as use spreads within the population. However, in terms of how diffusion
occurs with regard to public bicycle sharing as an intervention, little is known.
Through primarily qualitative methods, the present research explores the
motivational reasoning behind the behaviors of innovators and early adopters of Denver
B-cycle to determine if and how diffusion to the broader population is occurring. Because
many users in the initial year will be innovators and early adopters, members of this
group will likely be at the fore of the adoption curve, actively promoting the intervention.
Some of the initial users will be popular opinion leaders, with the power to influence their
peers. These early users will not only be participants in the intervention, they will
comprise a cadre of socially connected leaders who play an important role in the
diffusion of key concepts and uptake of active transportation behaviors.
Summary Of Literature Review
Evidence established in the literature provides an overview of how, over the
course of the past century, the population of United States has become increasingly
dependent on car use to satisfy transportation needs. Infrastructure, policies and social
norms have adapted to support cars at the expense of active transportation modes. During
the same time, physical activity levels among the population have declined, to the present
point at which the average person does not meet minimum recommendations. The
45


intensity of car dependence has coincided with a rise in the incidence of obesity, which
affects a majority of the population.
A substantial body of literature has detailed the link between inadequate physical
activity and obesity, and associated chronic health conditions, many of which are
expensive or difficult to treat. Past interventions to affect physical activity have often
used individually tailored strategies concentrated primarily on homogenous subgroups
within populations. However, these strategies are less well suited to counter a population-
scale problem. Instead, a broad engagement of the general population is more appropriate
for addressing the obesity problem, which, directly or indirectly, affects everyone.
A small but growing body of literature suggests that elements of the built
environment may be used to deliver an intervention to affect physical activity. Other,
related literature points to the effectiveness of lifestyle interventions for active
transportation as readily supportive of maintainable changes to behavior. A public bicycle
sharing system, such as Denver B-cycle, is a lifestyle intervention for physical activity
applied through infrastructure. In such an intervention, individuals develop their own
motivational reasoning to participate. By positioning shared bikes as a convenient active
transportation alternative, it is possible to shift car use to transportation bicycling,
resulting in a net increase of physical activity.
However, as a recently developed concept, little is known about public bicycle
sharing in application, and many specific gaps in knowledge exist. Theories of behavior
change including SCT, DI and complex systems theory have helped to identify gaps in
knowledge and to inform the development of methods of the present research by framing
the complex environment in which Denver B-cycle is installed. Each of these theories
46


contributes to informing the present research as to the social contexts in which
transportation decisions are made. These theories facilitate the detection and exploration
of the processes by which behavior among individuals achieves motivation to action, is
displayed and observed in social context, and is adopted by and/or transmitted to other
individuals. Interconnections between physical and social elements in the surrounding
environment support the operation of Denver B-cycle as a multilevel intervention.
Substantial gaps addressed by this research include:
Understanding in more depth how a health intervention applied through the
built environment is supported by existing infrastructure
Identifying participation in the intervention among demographic groups of the
general population
Understanding the impact of environmental behavioral influences on users of
Denver B-cycle
Assessing the impact of Denver B-cycle on the active transportation behavior
of the population
Understanding the motivational reasoning for participation in the intervention
Examining the diffusion of Denver B-cycle as an intervention over time
47


CHAPTER III
METHODS OF DATA COLLECTION AND EVALUATION
Study Design
A mixed methods research design was used for this study, with initial emphasis
on quantitative data collection and subsequent qualitative data collection to inform
quantitative findings. Quantitative data were collected from multiple sources, including
the Denver B-cycle system database, geographic information system (GIS) layers, a
survey of Denver B-cycle users, and from population resources, such as the U.S. Census
bureau, the Colorado Department of Public Health and Environment, and the Centers for
Disease Control and Prevention. Qualitative data were collected through site observations
and semi-structured in-depth interviews of select Denver B-cycle participants.
Use of mixed methods strengthens analytical power and affords cross-validation
of findings (Neuman 2003). Initial findings are used to inductively generate hypotheses
addressed through data analysis. Qualitative data serve to enrich understanding and to
apply contextual meaning to findings derived from quantitative data. This is especially
important in helping to determine why users exhibit particular behaviors. Answering
questions as to why something occurs is essential when considering behaviors and
processes that occur in complex adaptive systems because of the vast array of possible
causal pathways.
48


Brief Glossary Of Terms Specific To This Research
B-cycle: B-cycle is an equipment vendor for public bicycle sharing systems. B-
cycle is a consortium originally comprised of Trek Bicycle Company, Humana
Healthcare and Crispin Porter+Bogusky, an advertising firm.
Denver B-cvcle: A separate entity from B-cycle, Denver B-cycle is the name of
the public bicycle sharing system in Denver
Denver Bike Sharing: A non-profit organization that owns and operates Denver
B-cycle.
Study Setting
The study setting of the present research is the service area of the Denver B-cycle
public bicycle sharing system, located in Denver, Colorado, shown in Figure III. 1. The 40
Denver B-cycle stations located in the central business district of downtown Denver were
the primary focus of this study. Outlying groupings of Denver B-cycle stations in the
Cherry Creek and Denver University neighborhoods, each encapsulating five stations,
were not principal to the investigation. These smaller groupings are not well connected
with the downtown major group of stations, and exhibit different operational
characteristics than stations in the downtown group. GIS and other geographic data
collection concentrated on the City and County of Denver and the operational area of
Denver B-cycle.
49


Figure IIE.l Map of Denver B-cycle Stations in the Central Downtown Area, within
the City and County of Denver (2010).
50


Data Sources
A number of data sources were used in the present research. Table III. 1 shows
data sources for specific aims of the study.
Table III.1 Data Sources Associated with Study Aims And Hypotheses.
Study Aims Data Sources
Aim 1: Examine the physical characteristics and human activity in the built environment in which Denver B-cycle stations are situated. Station site observational data
Aim 1.1: Identify features that distinguish low and high performing sites. Denver B-cycle system usage dataset Station site observational data
Aim 1.2: Investigate the integration of Denver B-cycle within the existing urban transportation system. Hypothesis 1.2: Denver B-cycle stations having greater integration with transportation infrastructure, as indicated by proximity to public transportation stops, on and off-street bicycle facilities, and other Denver B-cycle stations in the network, will experience greater numbers of checkouts. City of Denver geographic information system (GIS) data Regional Transportation District (RTD) GIS data
Aim 2: Describe the characteristics of Denver B-cycle users. Denver B-cycle system usage dataset City of Denver geographic information system (GIS) data
Aim 2.1: Determine demographic characteristics of Denver B-cycle users. Hypothesis 2.1: Denver B-cycle users will be more likely to be male, Caucasian, more highly educated, and have a higher household income than the general population. Denver B-cycle user survey dataset 2010U.S. Census Denver County demographic data 2010 U.S. Census American Community Survey Denver County data 2010 Colorado Behavioral Risk Factor Surveillance System (BRFSS) dataset
51


Aim 2.2: Investigate factors influencing Denver B-cycle use. Hypothesis 2.2: The number of Denver bi- cycle checkouts by annual members will be related to lifestyle factors indicating ability to replace car use with shared bike use. Denver B-cycle user survey dataset
Aim 2.3: Determine impacts of Denver B- cycle on shifts toward active transportation. Hypothesis 2.3: Denver B-cycle annual members will shift mode choice away from car use and toward active transportation via shared bicycles. Denver B-cycle user survey dataset
Aim 2.4: Identify active transportation benefits for Denver B-cycle annual members. Hypothesis 2.4: Denver B-cycle annual members will exhibit a net increase in quantity of active transportation. Denver B-cycle system usage dataset Denver B-cycle user survey dataset
Aim 3: Provide an in-depth description of Denver B-cycle and its impacts as a public health intervention. Semi-structured in-depth interviews of select participants Denver B-cycle user survey dataset
52


Data Variables
Each data source used in the present research included numerous variables, shown
in Table III.2.
Table III.2 Data Variables Used in this Research, Arranged by Source.
Data Sources Variables
Station site observational data (Collected by PI) See Appendix A for site observation data collection instrument Collected during four one-hour periods at each station site: Bicycle count Presence of graffiti, litter and vandalism Notes on observed activities Description of area around site Behavior mapping of pedestrian and bicycle routes through site area
Denver B-cycle system usage dataset (Collected automatically through Denver B-cycle system software. De-identified dataset provided by Denver Bike Sharing) Collected during subscription: Date of purchase Type of subscription (24-hour kiosk, 24- hour online, 7-day, 30-day, annual) De-identified addresses of annual members Collected for each checkout session: User identification number Date and time of checkout Date and time of check-in Duration of checkout Station of checkout Station of check-in Derived from Denver B-cycle system usage dataset (system-wide): Total checkouts, by station Total subscriptions by type GPS data from a subset of shared bikes equipped with GPS units GPS speed and motion logs
53


City of Denver geographic information system (GIS) data (Retrieved from DenverGIS online repository) Denver County GIS layers: Base geographic features Streets and highways Street names and address locations Bicycle facilities Denver B-cycle station locations Layers created using ESRI ArcGIS 9.0 software Network distance to Denver B-cycle stations at 150, 300 and 500 meters Network density of Denver B-cycle stations at 150, 300 and 500 meters Geo-coded addresses of Denver B-cycle annual members
Regional Transportation District (RTD) GIS data (Retrieved from RTD online repository) RTD GIS layers: Transit stops: o Bus stops o Light rail stations
54


Denver B-cycle user survey dataset (Collected through online SurveyGizmo survey via link sent from Denver Bike Sharing to Denver B- cycle users) See Appendix B for Denver B- cycle User Survey instrument Demographic data: Gender Age group Home ZIP Code Ethnicity Race Educational attainment Household income Health data: Self-assessed health status Self-reported weight Self-reported height Self-assessed poor physical health in past 30 days Self-assessed poor mental health in past 30 days Self-assessed days disrupted by poor physical or mental health in past 30 days Bicycle ownership and skill level data: Bicycle ownership Self-assessed bicycle skill level Bicycle transportation data: Bicycle transportation activity Number of Denver B-cycle checkouts per week Types of Denver B-cycle subscriptions purchased Active transportation behavior: Shifts from car use to Denver B-cycle, bicycling and walking Active transportation activity in previous 30 days Transit-related data: Multimodal bicycle/transit use Multimodal B-cycle/transit use Transit use dependent on bicycle access Bicycle safety data: Helmet use on private bicycles and Denver B-cycle bikes Other transportation data: Car ownership/access Commute mode choice
2010 Colorado Behavioral Risk Factor Surveillance System (BRFSS) dataset Health status Poor physical health days in past 30 days Poor mental health days in past 30 days
55


2010 U.S. Census Denver County demographic data Population total Gender Ethnicity Race
2010 U.S. Census American Population by age groups
Community Survey Denver County Educational attainment
data Household income
Semi-structured in-depth interviews of Transcribed and coded text from interviews
select participants
Geographic Data Overview
GIS data were retrieved from online repositories maintained by DenverGIS, an
agency of the City and County of Denver, and the Regional Transportation District
(RTD), the local transit authority. The PI collected geographic observational data at
Denver B-cycle station sites, outlined in Table III.2. Refer to Appendix A for the station
observation data collection instrument.
Denver B-Cycle System Usage Dataset Overview
As shown in Table III.2, variables used for this research are generated and
collected through normal operation of the Denver B-cycle system. Denver B-cycle has
two major subscription categories, annual members and short-term users. Annual
members must register online through the Denver B-cycle website, during which time
they are required to provide contact data, including an address of record. Short-term user
types include a 24-hour pass purchased at a kiosk, a 24-hour pass purchased online, a 7-
day pass purchased online, and a 30-day pass purchased online. The Denver B-cycle
system database automatically records the subscription data, including the date, time and
subscription type of each sale. The system automatically aggregates the number of
subscriptions sold by type.
56


Annual members are given a B-card with an integrated RFID chip that enables
quick, automated checkouts of shared bikes at any station. Short-term users must check
out bikes through the station kiosk using a credit card. Each time a bicycle is checked out,
the system software associates the specific user with a specific bicycle and automatically
records the time, date, and station location of checkout and check-in, as well as the
duration of checkout. These data are collected and aggregated every time a user checks
out a shared bike, per individual user, per station and system-wide, maintained with
proprietary software and stored on the B-cycle system secured data server.
Each of the preceding data items were extracted from the B-cycle database into a
de-identified dataset by Denver Bike Sharing, and provided to the PI for the present
research as comma separated values (CSV) documents.
The B-cycle branded model of bicycle owned and used by Denver Bike Sharing
was designed to have an integral GPS unit. However, during the initial year, most of the
bikes did not have operational GPS units on-board. However, during September and
October 2010, B-cycle equipped a test group of seven shared bikes with GPS units that
logged route, direction and speed data during checkout. These test bikes were put to
service in the fleet of shared bikes and randomly logged GPS data. These GPS data were
aggregated and provided to the PI by B-cycle.
Denver B-Cycle User Survey Dataset Overview
In September 2010, an online survey was administered to Denver B-cycle users
who had registered contact information through Denver Bike Sharing. Data variables are
shown in Table III.2, and the complete survey instrument is included in Appendix B. A
link to the survey conducted online through Survey Gizmo was distributed via the Denver
57


Bike Sharing email newsletter contact list, amounting to approximately 2,000 individuals.
Valid survey responses totaled 599, for a response rate of 30%. The survey collected
information regarding demography and other characteristics of Denver B-cycle users. An
overview of subsections of the survey follows, with a description of the derivation of
survey questions, as well as reasons for inclusion in the survey.
Demographic information: Denver B-cycle User Survey participants made up the
demographic sample of Denver B-cycle users. Respondents to the survey were
largely comprised of annual members, with a smaller portion of short-term users
who subscribed to Denver B-cycle online, or otherwise registered to receive the
Denver Bike Sharing monthly email newsletter. Survey participants self-selected
to participate in the study by completing an online survey of users administered in
September 2010. Questions and response categories were adapted from similar
questions asked by the U.S. Census Bureau, so that survey data would be
comparable with 2010 U.S. Census data and 2010 U.S. Census American
Community Survey data, as listed in Table III.2.
Health status and health-related quality of life: Questions and response categories
were the same as asked in the 2010 Behavioral Risk Factor Surveillance
System (BRFSS) Questionnaire (Centers for Disease Control and Prevention
(CDC) 2009), so that survey data would be comparable. Health data elements
used included self-assessed health status, poor physical health in the past 30 days,
and poor mental health in the past 30 days. Data variables are presented in Table
III. 2.
Bicycle ownership and skill level: The PI developed questions and response
58


categories for this section regarding bicycle ownership and self-assessed bicycling
skill level. Bicycle ownership indicates whether users have the option to use their
own bikes instead of shared bikes. The ability to competently ride a bicycle in
various environments, including in traffic, are necessary and valuable skills for
using a public bicycle sharing system. The confidence of participants in their
bicycling abilities is an important indicator of competency. Table III.2 shows data
variables for this category.
Bicycle transportation use: The PI developed questions and response categories
for this section regarding use of bicycles as transportation. Measures of the use of
bicycles for transportation are important in the evaluation of the intervention on
active transportation behavior. The survey response dataset is not linked to the
Denver B-cycle system usage dataset for specific individuals. Therefore, as an
indicator of use, survey respondents were asked to report their number of weekly
checkouts of Denver B-cycle. Data variables associated with this category are
presented in Table III.2. The questions were developed because no appropriate
questions could be located in existing survey instruments.
Active transportation: Questions and response categories the same as those used
in a survey conducted by the New York Department of Health and Mental
Hygiene (New York City Department of Health and Mental Hygiene 2009). An
understanding of existing active transportation behavior among survey
respondents was important in order to identify established tendencies of
participants toward engaging in active transportation. Data from this section were
used to develop a car use replacement modifier to assess whether Denver B-cycle
59


reduced car use in favor of shared bicycles. A listing of data variables for this
category appears in Table III.2.
Bicycling to transit: The PI developed questions and response categories for this
section regarding joint use of bicycles and transit. Data variables are shown in
Table III.2, and include items assessing whether respondents accessed transit via
Denver B-cycle or other bicycles, and if a bicycle/transit link enabled transit use
among respondents. Literature suggests mutual support between bicycling and
transit (Martens 2004). Therefore, data collection from users was important to
detect multimodal linkages in the behavior of Denver B-cycle users. The
questions were developed because no appropriate questions could be located in
existing survey instruments.
Bicycle safety: The PI developed a question and response categories for this
section regarding bicycle helmet use. Data variables for this category are listed in
Table III.2, and assess whether respondents wear helmets while riding bicycles.
The use of helmets to improve safety while bicycling has long been a subject of
debate (Thompson, Sleet et al. 2002). However, during observation, helmet usage
among Denver B-cycle users was observed as being low. Including this section on
the survey allows for inquiry into helmet use behavior among Denver B-cycle
users. The question was developed because no appropriate question could be
located in existing survey instruments.
Other transportation: The PI developed a question for this section regarding car
ownership. Establishing car ownership status among respondents reveals whether
Denver B-cycle users have access to a car when choosing a transportation mode.
60


A question regarding commute mode choice was adapted from the 2010 U.S.
Census American Community Survey questionnaire (U.S. Census Bureau 2010),
in order to identify commute characteristics among Denver B-cycle users for
comparison with the general population. Data variables appear in Table III.2.
Select Participants
Following the survey, a subset of individuals, referred to as select participants,
was identified based on survey responses and were recruited for additional data
collection. Of 599 completed surveys, 317 respondents (52.9%) indicated a willingness to
be contacted for in further research. From this group, twenty-two individuals were
recruited through email and telephone. Select participants were purposively chosen based
on survey responses to weekly number of Denver B-cycle checkouts.
Human Subjects Review
The Colorado Multiple Institutions Review Board granted approval to this project
as Protocol #10-0690. One amendment to the original protocol was filed and approved to
extend the operational dates of the study.
In accordance with the study protocol, informed consent was obtained from all
participants selected for interviews as they were enrolled. Signed consent forms were
stored in a secure file cabinet. Randomly selected five digit numbers were used to
identify study participants. A secured server was used for storage of digital data.
Methods of Data Collection and Evaluation, by Specific Aim
In the following, the methods of data collection and analysis for this investigation
61


are presented, organized according to the Aims of the study.
Aim 1: Examine the physical characteristics and human activity in the built
environment in which Denver B-cycle stations are situated.
Aim 1.1: Identify features that distinguish low and high performing sites.
Data collection: Aim 1 and Aim 1.1 were concurrently addressed using station
site observational data. Following the first three months of operation, Denver B-cycle
system usage data for the total checkouts per station were compiled and evaluated to
identify high and low performing stations based on quantity of checkouts up to that time
The subset of select stations was chosen for closer examination of characteristics of the
surrounding built environment and observable behaviors, in order to understand in more
depth some of the factors that may contribute to station performance. The five highest
and five lowest performing stations in the downtown group were designated as select
stations, listed in Table III.3.
Table III.3 Denver B-cycle Select Stations.
Station Name Select Station Performance Designation
Market Street Station High
16th and Little Raven High
REI High
19th and Pearl Street High
Union Station High
Denver Health Low
Five Points Low
15th and Tremont Low
25th and Lawrence Low
Pepsi Center Low
Over the course of the 2010 season, station rankings by checkout changed
somewhat. For consistency, select stations remained the same as those initially chosen.


Site observations of each select station during observation intervals were conducted on
Tuesday, Wednesday or Thursday, so that normal weekday activities could be observed.
The four intervals of site observation were, August 25 and 26, September 28 and 30,
October 28 and 29 and December 14 and 15. Observations were conducted for one hour
at each site during each interval, between the hours of 7:00am and 6:00pm. Individual
stations were observed at different times during successive intervals so that activities at
different times of the day could be recorded. Days with adverse weather were avoided to
reduce weather related effects on observed activities.
Appendix A contains the instrument for select station observation. A satellite
photo of each station location was retrieved from Google Maps and inserted into the map
area on the observation sheet. During each observation, data items collected included
behavior mapping of routes taken by proximate pedestrian and bicycle traffic, trace
measures indicating presence of graffiti, litter and vandalism, and a written description of
the built environment surrounding each select station site, as well as activities of human
presence.
Pedestrian and bicycle routes through the vicinity of station sites were recorded
during each observation. Behavior maps of observed adjacent pedestrian and bicycle
routes were drawn over the Google Map on the observation sheets. For pedestrian routes,
low traffic routes were designated as having fewer than 10 individuals per hour, medium
traffic routes had up to 100 individuals per hour, and high traffic routes had more than
100 individuals per hour. For bicycle routes, low traffic routes had fewer than 5 bicyclists
per hour, medium traffic routes had up to 20 bicyclists per hour, and high traffic routes
had more than 20 bicyclists per hour.
63


Data analysis: Grounded visualization is a qualitative analytical technique that
allows for integrative analysis of quantitative GIS data and qualitative spatial data
(Knigge and Cope 2009). Descriptive analysis and grounded visualization of behavior
maps and other observations of the built environment collected at select station locations
were used to examine the interaction of human activity, Denver B-cycle operations,
transit and bicycling-supportive facilities. Particular attention was paid to evaluating the
quantity and type of pedestrian traffic near each select station, as well as the nature of any
activities observed in the vicinity of station sites. Findings from grounded visualization
analysis were compared with GIS data to inform and validate quantitative observations.
Aim 1.2: Investigate the integration of Denver B-cycle within the existing urban
transportation system.
Hypothesis 1.2: Denver B-cycle stations having greater integration with transportation
infrastructure, as indicated by proximity to public transportation stops, on and off-street
bicycle facilities, and other Denver B-cycle stations in the network will experience
greater numbers of checkouts.
Data collection: GIS were retrieved from repositories maintained by DenverGIS
and RTD as detailed in Table III.2. Base layers for Denver County, including layers
showing streets and highways, street names and address locations, bicycle facilities,
Denver B-cycle station locations, and light rail and bus transit stops. Additional layers
were created to show network distance to Denver B-cycle stations of 150, 300 and 500
meters using ESRI ArcGIS 9.0 software. GIS layers were assembled for analysis in ESRI
ArcGIS 9.0. The data variables B-cycle/transit use and transit use dependent on bicycle
access from the Denver B-cycle user survey dataset were also used.
64


Data analysis: Geographic network distance tools are useful to describe
interconnectivity of access to facilities in context of non-motorized transportation
(Oliver, Schuurman et al. 2007). A geographic framework can assist in elucidating
patterns within dynamic systems of urban transportation (Shaw and Xin 2003). Network
distance of the proximity of Denver B-cycle stations to other Denver B-cycle stations, to
bus and train transit stops, and to on and off street bicycle facilities was used to evaluate
the level of integration of Denver B-cycle with itself, and within the existing system of
public and non-motorized transportation infrastructure.
Using ESRI ArcGIS 9.0 software, network distances of 150 meters, 300 meters
and 500 meters were used to examine the area proximate to Denver B-cycle stations. The
network distances of 150 and 300 meters are derived as reference distances from the
stated goals and emerging practices of station placement within large and established
public bicycle sharing systems in Europe and elsewhere. The 150 meter network distance
(the maximum distance from a station in the service area given the target of 300 meter
distance between stations) is similar to the distance in several large public bicycle sharing
systems, and 300 meters is a doubling of that distance. Additionally, a network distance
of 500 meters was used as a reference distance to compensate for the lower geographic
density of Denver and the Denver B-cycle system as compared to elsewhere.
Counts were made for each Denver B-cycle station of the number of transit stops,
length of bicycle facilities, and the number of other Denver B-cycle stations within each
of the three network distance zones. An examination of the interconnectedness of Denver
B-cycle stations with network distance zones of other Denver B-cycle stations was used
to determine the level of network cohesion within a mutually supportive system.
65


Multiple regression analysis using Stata 11 software was performed to evaluate
Hypothesis 1.2. The outcome variable was total checkouts per station for the season.
Predictor variables for each station included the number of transit stops (light rail and
bus) within 150 meters network distance, presence of adjacent bicycle facilities (on or
off-street lanes, routes or paths), and other Denver B-cycle stations within 300 meter
network density (intersecting 150 meter network distance perimeters between stations).
The data variables B-cycle/transit use and transit use dependent on bicycle access
from the Denver B-cycle user survey dataset were descriptively analyzed using Microsoft
Excel.
Aim 2: Describe the characteristics of Denver B-cycle users.
Data collection: To address this Aim, data from the Denver B-cycle system usage
database were used, described in detail earlier in this chapter. Data variables for
subscription included: date of purchase and type of subscription (annual, 24-hour kiosk,
24-hour online, 7-day, 30-day), compiled weekly. Data used for evaluation of checkout
activity included weekly numbers of checkouts per station and total year-end checkouts
per station. De-identified addresses of annual members were extracted from the Denver
B-cycle system usage database. The Denver street address layer was retrieved from the
DenverGIS online repository to facilitate geocoding of the addresses of Denver B-cycle
annual members.
Data Analysis: To address Aim 2, the Denver B-cycle system usage dataset was
imported into Microsoft Excel for evaluation. Descriptive analyses of the following items
were conducted: weekly cumulative Denver B-cycle checkouts in central downtown,
number of Denver B-cycle checkouts per week from stations in central downtown, and
66


total year-end Denver B-cycle checkouts per station in central downtown.
Stations in the central downtown group were organized geographically into
subgroups of neighborhood service areas, as shown in Figure III. 3.
Figure IIE.2 Denver B-cycle station Groupings by Geographic Service Areas.
Weekly average checkouts per station delineated by neighborhood subgroup were
tabulated from the Denver B-cycle system usage dataset and descriptively analyzed using
Microsoft Excel.
Membership sales by type per week were extracted from the Denver B-cycle
system usage dataset and descriptively analyzed using Microsoft Excel.
The addresses of annual members collected during registration, and contained in
the Denver B-cycle system usage dataset were extracted using Microsoft Excel and
geocoded into a GIS map layer using ESRI ArcGIS 9.0 software. This map layer was
67


then overlaid with other GIS layers to examine the dispersal of annual users relative to
Denver B-cycle stations, and analyzed using grounded visualization.
Aim 2.1: Determine demographic characteristics of Denver B-cycle users.
Hypothesis 2.1: Denver B-cycle users will be more likely to be male, Caucasian, more
highly educated, and have a higher household income than the general population.
Data collection: The Denver B-cycle user survey dataset, described in detail
earlier in this chapter, was used to address Aim 2.1. Data variables from this dataset
included: gender, age group, home ZIP Code, ethnicity, race, educational attainment,
household income, self-assessed health status, self-reported weight, self-reported height,
self-assessed poor physical health in past 30 days, self-assessed poor mental health in
past 30 days, bicycle ownership, self-assessed bicycle skill level, active transportation
activity in previous 30 days, car ownership/access, and commute mode choice. The
variable BMI was calculated from the variables self-reported weight and self-reported
height using Stata 11 software.
Population level data were also collected from 2010 U.S. Census Denver County
demographic data for gender, ethnicity and race, 2010 U.S. Census American
Community Survey Denver County data for age group, educational attainment and
household income. Data were also collected from the 2010 Colorado BRFSS Denver
County for population weight categories, health status, days of poor physical health in the
preceding 30 days, and days of poor mental health in the preceding 30 days
Data analysis: To evaluate Hypothesis 2.1, demographic variables from the
Denver B-cycle user survey dataset were compared to 2010 U.S. Census Denver County
demographic data regarding gender, ethnicity and race, and to 2010 U.S. Census
68


American Community Survey for Denver County data regarding age group, educational
attainment, and household income using chi square tests in Microsoft Excel.
The variable BMI of each survey respondent was calculated using Stata 11
software from data variables self-reported height and self-reported weight taken from the
Denver B-cycle user survey dataset, using the formula (Centers for Disease Control and
Prevention 2011):
BMI = weight (lb)/[height (in)]2 x 703
Descriptive statistical analysis was conducted using Stata 11 software to ascertain
the mean, minimum, maximum and standard deviation of the calculated BMI scores of
Denver B-cycle annual members. The BMI scores of Denver B-cycle annual members
were organized according to normal, overweight and obese weight categories and the
distributions were compared with equivalent data from the 2010 Colorado BRFSS for the
general adult population of Denver County using chi square tests in Microsoft Excel.
Health self-assessment data of annual members taken from the Denver B-cycle
user survey dataset, including health status, days of poor physical health in the preceding
30 days, and days of poor mental health in the preceding 30 days were compared with
equivalent data from the 2010 Colorado BRFSS for the general adult population of
Denver County using chi square tests in Microsoft Excel.
The data variable active transportation behavior from the Denver B-cycle user
survey dataset was organized and descriptively analyzed using Stata 11 software.
The data variables self-assessed bicycling ability and bicycle ownership from the
Denver B-cycle user survey dataset were organized and descriptively analyzed using
Microsoft Excel.
69


Data on car and bicycle commuting behavior of Denver B-cycle users taken from
the Denver B-cycle user survey dataset were descriptively analyzed for comparison with
2010 U.S. Census American Community Survey data for Denver County, the State of
Colorado, and the United States.
Aim 2.2: Investigate factors influencing Denver B-cycle use.
Hypothesis 2.2: The number of Denver B-cycle checkouts by annual members will be
related to lifestyle factors indicating ability to replace car use with shared bike use.
Data collection: The Denver B-cycle user survey dataset was used to address this
Aim. Specifics of this dataset are described in detail earlier in this chapter. Data variables
used included number of Denver B-cycle checkouts per week, gender, age group, home
ZIP Code, household income, self-assessed health status, self-reported weight, self-
reported height, commuting behavior, shift from car use to Denver B-cycle, and bicycle
ownership.
Some data variables were derived from existing data variables in the Denver B-
cycle user survey dataset using Stata 11 software. A dichotomous variable proximity was
derived from home ZIP Code variable, indicating in-service area or not in-service area
status. Table III.4 shows ZIP Codes that contain Denver B-cycle stations, which are
classified as in the Denver B-cycle service area.
Table III.4 Denver B-cycle In-Service Area ZIP Codes.
80202 80205 80264 80293 80206
80203 80211 80265 80294 80210
80204 80218 80290 80209
A variable BMI was calculated from self-reported height and self-reported weight
70


variables, using the formula cited earlier in this chapter. A dichotomous variable
commute via Denver B-cycle was derived by extracting responses for commuting in part
or in whole by using Denver B-cycle from the commuting behavior variable. A
dichotomous variable car replacement was derived from the variable shift from car use to
Denver B-cycle.
Data analysis: To evaluate Hypothesis 2.2, ordered logistic regression analysis
using Stata 11 software was performed to develop a multivariate model to predict the
dependent variable number of Denver B-cycle checkouts per week. Independent variables
included in the model were commute via Denver B-cycle, car replacement, proximity,
bicycle ownership, gender, age group, household income, self-assessed health status, and
BMI.
A second model using multiple logistic regression analysis through was
performed to develop a multivariate model to predict the dependent variable commute via
Denver B-cycle. Independent variables included in the model were car replacement,
proximity, bicycle ownership, gender, age group, household income, self-assessed health
status, and BMI.
Prior to applying the regression models, the assumptions of observations being
independent and independent variables being linearly related to the logit were tested and
met using the omodel command in Stata 11 (UCLA: Academic Technology Services
Statistical Consulting Group 2011).
Aim 2.3: Determine impacts of Denver B-cycle on shifts toward active
transportation.
Hypothesis 2.3: Denver B-cycle annual members will shift mode choice away from car
71


use and toward active transportation via shared bicycles.
Data collection: The Denver B-cycle user survey dataset, described in detail
earlier in this chapter, was used to address Aim 2.3. The data variable used to address
Aim 2.3 was shift from car use to Denver B-cycle. Data in this variable came from a
survey question that asked participants to estimate the frequency of trips they make using
Denver B-cycle that replace car trips. Response categories for the question were Never,
Rarely, Sometimes, Most of the time, and Always. Responses to this question
contributed to a model estimating the frequency with which annual members replace car
trips with Denver B-cycle trips. Because such trips replace car use with bicycle use, any
positive response comprises a net increase in active transportation behavior regardless of
any existing active transportation behavior of individual users.
Data analysis: In evaluation of Hypothesis 2.3, a weighted car trip replacement
multiplier (Mctr) was derived from the shift from car use to Denver B-cycle variable,
using the formula:
Mctr = [(n1*0.00)+(n2*0.25)+(n3*0.50)+(n4*0.75)+(n5*1.00)]/N
The car trip replacement multiplier returned a figure estimating the average
percentage of Denver B-cycle trips made by annual members that replace car trips.
Aim 2.4: Identify active transportation benefits for Denver B-cycle annual members.
Hypothesis 2.4: Denver B-cycle annual members will exhibit a net increase in quantity of
active transportation.
Data collection: To address Aim 2.4, the Denver B-cycle system usage dataset
was used, described in detail earlier in this chapter. Data variables used included
date and time of checkout and duration of checkout for each checkout session logged for
72


each annual member for the year. In addition, GPS data for speed while in motion from
the Denver B-cycle system usage dataset was used to calculate an average estimated
speed during checkout.
Data analysis: For each annual member, totals were calculated from the data
variables date and time of checkout and duration of checkout from the Denver B-cycle
system usage dataset, to obtain total checkouts, total minutes of checkout time, total
weeks of activity (the number of weeks in which a checkout was logged) and total
estimated miles ridden.
The figure for total estimated miles ridden was derived from the average duration
of checkout time and the average estimated speed during checkout, as calculated from
GPS data generated by a test group of 10 shared bikes equipped with GPS units,
described earlier in this chapter.
The totals for all individual annual members for total checkouts, total minutes of
checkout time, total weeks of activity, and total estimated miles ridden were summed and
averaged to obtain average totals for annual members. The car trip replacement multiplier
developed in Aim 2.3 was then applied to average weekly and year-end totals to
calculate net increases in active transportation behavior.
The average weekly number of checkouts per active annual member was
charted over time. The chart was examined to identify trends of use over the course
of the season. This inquiry was to determine if active transportation became an
established behavior among annual members.
Aim 3: Provide an in-depth description of the impacts of Denver B-cycle as a public
health intervention.
73


Data collection: Twenty-two select participants were recruited, of which 9 were
female and 13 were male. All select participants were users of Denver B-cycle, at
differing levels of involvement. Eight were occasional users (<1 checkout per week) and
14 were regular users of the system (>2 checkouts per week). The select participants were
split evenly by locality status, between in service area and out of service area ZIP Code
of home address, as listed in Table III.4. Of the group, two were of Hispanic ethnicity.
Education and household income levels of select participants encompassed the spectrum
of survey respondents. The gender, age, occupation, locality status, and level of use for
each select participant are shown in Table III. 5.
74


Table III.5 Characteristics of Select Participants
Gender Age Occupation In service area Out of service area Occasiona l User (<1 checkouts per week) Regular user (>2 checkouts per week)
Male 30 financial analyst X X
Female 34 senior metrics consultant X X
Female 43 medical researcher, graduate student X X
Male 35 environmental planner X X
Male 67 retired mechanical engineer X X
Male 39 software developer, graduate student X X
Female 21 student X X
Female 62 lawyer X X
Female 23 communications assistant X X
Male 40 systems analyst X X
Male 44 mechanic X X
Male 52 chief financial officer X X
Female 27 marketing manager X X
Male 41 small business owner X X
Male 44 software engineer X X
Female 24 accountant X X
Male 46 office manager X X
Female 33 high school teacher X X
Male 48 telecommunications analyst X X
Male 33 software engineer X X
Male 58 IT manager X X
Female 33 economic development consultant X X
75


Semi-structured in-depth interviews of select participants were conducted
between December 2010 and May 2011. Interviews typically lasted between 30 and 60
minutes. Select participants were given gift cards of $25, either from REI or Whole
Foods in appreciation of their time, after the completion of the interview. The question
guide for the interviews appears in Appendix C. Interview questions were developed to
probe select participants regarding opinions and use of Denver B-cycle, physical activity
behavior, and bicycling behavior, bicycling comfort and perceptions of safety. Questions
were ordered into the following groups:
Introduction: Brief description of self, proximity of closest Denver B-cycle
station
Denver B-cycle initial experience: Reasons for first using Denver B-cycle, and
initial impressions.
Activities integrated with Denver B-cycle: The purposes for which the select
participant uses Denver B-cycle, destinations visited, self-assessment as to
whether Denver B-cycle fits their lifestyle.
Bicycling behavior: How bicycle riding habits have changed while or since
using B-cycle, how general bicycle riding habits have changed, and whether
select participants have influenced the behavioral actions of other
individuals.
Physical activity/active transportation: Behavior regarding physical activity,
opinions regarding the use of bicycles for transportation, opinions regarding
the social status and acceptability of bicycling for transportation.
76


Perceived quality of life/benefits: Opinions regarding how Denver B-cycle
may affect quality of life, and health and economic effects attributable to
Denver B-cycle.
Effects on other modes: Description of how use of Denver B-cycle may have
affected the use of other modes of transportation.
Perceived bicycling comfort and safety: The Portland Bureau of
Transportation in Portland, Oregon developed a categorization system for
transportation bicyclists. They determined four levels, in descending order of
comfort while riding in traffic, including the strong and fearless, the enthused
and confident, the interested but concerned, and the no way no how groups
(Geller 2009). Detailed descriptions of each group appear in Appendix C, and
were presented to participants during interviews. They were then asked to
identify in which category they felt most comfortable. Interviewees were also
asked to identify concerns about safety they have experienced while using
Denver B-cycle.
Impressions of Denver B-cycle: Open ended opinions regarding the function
and usefulness of Denver B-cycle.
Data analysis: Digital recordings of interviews were transcribed into Microsoft
Word. Qualitative Causal Analysis (QCA) is an analytical method used to identify and
explore causal conditions and pathways using qualitative data (Ragin 1999). Within the
framework of QCA, matrices of response categories are developed from coded
interviews, into which groupings of responses with similar context or logic are organized
77


into fuzzy sets (Rihoux 2006). Key words, concepts, and themes were identified and
inductively coded into categories organized by thematic question groups, outlined in the
preceding. Repeated words, phases and themes, situational context, and commonalities
among respondents were noted, an interpretive framing guided by a grounded theory
approach (Bernard 2011). A summary table of code definitions appears in Appendix D.
The findings of this analysis were compared among and between subgroups of select
participants. The coded data matrices were then used to develop deterministic functional
models, an analytic method of QCA to describe data and to construct groupings of
generalized findings (Rohwer 2011).
Following QCA, generalized findings were examined using analytic induction.
Analytic induction is a method of examination of social phenomena through the
development of hypothetical explanations, which are inductively revised and refined in
an effort to achieve practical certainty of a hypothesis to fit observations (Robinson
1951). In interpretation of the qualitative findings, quantitative findings were revisited to
inform and hone the development of hypotheses derived through analytic induction.
Quantitative and qualitative findings, when used to cross validate each other, allow for
greater breadth and depth of understanding (Neuman 2003). Once qualitative analyses
shaped hypotheses with sufficient practical certainty, these findings were used to form an
interpretive lens guided by a grounded theory approach, through which quantitative
findings were examined.
78


CHAPTER IV
RESULTS
Introduction
The results of this research are organized by study aim. The specific aim
addressed is presented at the beginning of each part, followed by associated hypotheses
and analyses. Summaries of findings are offered at the end of each segment.
Aim 1: Examine the physical characteristics and human activity in the built
environment in which Denver B-cycle stations are situated.
Aim 1.1: Identify features that distinguish low and high performing sites.
Characteristics Of The Built Environment At Select Stations
As a public health intervention applied through infrastructure, it is important to
understand how Denver B-cycle has been incorporated into the space and activities of the
surrounding environment. In this intervention, individuals self-select to participate, and
determine on their own the frequency and continuity of participation. Examination of the
encapsulating space around stations can reveal if, or how, the intervention is being
accepted and incorporated.
Ten stations of the 40 in central downtown were selected for observation, shown
in Figure IV. 1. The five most and five least used stations, as ranked following the first
three months of operation, were designated select stations. Select stations observed in the
study represent one quarter of the 40 Denver B-cycle stations of central downtown.
79


Figure IV. 1 Denver B-cycle Select Station Locations.
Features of the built environment act in conjunction to attract human presence to
the station sites in varying degrees. A foundation of literature has identified the strong
influence of the built environment on physically active behavior (Flandy, Boarnet et al.
2002; Lopez-Zetina, Lee et al. 2006; Lee, Ewing et al. 2009), but using the built
environment to proactively influence behavior through operation of a PBSS is largely
unexplored. Observations of select stations revealed characteristics of the built
environment, as well as activities at or near the sites of select stations.
A description and summary of observations of each select station appears below.
The five high-use select stations appear first, followed by the five low-use select stations.
A summary of findings follows.
80


Market Street Station
Description of the built environment: The Denver B-cycle station at Market Street
Station is located adjacent to Market Street and the 16th Street Mall, to the East of the
southern entrance of the bus station. Market Street Station is a major regional bus station
in central downtown Denver. In addition to the bus station, several buses stop along the
East side of Market Street, across from the station location. Market Street is one-way
northbound, with two travel lanes of nearly constant car traffic and a large amount of bus
traffic. The Denver B-cycle station is readily visible from the street and nearby pedestrian
areas.
The 16th Street Mall is a major center of activity in the downtown area, including
employment, entertainment, retail and commerce. It is a pedestrian/transit mall, served by
free shuttle buses in both directions at frequent intervals. A stop for the 16th Street Mall
shuttle bus is West of the station site.
Several sidewalks converge in the entrance area to the bus terminal to the West of
the station site. Many employment centers, hotels, restaurants and retail shops are within
view of the station site. A designated parking space for a car-sharing vehicle is located a
few meters from the Denver B-cycle station along Market Street. Bicycle riding and
parking are illegal on the 16th Street Mall. Several other Denver B-cycle stations are
located within a few blocks of the station site.
Trace measures: During site observations, small scraps of paper and other debris
were sometimes seen in the vicinity of the station site. During at least one observation, a
maintenance person was seen sweeping and removing litter from the area. No evidence of
graffiti or vandalism at the station was detected during any of the observations.
81


Behavior mapping of pedestrian and bicycle activity: A tremendous amount of
pedestrian activity was continually present at the Market Street Station site, regardless of
time of day. During morning and evening commute times, pedestrian activity was very
high, with hundreds of pedestrians alighting from or entering buses every few minutes.
Foot traffic crisscrossed every available space on sidewalks and much of the adjacent
street, as pedestrians entered and exited the bus terminal, or walked along the 16th Street
Mall or Market Street. Midday and lunchtime activity in the area was also high, if
somewhat lighter than during commute times. Street furniture and public shaded areas
attracted numerous people during each observation. Figure IV.2 shows a map of
pedestrian and bicycle tracks observed at the station site.
Bicycle activity in the area was fairly active, with between 18 and 44 bicycles
observed being ridden during each hour-long observation at this site. Nearly all the
available bicycle racks for privately owned bikes were occupied during all observation
periods. Other privately owned bicycles were parked on signposts, streetlamps and trees
in the area around the station site.
82


Figure IV.2 Behavior Mapping: Market Street Station.
Although illegal on the 16th Street Mall, bicycle activity on the Mall was not
uncommon. There were no on-street bicycle facilities along Market Street. Many
bicyclists rode northbound in the travel lanes of Market Street while a lower quantity
rode southbound along the sidewalks.
Some of the bicycle traffic in the area was due to people entering and exiting the
bus terminal with bicycles. Bicycle activity into and out of the bus terminal was most
prevalent during commute times. During each observation period, a few people were seen
to transition between bus and Denver B-cycle.
83


16th and Little Raven
Description of the built environment: The Denver B-cycle station at 16th and
Little Raven is located on the East side of Commons Park, near the intersection of Little
Raven Street and 16th Street. The station site is close to several high-rise residential
buildings with ground level restaurants, retail and commerce, and is readily visible from
the street and surrounding area. The area is West of the Millennium Bridge, which
connects the area to Union Station and central downtown across several freight railroad
tracks. Little Raven Street has a light but frequent car traffic presence.
To the West of Commons Park are the Platte River and numerous mixed-use
commercial and residential buildings. Although the land use in the area surrounding the
station site is mixed, many of the buildings in the area are primarily dedicated to
residential space. Many of the units appear to be high-end apartments and condominiums.
A major pathway through the park is immediately adjacent to the station site.
Several other pathways and sidewalks converge near the station site, and a major
pedestrian crossing for Little Raven Street is a few meters to the East.
Trace measures: During each site observation, litter was observed at the station
site, ranging from small bits of paper and cigarette butts to food wrappers and drink
containers. Some of the bicycles at the station occasionally had advertisements for
entertainment events taped to their baskets. The map board of the station had been tagged
with pink paint prior to the August observation, but had been removed by the following
observation. The pink tag later reappeared on the stations map.
Behavior mapping of pedestrian and bicycle activity: Pedestrian activity near the
84


station site remained light but consistent during each of the observation periods, with the
majority of foot traffic passing close to the station site. There were many runners and dog
walkers on the paths and sidewalks near the station site. The street crossings near the
station site maintained a nearly constant pedestrian presence. Figure IV.3 shows a map of
pedestrian and bicycle tracks mapped at the station site.
At an observation during morning commute time, a few pedestrians who were
apparent annual members in possession of a Denver B-cycle B-card were seen to check
out bikes from the station then ride in the direction of central downtown. Many of these
riders wore business attire and placed handbags or briefcases into bike baskets before
leaving. Each appeared to be quite familiar with the operation of the docks and the
bicycles.
85


Full Text

PAGE 1

PUBLIC BICYCLE SHARING AS A POPULATION SCALE HEALTH INTERVENTION FOR ACTIVE TRANSPORTATION IN DENVER, COLORADO by Andrew L. Duvall B.A., University of Wyoming, Laramie, 1994 A thesis submitted to the Faculty of the Graduate School of the University of Colorado Denver in partial fulfillment of the requirements for the degree of Doctor of Philosophy Health and Behavioral Sciences 2012

PAGE 2

ii This thesis for the Doctor of Philosophy degree by Andrew L. Duvall has been approved for the Health and Behavioral Sciences by Debbi Main, Chair and Advisor John Brett Wesley Marshall Eric France Date: April 6, 2012

PAGE 3

iii Duvall, Andrew L. (Ph.D., Health and Behavioral Sciences) Public Bicycle Sharing as a Population Scale Health Intervention for Active Transportation in Denver, Colorado Thesis directed by Professor Debbi Main ABSTRACT Inadequate physical activity associated with a cute car dependence is linked to increased risk of obesity and related chronic health conditions, including heart disease, diabetes, hypertension, depression and cancer. More than half of U.S. adults do not meet the minimum recommended levels of physical activity. This research is a mixed methods quantitative/qualitative st udy of the impact of a public bicycle sharing system as a health intervention to induce active transportation behavior. The primary quantitative outcome was a net increase in active transportation through shared bike use. Qualitative outcomes included inve stigation of the effects of the intervention on the behaviors of individual users, sources of motivation to participate and continue use, indicators and pathways of diffusion of the intervention, and broader impacts of the intervention. Denver B cycle annu al members logged an average 60.3 minutes of weekly checkout time, of which an estimated 35.5% to 50.0% replaced car trips. Annual members differed significantly from the general population, being more likely to be male, non Hispanic, Caucasian, aged 25 an d 44, more educated !"#$%&"&$'&()"$*+,-(!" higher self reported health status, and tended to be of normal weight. M ultivariate model s of influencing factors found two key variables associated with the number of checkouts among annual members Commuting via shared bikes and

PAGE 4

iv the ability to replace car use increased the odds of higher checkouts. Proximity of residence to Denver B cycle stations was not a significant predictor. Women were as likely as men to commute via shared bikes. The use of Denver B cycle in creased net active transportation among annual members. Users discovered their own meaningful ways in which the use of shared bicycle best served their needs. Some participants best able to integrate Denver B cycle into their lifestyles reported weight los s, increased fitness, reduced car dependence, and economic benefits. The central implication of this research is that public bicycle sharing can be effectively applied to increase active transportation behavior. The findings are consequential for two tren ds: shifts away from car use toward use of shared bikes, and increases in overall active transportation. The form and content of this abstract are approved. I recommend its publication. Approved: Debbi Main

PAGE 5

v DEDICATION I dedicate this work to my wife Jul ie, my daughters Stella and Piper, and to my family and friends. Their support was the wind at my back, and their love helped my wheels to spin.

PAGE 6

vi ACKNOWLEDGMENTS The author is greatly appreciative of the U.S. National Science Foundation Integrative Graduate Education and Resea rch Traineeship (IGERT) program, "#$%& '()*+!,-.!/#$ 0 1234567 ! The Association of Schools of Public Health / Centers for Disease Control and Prevention Environmental Health Commu nity Design scholarship program, and the Community Benefit Initiatives Committee of the 8)9:;*!<;*=)>;>?;!@-A-*)+-!">:?9?B?;! C-*!D;)A?E!%;:;)*FE!G*-H9+;+!I;>;*-B:!F->?9>B;+!:BGG-*?. Scholarship and camaraderie fostered by the faculty, staff and students of the University of Colorado Denver Department of Health and Behavioral Sciences, and the Center for Sustainable Infrastructure Systems at the University of Colorado Denver, directed by Dr. Anu Ramaswami, has been of substantial and ongoing value. The autho r wishes to express deep gratitude for the opportunities and support afforded by d isserta tion committee chair Dr. Debbi Main, and committee members Dr. John Brett, Dr. Wesley Marshall and Dr. Eric France These talented and committed educators contributed insight, encouragement, direction and advice, which have greatly enhanced the quality of this work This research would not have been possible without a great many people. The author wishes to thank those who directly affected the advancement of this work Organizations : The City and County of Denver, Denver Bike Sharing, BikeDenver, the Denver Mayor's Bicycle Advisory Committee, GreenPrint Denver, Bikes Belong, and B cycle. Individuals : Parry Burnap, Steve Sander, Emily Snyder, Piep van Heuven, Tracy Halasinski, Dick Gannon, Karen Good, Nick Bohnenkamp, Cindy Bosco, Ken Gart, Corina Lindley, Nate Kvamme, Tim Blumenthal, Bob Burns, Governor John Hickenlooper, Mayor Guillermo Vidal, and the volunteers, employees and board of Denver Bike Sharing. Speci al thanks goes to everyone who rides or aspires to ride a bicycle, for recreation, for transportation, for the betterment of themselves and their community, and especially for enjoyment.

PAGE 7

vii TABLE OF CONTENTS CHAPTER 1. INTRODUCTION ................................ ................................ ................................ .......... 1 Background ................................ ................................ ................................ ............. 1 Significance Of The Study ................................ ................................ ...................... 2 Research Question and Specific Aims ................................ ................................ .... 7 Overview of Methods ................................ ................................ ............................. 8 2. LITERATURE REVIEW ................................ ................................ ............................. 10 Development Of The Current Built Environment ................................ ................ 10 The Emergence Of Car Dependence ................................ ............................... 10 Decoupling Transportation From Physical Activity ................................ ....... 11 The State Of Obesity And Why It Matters ................................ ..................... 12 Obesity Risk Through The Urban Form ................................ ......................... 14 Beyond The Car: Public Bike Sharing ................................ ............................ 16 Guiding And Applied Theories For This Research ................................ .............. 19 Informing the Study ................................ ................................ .............................. 21 Intervention Through The Built Environment ................................ ................ 22 Intervention And The General Populations ................................ .................... 26 The effects of environmental context ................................ ............................. 32 Public Bicycle Sharing And Active Transportation ................................ ........ 37 Public Bicycle Sharing Participation ................................ .............................. 41 Motivation To Use Public Bicycle Sharing ................................ .................... 44 Summary Of Literatu re Review ................................ ................................ ............ 45 3. METHODS OF DATA COLLECTION AND EVALUATION ................................ .. 48 Study Design ................................ ................................ ................................ ......... 48

PAGE 8

viii Brief Glossary Of Terms Specific To This Research ................................ ........... 49 Study Sett ing ................................ ................................ ................................ ......... 49 Data Sources ................................ ................................ ................................ ......... 51 Data Variables ................................ ................................ ................................ ....... 53 Geographic Data Overview ................................ ................................ ............. 56 Denver B Cycle System Usage Dataset Overview ................................ ......... 56 Denver B Cycle User Survey Dataset Overview ................................ ............ 57 Select Participants ................................ ................................ ................................ 61 Human Subjects Review ................................ ................................ ....................... 61 Methods of Data Collection and Evaluation, by Specific Aim ............................. 61 4. RESULTS ................................ ................................ ................................ ..................... 79 Introduction ................................ ................................ ................................ ........... 79 Aim 1 : Examine the physical characteristics and human activity in the built environment in which Denver B cycle stations are situated ................................ 79 Aim 1.1 : Identify features that distinguish low and high performing sites. ......... 79 Characteristics Of The Built Environment At Select Stations ........................ 79 Market Street Station ................................ ................................ ...................... 81 16 th and Little Raven ................................ ................................ ....................... 84 REI ................................ ................................ ................................ .................. 87 19 th and Pearl Street ................................ ................................ ........................ 91 Union Station ................................ ................................ ................................ .. 93 Denver Heal th ................................ ................................ ................................ 96 Five Points ................................ ................................ ................................ ...... 99 15 th and Tremont ................................ ................................ ........................... 102 25 th and Lawrence ................................ ................................ ......................... 105 Pepsi Center ................................ ................................ ................................ .. 108

PAGE 9

ix Analysis of Characteristics of the Built Environment at Select Stations ...... 111 Aim 1.2 : Investigate the integration of Denver B cycle within the existing urban transportation system ................................ ................................ ......................... 115 Station Network Density of Denver B cycle ................................ ................ 116 Proximity To Transit ................................ ................................ ..................... 117 Proximity To Bicycle Facilities ................................ ................................ .... 119 Summary Of Denver B Cycle Wit hin The Urban Transportation System ... 121 Aim 2: Describe the characteristics of Denver B cycle users. ............................ 123 Description Of System Wide Patterns Of Use ................................ .............. 123 Checkout Activity ................................ ................................ ......................... 123 Membership Characteristics ................................ ................................ .......... 128 Aim 2.1: Determine demographic characteristics of Denver B cycle users. ...... 130 Demographic data summary ................................ ................................ ......... 130 Evaluating Gender, Ethnicity and Race ................................ ........................ 134 Evaluating Age Group, Educational Attainment and Household Income .... 135 Health Related Survey Data ................................ ................................ ......... 137 Summary: Denver B cycle U sers Versus the General Population ................ 146 Aim 2.2: Investigate factors influencing Denver B cycle use. ........................... 148 Multivariate Modeling Of Number Of Checkouts By Denver B Cycle Annual Members ................................ ................................ ................................ ....... 149 Commute Via Denver B Cycle ................................ ................................ ..... 153 Summary Of Factors Influencing Use ................................ .......................... 156 Aim 2.3: Determine impacts of Denver B cycle on shifts toward active transportation. ................................ ................................ ................................ ..... 158 Summary Of The Impact Of Denver B Cycle On Active Transportation .... 160 Aim 2.4: Identify active transportation benefits for Denver B cycle annual members. ................................ ................................ ................................ ............. 161

PAGE 10

x Physical Activity Benefits ................................ ................................ ............. 161 Summary Of Active Transportation Benefits For Denver B Cycle Annual Members ................................ ................................ ................................ ....... 167 Aim 3: Pr ovide an in depth description of the impacts of Denver B cycle as a public health intervention ................................ ................................ .................. 168 Motivation for Active Transportation Behavior ................................ ........... 168 Health outcomes ................................ ................................ ............................ 169 Economic Outcomes ................................ ................................ ..................... 174 Factors For Incorporating Behavior Change Into Lifestyle .......................... 176 Diffusion of Use within Social Networks ................................ ..................... 181 Social change ................................ ................................ ................................ 187 5. DISCUSSION AND LIMITATIONS ................................ ................................ ........ 195 Discussion ................................ ................................ ................................ ........... 195 Interpreting The Findings ................................ ................................ ............. 195 Denver B Cycle As A Lifestyle Intervention ................................ ............... 196 "Stealth Intervention ................................ ................................ ................... 200 Population Reached Through The Intervention ................................ ............ 204 Broader Impacts on Bicycling Activity In Denver ................................ ....... 206 Long Term Outcomes ................................ ................................ ................... 210 Ongoing Intervention ................................ ................................ .................... 211 Informing Application: Policy And Infrastructure ................................ ....... 214 Theoretical Implications ................................ ................................ ............... 218 Opportunities For Future Research ................................ ............................... 224 Study Limitations ................................ ................................ ................................ 228 Conclusions ................................ ................................ ................................ ......... 230

PAGE 11

xi REFERENCES ................................ ................................ ................................ ............... 232 APPENDIX ................................ ................................ ................................ ..................... 244 "#! Denver B cycle Select' Station Observation Sheet ................................ ....... 245 B. Denver B cycle User Survey Instrument ................................ ....................... 246 C. Semi structured Interview Question Guide for Denver B cycle Select' Participants ................................ ................................ ................................ .......... 256 D. Code Definitions ................................ ................................ ............................ 259

PAGE 12

xii LIST OF TABLES Table II.1 Obesity: Chronic Health Conditions and Risks. ................................ ......................... 14 III.1 Data Sources Associated with Study Aims And Hypotheses. ................................ .. 51 III.2 Data Variables Used in this Research, Arranged by Source. ................................ .... 53 III.3 Denver B cycle Select Stations. ................................ ................................ ................ 62 III.4 Denver B cycle In Service Area ZIP Codes. ................................ ............................ 70 III.5 Characteristics of Select Participants ................................ ................................ ........ 75 IV.1 Characteristics of High and Low Use Denver B cycle Select Stations. ................. 111 IV.2 Means, Standard Deviations and Intercorrelations for Total Checkouts and Predictor Variables (N=32). ................................ ................................ ................................ ........... 120 IV.3 Simultaneous Multiple Regression Summary for Transit Stops, Bicycle Facilities, and Network Density Predicting Total Checkouts (N=32). ................................ ............ 121 IV.4 2010 Denver B cycle Checkouts Per Station in Central Downtown. ..................... 126 IV.5 Demography and Characteristics of Denver B cycle Users. ................................ .. 132 IV.6 Gender, Ethnicity and Race of Denver B cycle Annual Members and Denver County Population (U.S. Census 2010). ................................ ................................ ......... 134 IV.7 Age Group, Educational Attainment and Household Income of Denver B cycle Annual Members and Denve r County Population (2010 U.S. Census American Community Survey 1 Year Estimates). ................................ ................................ .......... 136 IV.8 Health Characteristics of Denver B cycle Annual Members Versus the Denver County Adult Population. ................................ ................................ ............................... 138 IV. 9 Calculated BMI Figures from Self Reported Height and Weight of Denver B cycle Annual Members. ................................ ................................ ................................ ............ 140 IV.10 Percent within Weight Class Groups as Determined by BMI, Denver B cycle Annual Members Versus Denver County Population. ................................ .................... 140 IV.11 Active Transportation Behavior of Denver B cycle Users. ................................ .. 142

PAGE 13

xiii IV.12 Ordered Logistic Regression Predicting Number of Denver B cycle Checkouts by Annual Members. ................................ ................................ ................................ ............ 150 IV.13 Multiple Logistic Regression Predicting Commute Via Denver B cycle By Annual Members. ................................ ................................ ................................ ........................ 154 IV.13 Year end Average Totals and Estimates of Denver B cycle Use Per Annual Member. ................................ ................................ ................................ .......................... 162 IV.14 Averag e Frequencies of Use Per Denver B cycle Annual Member Per Week of Use. ................................ ................................ ................................ ................................ 162 IV.15 Average Net Year end Indicators of Active Transportation Behavior Per Annual Member Attributo Denver B cycle Use. ................................ ................................ ......... 164 IV.16 Average Indicators of Active Transportation Behavior Per Annual Member Per Week of Use Attributo Denver B cycle Use. ................................ ................................ .. 164 V.1 Denver Commuter Population and Estimated Numbers of Bicycle Commuters, 2009 versus 2010. ................................ ................................ ................................ .................... 208 V.2 Estimated Effects of Denver B cycle Commuters on the Increase in Bicycle Commuting in Denver, 2009 to 2010 ................................ ................................ ............. 208 D.1 Code Definitions ................................ ................................ ................................ ...... 259

PAGE 14

xiv LIST OF FIGURES Figure II.1 U.S. Cities Larger than 400,000, Ranked by Bicycle Commuting. ........................... 34 III.1 Map of Denver B cycle Stations in the Central Downtown Area, within the City and County of Denver (2010). ................................ ................................ ................................ 50 III.2 Denver B cycle station Groupings by Geographic Service Areas. ........................... 67 IV.1 Denver B cycle Select Station Locations. ................................ ................................ 80 IV.2 Behavior Mapping: Market Street Station. ................................ ............................... 83 IV.3 Behavior Mapping: 16 th and Little Raven. ................................ ............................... 86 IV.4 Behavior Mapping: REI. ................................ ................................ ........................... 90 IV.5 Behavior Mapping: 19 th and Pearl Street. ................................ ................................ 93 IV.6 Behavior Mapping: Union Station. ................................ ................................ ........... 95 IV.7 Behavior Mapping: Denver Health. ................................ ................................ .......... 97 IV.8 Behavior Mapping: Five Points. ................................ ................................ ............. 101 IV.9 Behavior Mapping: 15 th and Tremont Street. ................................ ......................... 104 IV.10 Behavior Mapping: 25 th and Lawrence Street. ................................ ..................... 107 IV.11 Behavior Mapping: Pepsi Center. ................................ ................................ ......... 110 IV.12 Network Distances to Denver B cycle Stations. ................................ ................... 116 IV.13 Location o f Denver B cycle Stations Relative to Transit Stops and Stations in Central Downtown. ................................ ................................ ................................ ......... 118 IV.14 Denver On and Off Street Bicycle Facilities. ................................ ....................... 119 IV.15 Cumulative Denver B cycle Checkouts in Central Downtown, 2010. ................. 124 IV.16 Number of Denver B cycle Checkouts Per Week from Stations in Central Downtown. ................................ ................................ ................................ ...................... 125 IV.17 Weekly Average Checkouts Per Station by Neighborhood Service Areas. ......... 127

PAGE 15

xv IV.18 Activated Denver B cycle Memberships by Type, Per Week. ............................. 128 IV.19 Geocoded Addresses of Denver B cycle Annual Members in Central Downtown Denver. ................................ ................................ ................................ ............................ 129 IV.20 Self Assessed Bicycling Ability of Denver B cycle Users. ................................ 143 IV.21 Bicycle Ownership among Denver B cycle Users. ................................ .............. 144 IV.22 Car and Bicycle Commuting Mode Share: Denver B cycle Users, Denver County, Colorado and the United States (U.S Census Bureau 2011). ................................ ......... 145 IV.23 Annual Member Responses for Frequency of Denver B cycle Trips that Replace Car Trips. ................................ ................................ ................................ ........................ 159 IV.24 Percentage of Recommended Weekly Physical Activity Met by Average Wee kly Total Denver B cycle Minutes of Checkout: 90 Minute Recommendation Versus 150 Minute Recommendations. ................................ ................................ ............................. 163 IV.25 Percentage of Recommended Weekly Physical Activity Met by Average Weekly Minutes of Denver B cycle Checkout Tim e Replacing Car Trips: 90 Minute Recommendation Versus 150 Minute Recommendations. ................................ ............. 165 IV.26 Average Weekly Number of Checkouts Per Active Denver B cycle Annual Member. ................................ ................................ ................................ .......................... 166 IV.27 "#$%#&!'(#!)%*+,!-# +.#/!0 1 232$#!'(#/(!456$#!*+!7/6.8&#!069#(!:#/(;(!456$#! *+!-#+.#/!0 1 232$#!069#(< ................................ ................................ ............................... 191 IV.1 Denver Bicycle Commuter Mode Share, 2005 to 2010. ................................ ......... 207

PAGE 16

1 CHAPTER I INTRODUCTION Background A cute dependence on automobiles for transportation contributes to lack of physical activity; a condition identified with increased health risks. A well supported body of literature has identified links between high reliance on cars for transportation and increased risk of obesity and obesity related chronic conditions, w hich are associated with inadequate levels of physical activity (Handy, Boarnet et al. 2002; Frank, Andresen et al. 2004; Ewing, Brownson et al. 2006; Jones, Rutt et al. 2006; Lopez Zetina, Lee et al. 2006; Jacobson, Ki ng et al. 2011) The intensity of car use and air pollution also affects community livability and quality of life due to environmental degradation, pollution, and reduced access to land, rendering areas inhospitable to outdoor activities (MacKerron and Mourato 2009) !"#$%&#%&'($%#)*"+"#',-#*./'0$ "#'$$)*12$'1(&#$3#%*4%#0')#-&/&,-&,0&03.&# *,0)&'"*,4(5#+,36,7# strategies to mitigate and reduce the quantity of car use have become increasingly important. In recent years, some cities have introduced public bicycle sharing as an infrastructural tool to reduce car dependence while increasing opportunities for physical activity. Public bicycle sharing is a form of public transportation in which a fleet of bicycles within a network of automated stations is made available to users for short trips, for a min imal fee # Public bicycle sharing projects have been implemented in cities throughout the world primarily in countries in Europe and Asia already supportive of bicycling for

PAGE 17

2 transportation. However, in the past few years, public bicycle sharing systems ha ve begun to appear more broadly, including several car dependent cities in the United States (Shahe en and Guzman 2011) Still in its infancy, literature specific to public bicycle sharing is quite limited presently, with relatively little known about its associated health, social, behavioral, environmental and infrastructural impacts. The findings of this dissertation research are intended to inform public health policy through examin ation of the operation of a public bicycle sharing system and its effects on the user population and the city in which it operates. This dissertation research focuses on Denver B cycle, a public bicycle sharing system in Denver, Colorado, which opened to the public on April 22, 2010 Denver B cycle was the first large scale public bicycle sharing system in the United States, primarily designed to target the behavior of pe ople who live and/or work in downtown Denver. During its initial season, spanning from April to December 2010, Denver B Cycle was comprised of 500 bicycles available from 50 stations distributed predominantly in the central downtown business district of th e city, with two outlying smaller groupings of stations in the Cherry Creek and Denver University neighborhoods. This dissertation research focuses on the initial 2010 Denver B cycle season of operations. Significance Of The Study Inadequate physical activ ity is a modifiable risk factor contributing to the current problem of obesity in the United States. The costs associated with treating obesity related diseases are expected to rise dramatically, with t otal health car e costs attributable to

PAGE 18

3 obesity and overweight doubling every decade accounting for 16 18% of total US health care costs by 2030 (Wang, Beydoun et al. 2008) Health care expenditures for obese people are 42% more than for people of healthy weight, with obesity related medical costs accounting for $147 billion in U.S. annual medical costs (Levi, Vinter et al. 2010) Unfortunately, more than half of U.S. adults do not meet the minimum recommended levels of physical activity (Centers for Disease Control and Prevention 2005) even though an increase in regular physical activity can significantly reduce their relative risk for related chronic illnesses (Warburton, Nicol et al. 2006; Haskell, Lee et al. 2007) Even modest i mprovements to physical activity can reduce morbidity and mortality (Wen, Wai et al. 2011) with the greatest benefit to individuals who shift from no activity to low levels of physical activity (Woodcock, Franco et al. 2010) Reduction of preventable chronic conditions and associated costs is of paramount importance if the health care system is to sustain any ability to meet projected needs. One study predicts that b y 2 015, 75% of adu lts in the U.S. will be overweight or obese (Wang and Beydoun 2007) Faced with these staggering health and economic figures, broadly applied and effective interventions to increase physical activity and reduce incid ence of obesity in populations are critically needed to avoid future catastrophic health and economic outcomes. Obesity affects everyone, as a majority of the population is either presently obese or at elevated risk of obesity, and the entirety of the popu lation bears the associated costs. A key barrier to reducing the health and economic costs of obesity is the high dependence on automobiles for transportation, a norm strongly established in much of the

PAGE 19

4 United States. In 2009, 40% of all trips in the U.S. were shorter than two miles, though cars were the dominant mode of choice for these short distances (Alliance for Biking & Walking 2012) During the rapid spread of urban areas in the post World War II period, car use has risen steadily, con current with a decrease in bicycling and walking trips (Ogden and Carroll 2010) Over the same period of time, per capita vehicle miles traveled has also climbed, also attributed as a factor influencing risk of obesity (Jacobson, King et al. 2011) Time spent in a car is time that cannot be spent being physically active. One study of the effects of car d ependence on obesity found a 6% increase in obesity risk for each hour spent in a car daily (Frank, Andresen et al. 2004) As a result, replacing behaviors of car dependence with physically active transportation has become a more recent focus of hea lth promotion efforts. The impact of the physical form of the built environment on behavior has been widely studied, perhaps most notably in how the types and form of infrastructure encourages car use at the expense of physically active transportation (Handy, Boarnet et al. 2002; Frank, Saelens et al. 2007; Pendola and Gen 2007) Socioeconomic factors of the built environment, such as access to opportunities for more healthy choices are also important, as populations in lower resource communities are at increased risk of obesity (Babey, Hastert et al. 2009; Sallis, Saelens et al. 2009; Beech, Fitzgibbon et al. 2011) Although research points to how factors of the built environment contribute to behaviors associa ted with risk of obesity, it also suggests the potential for modifying features of the built environment to induce health promoting behaviors. Elements of the built environment have been intentionally designed to target physical activity behaviors at a li mited scale in community settings, such as the

PAGE 20

5 development of walking routes in a community (Krieger, Rabkin et al. 2009) Behavioral health interventi ons applied through infrastructure to affect active transportation at a large population scale such as bike sharing are not common, and have not been well studied. The implementation of public bicycle sharing is an opportunity to evaluate the viability of a population scale active transportation intervention. A public bicycle sharing system by design introduces a readily accessible alternative to car use for short trips, while simultaneously introducing an opportunity for low to moderate levels of physical activity to users. Behavioral health interventions for physical activity have largely been conducted with the assumption that information access will lead to motivation and action. Such minimal intervention approaches have not proven successful (Keyserling, Hodge et al. 2008) Some smaller scale interventions use en hanced approaches, including individual level strategies involving high intensity tailored coaching and/or peer counseling to realize impact (Dutton, Provost et al. 2008) Such enhanced approaches, however, are too logistically challenging and costly as a population level solution As an intervention, p ublic bicycle sharing is a broad based approach to reach a large population The designed intent is for use of the system to fit the lifestyles of users, who develop their own motivational reasoning to become participants. The users themselves are important operational elements of the intervention, self selecting for participation, self moderating continued activity, and playing roles as recruiters and transmit ters of information to engage additional participants. The design and outlay of the public bicycle sharing system is intended to facilitate development of individual motivation by meeting basic transportation needs. Through

PAGE 21

6 supportive placement of public bicycle sharing stations near places of residence and desirable destinations, a wide variety of urban trips for which a car might have previously been used can be replaced with shared bicycle trips. Additionally, with access to public bicycle sharing syste m bicycles, transit use, which has been identified to encourage active transportation (Besser and Dannenberg 2005) might be made more attractive and accessible to people who do not currently use transit due to inconvenience. Havin g the option for quicker trip completion than walking without the need to use a car positions public bicycle sharing as an attractive transportation mode. Locations too distant from transit stops to access by walking, but within the public bicycle sharing system service area are prime for shared bike/transit trips. In summary, this r esearch begin s to fill substantial gaps in current knowledge regarding population scale interventions to increase physical activity. As public bicycle sharing system is deploye d in more cities, this information is of mounting importance to provide public health professionals and policy makers with information for evidence based health policy and community design decisions Obesity is in part a consequence of high car dependence and as such is a modifiable risk. Given that obesity and associated illnesses are a looming threat to an increasing majority of the population, and that costs to treat obesity related illnesses are projected to rise dramatically in the coming decades (Wang, Beydoun et al. 2008) it is a population scale problem that may best be addressed by a population scal e intervention. By using infrastructure to induce changes in behavior, a large number of people can be simultaneously targeted. Introducing shared bicycles for transportation has the potential both to increase active transportation and replace ca r trips, j oint goals of mediating health

PAGE 22

7 risks associated with obesity. Research Question and Specific Aims The primary purpose of this dissertation research is to address the following question: Does public bicycle sharing serve as a population scale intervention to catalyze active transportation within a population? The following specific Aims and hypotheses are designed to address this question: Aim 1 : E xamine the physical characteristics and human activity in the built environment in which Denver B cycle stations are situated Aim 1.1 : Identify features that distinguish low and high performing sites. Aim 1. 2 : I nvestigate the integration of Denver B cycle within the existing urban transportation system Hypothesis 1. 2: Denver B cycle stations having greater integration with transportation infrastructure, as indicated by proximity to public transportation stops, on and off street bicycle facilities, and other Denver B cycle stations in the network will experience greater numbers of checkouts. Aim 2 : De scribe the characteristics of Denver B cycle users. Aim 2 .1 : Determine demographic characteristics of Denver B cycle users. Hypothesis 2.1 : Denver B cycle users will be more likely to be male, Caucasian, more highly educated, and have a higher household in come than the general population Aim 2.2 : Investigate factors influencing Denver B cycle use

PAGE 23

8 Hypothesis 2.2: The number of Denver B cycle checkouts by annual members will be related to lifestyle factors indicating ability to replace car use with shared bike use Aim 2 .3 : Determine impacts of Denver B cycle on shifts toward active transportation Hypothesis 2.3 : Denver B cycle annual members will shift mode choice away from car use and toward active transportation via shared bicycles Aim 2 .4 : Identify ac tive transportation benefits for Denver B cycle annual members Hypothesis 2.4 : Denver B cycle annual members will exhibit a net increase in quantity of active transportation Aim 3 : Provide an in depth description of the impacts of Denver B cycle as a pub lic health intervention Overview of Methods This dissertation research is a mixed methods quantitative/qualitative study of the impact of a public bicycle sharing system as a health intervention for active transportation behavior. The timeframe for the s tudy was the first season of operation of Denver B cycle, in Denver, Colorado, from April to December 2010. The study uses geographic and spatial data to examine how Denver B cycle is integrated into the existing infrastructural and social components of th e city system, and to provide contextual understanding of how the system is used. Th is investigation also describes use of the system during its initial year, determines the socio demographic makeup of users in order to distinguish participating and non participating groups, and explores factors influencing use of the system. The study detects changes to active transportation behavior of annual members, identifies net active transportation increase due to use of the system, and

PAGE 24

9 desc ribes the impacts of the system on individuals. Quantitative findings derived from exploring the system usage of Denver B cycle are viewed through an evaluative lens informed by qualitative observations and in depth interviews of selected users of the sys tem. The combination of quantitative findings and more nuanced qualitative inquiry provides for a richer and more contextually sensitive understanding of the effects of Denver B cycle on the city.

PAGE 25

10 CHAPTER II LITERATURE REVIEW Development Of The Current Built Environment The Emergence Of Car Dependence The prevailing shape of the urban form in the United States did not emerge on its own; much has been designed to support private motorized transportation. Many of the elements of the built environment incorporate features that, by intention, are deferentia l to the primacy of the car. These physical features have induced the evolution of social preferences, observable in behaviors of the population. As evidence, an accounting of travel behavior in 2009 found that 40% of trips in the U.S. are less than two mi les, and 27% of trips are less than one mile in length, though cars are the mode of choice for 87% of trips under two miles and 62% of trips under one mile (Alliance for Biking & Walking 2012) The present acute dependence on cars for transp ortation is the product of deliberate efforts to change society. At the end of World War II, in his book entitled, When Democracy Builds visionary Frank Lloyd Wright voiced the aspirations of the times when he imagined a society in which the city was dece ntralized, expansive, and supported by ubiquitous personal car transportation (Wright 1945) An ideological dichotomy between progression and regression formed around the mode of personal mobility. Prevailing ideals framed the road to the future as broadly paved, populated with speeding cars, and nearly without limit. Crowded cities with narrow streets full of

PAGE 26

11 pedestrians were emblematic of the past. Ever since, this course of thinking has had vast effects on the scale of urban development. The ambitions of Wright's era have materialized in our present; today it is completely accepted that nearly any daily activity involves or requires the use of a car. In our society, expansiv e and omnipresent roads, along with cheap, plentiful and convenient car parking are not only expected, but are enshrined in the policy of building codes, municipal and regional plans. Decoupling Transportation From Physical Activity The city as a sprawli ng metropolis to support a culture of car dependence is a contemporary phenomenon. Prior to the twentieth century and since the dawn of humanity, most primary travel modes required physical activity. Physically active transportation is any mode in which an individual expends his or her own energy for locomotion, with walking and bicycling the most prevalent forms. Active transportation may be a stand alone activity or incorporated as part of a multi modal trip, such as is the case when an individual walks t o a bus stop or rides a bicycle to a train station. Active transportation fulfills two objectives at once; serving basic transportation needs while integrating healthy physical activity into daily life. Transportation modes have affected urban development throughout the course of time. Cities developed in accordance with the Marchetti Constant, which posits the span of a city is limited to no more than what can be traversed in one hour, historically by foot (Marchetti 1994) The Marchetti Constant has held up to retrospective examination of the evolutio n of cities throughout the world. However, in very recent times, the advent of the fossil fuel age dramatically increased the distance of a one hour travel budget, leading to

PAGE 27

12 striking changes to many of the parameters of cities. First came the train, then the car, each greatly expanding the breadth of cities (Newman and Kenworthy 2006) The city is a complex adaptive system of interconnected actions and reactions, between and among physical and social components. Changes in prevalent travel modes and subsequent changes to urban geography resulted in substantial impac t on perceptions of comfort, utility, safety and appropriateness of modes. An individual consciously or unconsciously considers these items when weighing active transportation against the range of possible modes for any given trip. With the ubiquity of car s, a feedback loop of rapid change developed between people choosing cars and a built environment to support car use. Ordinary travel needs were, in effect, decoupled from expenditure of physical activity. Over time, this decoupling has led to unfavorable health outcomes. As preferred travel mode choice shifted to cars, the daily amount of time spent being physically active dropped (Frank, Andresen et al. 2004; Lopez Zetina, Lee et al. 2006) This, in turn, increased the relative ri sk of obesity; a factor linked with serious health concerns. As active transportation activity dropped, obesity rates among adults skyrocketed, from 13% in 1960 to 35% in 2009 (Alliance for Biking & Walking 2012) The State Of Obesity And Why It Matters The shadow of obesity looms large over the Unites States. Many of the factors influencing the risk of obesity are associated with lifestyle, not only including levels of physical activity and active transportation, but also a dramatically diffe rent nutritional landscape than existed in previous generations (Glanz 2009) The life styles in which many Americans live contribute to greatly increased risks. To put the situation in

PAGE 28

13 perspective, by 2015, an estimated 75% of U.S. adults may be overweight or obese (Wang and Beydoun 2007) A spectrum of elevated hea lth risks is associated with obesity, imposing considerable impact, not only on those individuals who are directly affected, but also on society. Obesity and related conditions contribute disproportionately to the challenge of maintaining a functioning sys tem of healthcare. As healthcare costs continue to increase and incidence of obesity within the population escalates, reduction of costs associated with this preventable condition is of paramount importance for economic sustainability. While ever more peo ple become obese, the expense of treating obesity related diseases are expected to rise dramatically. Total healthcare costs attributable to obesity are predicted to double every decade, to account for 16 to 18% of total US health care costs by 2030 (Wang, Beydoun et al. 2008) A key reason for the high expense of obesity is the many associated chronic health conditions, highlighted in Table II. 1. Several of these conditions are difficult or expensive to treat, and contribute to elevated morbidity and mortality. Some of the most serious and common obesity associated illnesses include c ardiovascular disease, ca ncer, diabetes, and osteoporosis (Warburton, Nicol et al. 2006) Faced with staggering health and economic figures, effective interventions to reduce risk of obesity among large portions of the pop ulation are essential.

PAGE 29

14 Table II. 1 Obesity: Chronic Health Conditions and Risks Premature death Type 2 diabetes Heart disease Stroke Hypertension Gallbladder disease Osteoarthritis (degeneration of cartilage and bone in joints) Sleep apnea Asthma Breathing problems Cancer (endometrial, colon, kidney, gallbladder, and postmenopausal breast cancer) High blood cholesterol Complications of pregnancy Menstrual irregularities Hirsutism (presence of excess body and facial hair) Stress incontinence (urine leakage caused by weak pelvic floor muscles) Increased surgical risk Psychological disorders such as depression Psychological difficulties due to social stigmatization Adapted from: The Surgeon General's Call To Action To Prevent and Decrease Overweight and Obesity 2001 (Office of Disease Prevention and Health Promotion, Cente rs for Disease Control and Prevention et al. 2001) Obesity and related risks are closely associated with inadequate physical activity, a modifiable risk factor. An increase in regular physical activity can significantly reduce the relative risk for obe sity related chronic illnesses (Warburton, Nicol et al. 2006; Haskell, Lee et al. 2007) The recommended quantity of daily moderate physical activity for adults is 30 minutes a day, five days a week (Haskell, Lee et al. 2007) but as little as 15 minutes a day or 90 minutes a week has been found to be of benefit, even for high risk individuals (Wen, Wai et al. 2011) Unfortunately, more than half of U.S. adults do not meet even the minimum recommended levels of physical activity (Centers for Disease Control and Prevention 2005) Obesity Risk Through The Urban Form As beneficial as it may be, opportunities to participate in physical activity can be elusive. Related to the consequences of car dependen ce, as previously discussed, the very

PAGE 30

15 places in which people live often act as a barrier. Numerous studies have established strong evidence between the physical makeup of the urban form and individual physical activity (Berrigan and Troiano 2002; Handy, Boarnet et al. 2002; Sallis, Frank et al. 2004) An ecological study using data from the Behavioral Risk Factor Surveillance System (BRFSS) found that the urban form affects engagement in physical activity and subsequ ently health outcomes (Ewing, Schmid et al. 2003) Infrastructural elements that re duce the viability of non motorized modes affect the quantity and accessibility of active transportation (Frank, Andresen et al. 2004; Frank, Saelens et al. 2007) Over the past 20 or more years, such studies have be come relatively plentiful, building a solid evidence base linking the built environment to levels of physical activity. In application of the Marchetti Constant, the distance a car can traverse in an hour is much greater than that of a pedestrian, so citie s have spread into the surrounding landscape. With burgeoning size, the scale of cities and the form of the built environment encourage automobile use at the expense of physically active transportation (Handy, Boarnet et al. 2002) Car use for most types and distances of trips has become the standard shaping social norms regarding transportation mode choice (Cervero and Radisch 1996) Feedback loops between the built environment and behavioral action have, over time, created cultural expectations for car use over act ive modes of transportation. These expectations are manifest in physical surroundings as high volume streets, acres of parking lots, absent or vestigial sidewalks, and large distances from residential areas to centers of employment. In sum, historically p lentiful opportunities for active transportation have all but

PAGE 31

16 disappeared in much of the modern extended urban setting. Communities that are highly car dependent exhibit limited supportive elements for active transportation behavior. Many metropolitan area s in the U.S. that matured with the advent of car culture are now typified by sprawl, accompanied by the byproduct of undesirable health outcomes among citizens. Beyond The Car: Public Bike Sharing For good or bad, high car dependence and associated impact s are part of modern life. However, elements in a complex adaptive system do not remain static. New solutions are constantly generated to address new problems. In the past few decades, a number of cities throughout the world have begun to meet the challeng e posed by obesity by promoting active transportation behaviors. Among a range of efforts in these cities, several have implemented public bicycle sharing systems. The introduction of shared bicycles into an urban environment is not a new concept. Progeni tors of public bicycle sharing appeared in postwar Europe (DeMaio 2003) Early bicycle sharing attempts were comprised of ad hoc assemblages of old or disused bicycles, often loosely organized by idealists or college students. No strategies for maintenance were incorporated; the bicycles were simply released into the urban land scape to be used freely by anyone. Unfortunately, most efforts of this type quickly fell victim to theft and vandalism. Beginning the late 20 th century, primarily in a few Western European cities, more innovative systems evolved (Sh aheen, Guzman et al. 2010) Experimentation and technological advancement (DeMaio and Gifford 2004) led to mechanisms to induce responsible behavior among users, greatly improving the viability of public bicycle

PAGE 32

17 sharing, and strengthening its reliability as a p ractical form of transportation. Through development, public bicycle sharing systems morphed from localized curiosities into items of interest to leaders of large cities. As of 2011, at least 165 cities around the world have implemented bike sharing (Shaheen and Guzman 2011) In the U.S., several cities, including Denver, Minneapolis, Washington, Bos ton, New York, San Francisco, Philadelphia, Chicago, and Portland have installed or expressed interest in public bicycle sharing systems. Although public bicycle sharing systems have received some high profile attention in the popular media, until very re cently, little academic research has been committed specifically to the subject. Previous to 2010, few peer reviewed articles mention public bicycle sharing either in relation to connection with transit (Pucher and Buehler 2009) or as a potential form o f transportation in the unspecified future (DeMaio 2003; DeMaio and Gifford 2004) Nevertheless, the situation is chan ging with rising interest, with literature related to the health and social impacts of public bicycle sharing growing. Articles on bike sharing outside the realms of transportation or planning are rare. The most substantial bike sharing publication to dat e that features a health component examined the public bicycle sharing system in Barcelona, known as Bicing, from an epidemiologic perspective (Rojas Rueda, Nazelle et al. 2011) The study focused on all cause population mortality rates between car drivers and Bicing users, and found that for Bicing users benefits outweighed risks, as determined by engagement in physical activity, exposure to air pol lution, and traffic incidents. Although the findings of the Barcelona study begin to fill in some of the very large gaps of understanding associated with public

PAGE 33

18 bicycle sharing systems, the authors acknowledged some serious methodological limitations, larg ely due to lack of availability of data. Considerable data gaps in the study include a demographic accounting of Bicing users and evidence based estimates of car use replaced by shared bicycle use. The Barcelona study is important as the first to take a su bstantive look at health effects of public bicycle sharing, but it only touched the surface of understanding the impact of public bicycle sharing on the health of users. Another recent publication explored transportation behaviors among users of a public bicycle sharing system in Hangzhou, China (Shaheen, Zhang et al. 2011) The authors of the study conducted intercept surveys and found that many users are car owners, suggesting th e potential of bike sharing to attract modal share awa y from cars. The Hangzhou study, among the first to have a substantial behavioral component, found that users of the system exhibited a higher rate of car ownership than non users, suggesting that bike sharing appeals to car owners. This finding also suggests selection effects, in that those with more money may be more likely to own cars as well as use shared bikes. Although this study begins to flesh out an understanding of use, the authors acknowledge a number of limitations, primarily that findings are not readily transferrable to other locations or populations. The health impacts of public bicycle sharing systems, and how such systems can begin to diminish societal path dependence on cars for transpor tation are topics that, at present, remain underrepresented in peer reviewed literature. In spite of the data limitation acknowledged in both the Barcelona and Hangzhou studies, the data generation possibilities of public bicycle sharing systems are intrin sically greater than that of traditional studies of bicycle use. Because the operation of bike sharing systems is

PAGE 34

19 conducted through computer databases that record and track individual users as well as the performance of the system at large, the data genera ted are much richer in detail than that afforded by typical methods such as bike counts or surveys alone. Current bike sharing systems now have the capacity to contribute new knowledge about active transportation behavior among urban populations. Guiding And Applied Theories For This Research Denver B cycle is an infrastructural element introduced into a city full of other physical and social elements, all part of a complex adaptive system. In examination of how individual agency results in outcomes, such as how Denver B cycle is used, a framework of supporting theory is helpful. This research uses systems theory for guidance at a conceptual level, suited for examination of multiple influences on population behaviors (Kay 2008) At the level of individual behavior, social cognitive theory (SCT) offers insight in how observed behavior and social interaction between individuals and within sm all groups affects behavior (Bandura 1989) and is used to inform research methods and evaluation. Diffusion of Innovations (DI) theory (Rogers 2003) fits as a connective theory between the broader conceptual and individual levels, to assist understanding of how Denver B cycle use transfers from person to person within the population, affecting systemic changes over time. The holistic nature of systems theory affords a useful perspective in understanding the complexity of interactions between multiple physical and social systems and components. This is particularly useful in understanding the effects of agency on resilience (Bohle, Etzold et al. 2009) in this case, the function of Denver B

PAGE 35

20 cycle over the course of the initial operating season. Th e ecological framework, which shares many conceptual elements with systems theory has been suggested for evaluation of infrastructural influences on active transportation behavior (Ogilvie, Bull et al. 2011) However, systems theory has methodological limit ations that make measurement and testing specific constructs difficult at the individual level. One study used SCT to investigate individual perceptions and social interactions as factors influencing physical activity in university students. The self regu lation skills of a group of 350 young adults attending a personal health class were assessed using a survey instrument, and their self reported physical activity was tracked over four weeks. The study found that a combination of factors, including outcome expectancy, exercise role identity, positive exercise experience, and social support among peers affect an individual's self regulation skills and therefore self efficacy in maintaining physical activity (Petosa, Suminski et al. 2003) These findings suggest that physical activity behavior is in part determined by internalized perceptions and social interactions to which individuals are exposed, either consciously or unconsciously. Internal and external experiences shape the ability of indiv iduals to learn and apply skills for self regulation, a key factor in achieving behavioral self efficacy. In the case of use of Denver B cycle, individuals may weigh many issues when considering whether to regularly use shared bikes for transportation. As in the case of the university student study, internalized perceptions of bicycling, as well as the actions of an individual's peer group are factors affecting active transportation behavior (Rovniak, Anderson et al. 2002; Robertson Wilson, Leatherdale et al. 2008) DI theory provides a lens to explore how decisions of individuals propagate as

PAGE 36

21 behaviors adopted within and among groups, as knowledge of or engagement in an activity makes its way through a population. DI theory expands on many of the constructs of SCT by identifying traits of innovators and early adopters; individuals who serve as popular opinion leaders and who exert asymmetric influence on other individuals within groups (Rogers 2003) In the context of Denver B cycle, the diffusion within the population of the concepts of shared bicycles in general, and how the system functions more specifically, are integral to the implementation of the intervention. Evaluating the effects of multiple social and physical factors on physical activity and/or active transportation behaviors is facilitated with a conjunction of the aforementioned theories. Methodologies incorporating combined theories as ecological approaches have been used to inform devel opment and evaluation of studies of active transportation behavior (Pikora, Giles Corti et al. 2003; Sallis, Cervero et al. 2006; Shannon, Giles Corti et al. 2006) to more fully appreciate influences on behavior at multiple levels. Informing the Study This review identifies literature relevant to bike sharing as an intervention and its potential impacts, and informs the study questions, aims and hypotheses. Relevant current findings are presented, and gaps in knowle dge of public bicycle sharing as a behavioral health intervention are identified. Much is unknown about how public bicycle sharing as is supported through existing elements of the built environment, the patterns of use and influences on participation among users of bike sharing, and the effects bike sharing has on the active transportation behavior and engagement in physical activity of users. The

PAGE 37

22 remainder of this chapter presents themes that are germane to the pursuit of the primary research question: Doe s public bicycle sharing serve as a population scale intervention to catalyze active transportation behavior within a population? Several threads of inquiry support this research. Each thread is supported by a review of relevant literature, in which expose d gaps in knowledge are identified. Intervention Through The Built Environment Studies have established a strong foundation supporting the idea that the form of the built environment influences transportation behaviors, which in turn affect health outcomes (Ewing and Cervero 2001; Handy, Boarnet et al. 2002; Cervero and Duncan 2003; Frank, Andresen et al. 2004; Beech, Fitzgibbon et al. 2011) The existence or absence of urban form elements supportive of active transpo rtation affects travel mode choice, and therefore the potential for inclusion of physical activity into regular habits. These effects are measurable in locales ranging from suburban (Ewing, Brownson et al. 2006) to urbanized areas (Pendola and Gen 2007) The fact that the majority of trips, whether urban and suburban, are conducted by car underscores the mode preference built into the present system, in which a hierarchy of mode choices establishes car transportation at its peak. A concentration on how elements of the built environment result in undesirable effects is a commonality among publications in this arena. What is missing, or at least understated in many studies of the built environment on p hysical activity, is that the findings also indirectly suggest that the built environment itself can be harnessed to induce desirable behaviors. A few interventions, however, have been designed around

PAGE 38

23 how infrastructural changes can affect behavior. One s uch study examined how improved public park facilities affected the local population (Cohen, Golinelli et al. 2009) The Cohen et al. study found that improvements to park in frastructure did not lead to increased physical activity. However, as a physical activity intervention, a weakness in the design was overreliance on an assumption that the targeted population would make the time to plan for and engage in leisure oriented p hysical activity at the park, or to otherwise become self motivated through the presence of park facilities. Although the study found the effects on increased physical activity to be marginal, the authors note that marketing or programmatic assistance to e ngage the community in park activities may have resulted in a more positive outcome. As was the case in the Cohen et al. study, physical elements often receive the lion's share of attention when considering effects of the built environment on physical acti vity. However, social components, such as time availability, perceived safety and other factors may also be important but are less well studied. James Sallis is a proponent of examining the ecological context of behavioral influences on active transportati on (Sallis, Cervero et al. 2006) This approach posits that many factors at many levels exert influence on behavioral action, and that the complexity of situational context should be taken into account. The importance of contextualizing the risk factors and fundamental causes of health outcome s (Link and Phel an 1995) is essential in the development of appropriate interventions that will be received within communities and taken up by populations at risk. Link and Phelan posit that examination of basic social conditions is integral to understanding how social factors affect risk. It is preferential that interventions

PAGE 39

24 designed to target a broad population don't unwittingly exacerbate rather than narrow health disparities. Integrating multilevel approaches into the design of an intervention to affect behavior at numerous leverage points within physical and social context can help to mitigate the potential for undesirable widening of health disparities. Multilevel approaches, incorporating intentionally designed elements of the built environment, and supported by social interaction can be used to target specific behavioral actions. One such study of the High Point public housing community in Seattle incorporated community based participatory research to develop a multilevel intervention to increase physical activit y, primarily for recreation and utilitarian purposes (Krieger, Rabkin et al. 2009) Members of the community identified barriers to participation in ph ysical activity, and then devised methods for remediation. Participants identified needed infrastructural improvements, such as restricted car parking near intersections to increase visibility and improve the safety of pedestrians, and the development of w alking oriented social group activities, effectively modifying social norms to include more physical activity. The High Point study revealed that an intervention involving members of the community could help to create social support and to instill a meani ngful sense of purpose that fit the lifestyles of participants. In the development of interventions that target populations, understanding the context in which the intervention is to occur is essential. However, a shortcoming of the High Point study was th at it was largely focused on recreational physical activity through walking groups, presumably comprised largely of those who had the leisure time availability to participate. The study did not improve understanding of how active transportation habits of r esidents may have been affected.

PAGE 40

25 Changes to the built environment are commonly implemented, ostensibly with goals to increase physical activity. Municipalities have long sought to encourage physical activity through development of parks, trails, on street bike lanes, walking paths and other features encouraging of physical activity, such as skate parks, and basketball or tennis courts. Many studies have identified the presence of these and similar features as being beneficial to health of a residential pop ulation (Ewing, Brownson et al. 2006; Pendola and Gen 2007; Sallis, Saelens et al. 2009) However, direct behavioral effects of specific infrastructure or amenities are seldom measured or studied. Considering parks, trails, and other facilities to encourage physical activity, infrastructure interventions designed to affect behavioral health through physical activity are not uncommon. However, the Cohen et al. study and the High Point study are set apart as examples o f infrastructure intervention s that include assessments of health impacts specific to defined projects. These studies have begun to sketch a picture of how interventions can be applied through infrastructure with the intent of affecting specific health out comes. Still, the Cohen et al. and High Point studies focus on small, defined and localized populations and are designed to affect physical activity through leisure activities. Moreover, unlike a public bicycle sharing system, they are not designed for bro ad integration across a gamut of existing urban structures. As discussed previously, obesity is a population scale problem that may benefit from a population scale intervention. However, for that to happen, it is important that an intervention be of a suf ficient size and type to actually affect a population As interventions conducted through infrastructure with measured health outcomes are not common, there are many important questions that remain For example, do existing

PAGE 41

26 physical and social elements of the built environment support an intervention delivered through infrastructure? The present research will address this question. On the social side of the continuum of influential factors, little is known about how Americans in a car dependent society will react to public bicycle sharing. In the High Point study, intervention efforts were designed to actively engage a defined community and tailored to a specific audience. Such is not the case with a public bicycle sharing system. Although the presence of pe destrians and human activity near potential station sites was a factor when the Denver B cycle operators considered station placement, whether people would engage the system was unknown. Some potential users are likely to quickly catch on to the concept; o thers may not. A public bicycle sharing system is an element of physical infrastructure. Stations and bikes are placed according to the plans of the administrators, but how well they integrate with the existing urban landscape is unknown until operation c ommences. Some literature has addressed various aspects of the use of privately owned bicycles in conjunction with public transit (Rietveld 2000; Pucher and Dijkstra 2003; Martens 2004; Givoni and Rietveld 2007; Martens 2007) Other research has focused on how urban environments might be designed to support bicycle trip choice (Dill and Carr 2003; Sallis, Frank et al. 2004; Dill 2009) but none directly address es the interrelation of bike sharing with existing transportation systems. Intervention And The General Populations The populations that have been subjects of study in physical activity interventions vary demographically. However, most studies to date have targeted children or narrowly defined adult populations. Denver B cycle has the potential to target broader populations;

PAGE 42

27 particularly young and middle aged adults. Studies of groups delineated by age are common, with school aged children frequently target ed. Investigation of active transportation behavior among children and teenagers has revealed connections between active commuting and risk of obesity. For example, a longitudinal cohort study of kindergarten and first grade children in Quebec found that a ctive transportation to school was predictive of lower body mass index (BMI) scores, an indicator of obesity (Pabayo, Gauvin et al. 2010) The Quebec study also found that sustaining active transportation reduced the risk of obesity for children in the early years of grade school. Another study used a social ecological approach to examine the influences on active commuting to school among high school students in Ontario, Canada. In the Ontario study, several variables, notably parental encouragement, smoking status, amount of sedent ary time, physical activity level and perceived weight status were significant indicators of active transportation behavior (Robertson Wilson, Leatherdale et al. 2008) An interesting findi ng from the Ontario study was that the percentage of students who commuted to school via active modes declined in grades 11 and 12, coinciding with the age at which students became eligible to obtain a driver's license. It is logical that children are ofte n the focus of physical activity interventions because of the importance of establishing healthy habits at a young age. At the other end of the spectrum, aging and elderly populations have also commonly been subjects of focus. As an example, a randomized c ontrolled trial examined the impact of two different variants of a physical activity intervention among population of Dutch adults aged 50 and over. One group received tailored letters with physical activity advice. Another group

PAGE 43

28 received enhanced informat ion, encapsulating the same information given to the other group, but were additionally provided with access to an online buddy forum with more tailored information, and maps of bicycling routes in their community highlighting walking and bicycling possibi lities (Stralen, Vries et al. 2009) The Dutch st udy found that the enhanced intervention strategy significantly improved bicycling behavior over the basic information strategy. The increase in individual awareness of nearby bicycling supportive facilities contributed to improved efficacy of the interven tion. Preventive interventions for groups at either end of the age continuum are important. As with the study of active transportation behavior of children, a focus on older populations is also logical. The percentage of the total population made up of agi ng adults is on the rise, and treatment of preventable conditions among this group are an area that may be targeted for health care cost reduction. It should be noted that Denver B cycle may be able to recruit users from the older population, but children under age 18 are not eligible to use the system. The majority of Denver B cycle users are likely to fall in age between young adults and the elderly. Physical activity interventions for the portion of the population in this range have also been studied. Ma ny interventions tar get specific demographic groups For example, a physical activity intervention that focused on middle aged urban women in Germany found that a group that was taught self regulation techniques was substantially more active than a group t hat was only provided with health information (Stadler, Oettingen et al. 2009) Groups targeted by race or ethnicity are also typical of intervention efforts. As an example, a physical activity intervention designed to be culturally and l inguistically adaptive to Latina women found that contextual tailoring of

PAGE 44

29 the intervention strategy was more effective in increasing physical activity than health information alone (Pekmezi, Neighbors et al. 2009) Some groups that have been the focus of physical activity interventions are demographic ally diverse but share other s imilarities within the group Groups defined by diagnosed health conditions, such as diabetes status are not uncommon. For example, participants of a physical activity intervention in Florida who had all been diagnosed with type II diabetes received tailor ed classes and information based on their current activity levels and motivational readiness (Dutton, Provost et al. 2008) The results of the Florida study showed no significant difference in physical activ ity or progression along the "stages of change" outcomes as compared to usual care. Other physical activity interventions that comprise more heterogeneous groups make assumptions of participants, which, in effect create homogeneity of a different sort. On e such study examined the effectiveness of an email/internet based diet and physical activity intervention among employees based at a large health care administrative complex (Sternfeld, Block et al. 2009) The employer based intervention found that the web delivered intervention, which centered on a personalized web page where each participant could track progress toward goals, increased self regulation and adherence to behavior over basic information alone. Although the employer based study population was mixed, all participants shared the same employer and work location. By design, most of th e aforementioned interventions include only people who have made conscious decisions to participate, and are aware of the goal of increasing physical activity. This is not an inherent weakness of such interventions, but overt objectives toward physical act ivity might not be sufficient to appeal to members of the

PAGE 45

30 general population at risk, or may even turn off a portion of the target population. Although the majority of the studies of physical activity interventions are thoughtfully designed and conducted, and are informative in reference to specific populations, conclusions obtained through examination of more homogenous often motivated groups are not readily applicable to broader groups of people. Therefore, findings from these studies may not necessarily apply to the present research. Among studies of physical activity interventions, the least represented population is that of mixed young to middle aged working adults of the general population. Coincidentally, this population is comprised of many who are presently healthy and who might most benefit from a preventive health intervention, before drifting toward increasing risk of obesity Young to middle aged adults in the U.S. are less active than counterpart populations in other industrialized countries (Bassett, Wyatt et al. 2010) The goal of Denver B cycle is to target a range of people to prevent or help slo w down rising rates of overweight and obesity. If successfully implemented, the impact of such an intervention will likely expand beyond the initial scope, because the group of young to middle aged adults encapsulates many decision makers, ranging in juri sdiction from the family level to the public policy level. Affecting change in the physical activity behaviors of the general adult population is essential to mediating the widening threat of obesity and related conditions. A potentially important result of Denver B cycle's reach to younger and middle aged adults is that they are often embedded within strong familial and/or social systems, where their beliefs and experiences can influence others. In fact, some intervention studies have examined how family and social context can be used to induce change in

PAGE 46

31 active transportation. Fitness and health can be linked to the familial environment, including activity preferences (Gruber and Haldeman 2009) Children who grow up observing and taking part in active transportation behaviors may continue to do so as an adult. Family resources are also a factor for active transportation. Children of families with lower household income have been found to be more likely to actively commute to school (Babey, Hastert et al. 2009) Much of the focus of these and similar studies centers on only a single trip type: commuting to school. Introducing active transportation to school commuting habits is certainly important, but b y concentrating primarily on school commuting, opportunities for many other family or social utilitarian trips in which active transportation can be encouraged may be missed. As the largest segment of the working population, most young and middle aged adu lts commute as part of employment. A physical activity intervention to reach this population would logically encourage active transportation as part or all of a commute, or for incidental trips during the day. Public bicycle sharing is well suited for this purpose. By integrating a public bicycle sharing into the transportation infrastructure of a city, a convenient alternative to car use for short trips is offered to a population, simultaneously introducing an opportunity for moderate physical activity amo ng users. The popularity of existing public bicycle sharing systems in many cities in the world leaves little doubt that people do participate, but to date no literature exists on the demographic makeup of users. Authors of the Barcelona study acknowledge d using central estimates of the population, as opposed to specific demographic data of users, noting this item as an objective for future research (R ojas Rueda, Nazelle et al. 2011) The authors of the Hangzhou study found that many bike sharing users were working age

PAGE 47

32 adults using the bikes for part or all of a commute (Shaheen Zhang et al. 2011) Yet, the cultural history of China in relation to utilitarian bicycling differs so greatly from the equivalent population group in the United States so as to make findings less generalizable Although the demographic makeup of public bicycle sharing system users has not been studied, some is known about the demography of bicycle commuters and others who use bicycles for utilitarian transportation, chiefly that the majority are male (Garrard, Rose et al. 2007; Dill 2009) tend to be Caucasian, and have higher household income than the general population (Alliance for Biking & Walking 2012) Beyond gender, race and household income, other known demographic characteristics of bicycle users are scarce. Denver B cycle requires a credit card for parti cipation, posing a barrier for those at the lower end of the socio economic strata. As commuting by bicycle is an activity in which only 2.2% of the Denver population engage (U.S. Census Bureau 2011) this alone is enough to set bicycle commuters apart from the general population, and underscores the likelihood that users of bike sharing will also exhibit differences. The gap in demographic knowledge of who uses a public bicycle sharing system in a major U.S. city is large. This research will provide early estimates of the populations that use and do not use Denver B cycle in its first year of operation. A lthough Denv er B cycle will reach a broader population of young and middle aged adults than other physical activity interventions, it is likely that users will not reflect the general population in important ways. The effects of environmental context

PAGE 48

33 Throughout much of the world, bicycles are very common vehicles for everyday transportation. In rapidly developing countries such as China and India, bicycles have historically been and remain an important vehicle, even as car ownership increases (Kenworthy and Hu 2002) In many countries, especially evident in Asia and Western Europe, commuters ride bicycles as a solitary mode, or as an initial or final leg of a journey in conjunction with public transportation (Martens 2004; Martens 2007) Bicycles are even important vehicles for commerce, being used to carry passengers and cargo (Pucher, Peng et al. 2007) In much of the world, in both developed and developing nations, bicycles are firmly established within transportation systems as a vital component. The situation for transportation bicycle use in the Unites States is very different. In the U.S., the percentage of commuters who choose to ride bicycles, referred to as the bicycle commuter mode share, remains consistently low. As shown in Figure II. 1, the 2010 American Community Survey found that just 0. 5% of U.S. commuters rode a bicycle to work, a figure that has remained mostly unchanged nationally for years (U.S. Census Bureau 2011)

PAGE 49

34 Source: U.S. Census 2010 American Community Survey (U.S. Census Bureau 2011) Figure II. 1 U.S. Cities Larger than 400,000, Ranked by Bicycle Commuting Exceptions to the low norm exist. Portland, Oregon achieved twelve times the national average with 6.0% bicycle commuter mode share. By comparison, Denver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

PAGE 50

35 ranked sixth in the same category with 2.2% bicycle commuter mode share, more than four times the nat ional average (U.S. Census Bureau 2011) As respectable as a 2.2% bicycle commuter mode share in Denver is compared to other U.S. cities, it pales in comparison with Western European and Asian cities where bicycles are used for 20% to 50% of all trips (Pucher and Dijkstra 2003; Pucher, Peng et al. 2007) Installation of a public bicycle sharing system in Denver is not likely to institute sweeping, immediate changes, but it will probably have some affect on bicycle use in the city. While it is impossible to consider all possible contributors, several key factors may affect the active transportation behavior of users of Denver B cycle. The perceived safety of bicycling is a factor that influences active transportation behavior (Dill 2009) Perceptions of safety are related to the type and availability of bicycle supportive infrastructure and the form and characteristics of the built environment. As evidence, higher levels of bi cycle supportive infrastructure have been found to be positively and significantly correlated with elevated rates of bicycle commuting (D ill and Carr 2003) Bicyclists prefer to use infrastructure and facilities that are supportive of bicycling, even to the point that increased travel time is considered to be worth the trade off to ride in a marked bike lane as opposed to an unmarked str eet (Tilahun, Levinson et al. 2007) Socio economic factors, such as fuel prices, have an influence on transportation mode choice. A study of recent dramatic escalations in fuel prices found that rises in fuel price corresponded to a small, yet statistically significant increase in transit ri dership (Lane 2009) Although th e fuel price study was focused on public transportation, the observed changes in behavior may well translate to higher use of other modes, including

PAGE 51

36 bicycles. The demand elasticity of fuel prices does not appear to carry large, long term effects, with rese arch showing that for every 10% increase in gasoline prices there is a corresponding 1.2% increase in transit use (Litman 2004; Currie and Phung 2008) However, a prolonged and significant change in fuel prices could yield systemic effects. As a type of public transportation of its own, and as a fe eder mode designed to support access to light rail and bus transit, Denver B cycle is an option for commuters seeking relief from fuel prices The influence of weather on bicycling habits is less simple to understand when comparing areas with similar climat es, such as the U.S. and Europe. Many countries with colder climates than much of the Unites States sometimes exhibit higher bicycle commuting mode shares. For instance, Canada's Yukon Territory, at the same latitude as Alaska, has more than two times the bicycle commuter mode share of California (Pucher and Buehler 2006) Although Copenhagen is subject to cold and inhospitable weather conditions, it is known for swarms of bicyclists, in contrast to U.S. cities at comparable latitudes. However, no population is immune to the cold when it come s to outdoor activities, and weather extremes have been found to reduce the propensity to ride a bicycle (Richardson 2000) Adverse weather, however, has been found to affect the quantity of recreation bicycle trips more than utilitarian bicycling trips (Koetse and Rietveld 2008) This finding may signal that people who depend on Denver B cycle for transportation may continue to use the system even after recreational riders have largely abandoned their use during cold weather. The preceding key factors are only a few that may influence Denver B cycle use. When taken in context of multiple and unpredictable emergent properties of a complex

PAGE 52

37 adaptive system, it is challenging to determine what will influence Denver B cycle use in i ts first season. However, given the paucity of research of bike sharing systems in the U.S., the present research will help address a few important gaps. Key gaps include understanding the rate at which use of shared bicycles will occur, which geographic sections of the service area will experience the most use, and whether there will be discernable differences in the way annual members use the system as compared to short term users. The present research will use quantitative and qualitative methods to ad dress these key gaps in knowledge. Public Bicycle Sharing And Active Transportation A societal shift from a long established pattern of integrated active transportation, to a high dependence on car transportation associated with urban sprawl has had uninte nded consequences. Evidence has found urban sprawl to be associated with elevated risk of obesity (Ewing, Brownson et al. 2006) Areas dominated by a single, prevailing land use variant, such as vast residential subdivisions without integrated retail or commercial elem ents, have also been found to have an adverse impact on obesity risk (Cervero 1995; Cervero and Radisch 1996) Wh en employment destinations are far removed from areas of residence, car dependent commutes are common. In the past, land use that comingled residential and commercial areas enabled walking or bicycling commutes, thus integrating beneficial physical activit y into daily life. The present state of commuting is dominated by car use, which contributes to sedentary time, decreases integrated physical activity, and increases health risks. In satisfying commute needs, a healthy behavior has been directly replaced b y an unhealthy behavior. The rising prevalence of obesity parallels the dominance of car culture. Between

PAGE 53

38 1960 and 2008, active transportation dropped, while obesity rates steadily rose (Alliance for Biking & Walking 2012) The effect appears localized largely to obesity status, rather than overweight status, as the percentage of overweight population remained fairly stable during the same period (Ogden and Carroll 2010) Links between car use and risk of obesity have been identified, with an important study finding a 6% increase in obesity risk for each hour spent in a car daily (Frank, Andresen et al. 2004) The quantity of car vehicle miles traveled (VMT) has expanded in tandem with obesity rates, a phenomenon not limited to the United States, but also observed in other countries with elevated car dependence (Jacobson, King et al. 201 1) The Jacobson et al. study found a six year lag time between increased VMT and increased obesity rates, displaying evidence of the cumulative risk effect posited in the Frank, Andresen et al. study of 2006. This is an environmental change in lifestyl e, and like that of an increase in caloric availability in the modern diets as compared to the past (Glanz 2009) contributes to cultural and contectual influences on the overall risk of obesity. Health risks attributable to the influence of car dependence extend beyond the physical occupation of a car, and are embedded within the physical and social fabric. Generations of people have been enculturated with social norms embracing the use of cars to fulfill nearly any transportation need. As evidence, areas designed or developed with integral car dependence have been found to exhibit higher r isks of obesity within the local population (Berrigan and Troiano 2002; Ewing, Schmid et al. 2003; Lopez Zetina, Lee et al. 2006) Such findings reveal, in part, the complexities that have developed surrounding socie tal car dependence, and suggest reducing that dependence as a primary intervention objective.

PAGE 54

39 Ongoing research has identified agendas for physical activity and active transportation that coincide (Sallis, Frank et al. 2004) In the disciplines of planning and public health, mixed land use is a strategy to improve conditions for walking and bicycling transportation by providing common destinatio ns close to residences. The concepts of Transit Oriented Development and New Urbanism incorporate mixed land use among residential, commercial and office spaces to ensure likely destinations are within range of non motorized modes, using the form of the bu ilt environment as a positive behavioral determinant (Handy, Boarnet et al. 2002) The incorporation of active transportation strategies within community design holds promise both for planners and engineers seeking pragmatic solutions to move people fr om place to place, as well as policy makers and public health professionals who desire to improve the health status of a population. A transition from motorized to non motorized modes results in health benefits to individuals and to communities, in terms o f physical activity, air quality and economics (Rabl and Nazelle 2012) although active transportation does increase risk of injury from collisions as compared to motor vehicle users (Reynolds, Winters et al. 2010) Bicyclists are indeed more physically at risk and exposed to exhaust pollutants as compared to car drivers, but when health benefits of being physically active are weighed against health risks, the results indicate, on average, a strong net benefit for participation in active transportation (Hartog, Boogaard et al. 2010) The Barcelona st udy assumed a very high rate of car trips replaced by shared bicycle trips. The key assumption by the researchers was that 90% of users were new bicyclists who shifted their use kilometer for kilometer from car trips (Rojas Rueda,

PAGE 55

40 Nazelle et al. 2011) While it is possible that this may be the case, the figure seems overly optimistic on two major counts: in assuming a vastly altered and solidly maint ained behavior of new bicyclists almost completely eschewing car use, and that an enormous majority of shared bicycle trips are replacing car trips. The Hangzhou study made no assumptions about the replacement of car use by shared bikes, finding that a shi ft from car trips to shared bikes occurred, but also that shifts from other modes including transit, walking and biking to shared biking also occurred (Shaheen, Zhang et al. 2011) Although the authors note that the historically high bicycle transportation rates in China have been declining, car ownership remains much lower than in the U.S. Although the findings of this study are interesting and useful within the context of the stud y environment, the cultural and infrastructural differences between China and the U.S. do not make direct comparison possible. It remains important to understand the modal shift capabilities of public bicycle sharing. However, in order to evaluate the out come of a public bicycle sharing as a health intervention to increase active transportation, identifying any net change in active transportation behavior among users is essential. In the present research, the focus is on how Denver B cycle has changed acti ve transportation behavior, that is, the extent to which shared bike trips are actually replacing car trips. S ome of the trips made by shared bike are likely to directly replace car trips, resulting in a net increase in active transportation behavior. Annu al members are more likely to be habitual users than short term users for a number of reasons, in part based on ability to replace car trips. Therefore, it is hypothesized that use among annual members will show a net increase in active transportation beha vior.

PAGE 56

41 Public Bicycle Sharing Participation An increased risk of obesity affects a large portion of the population. Any intervention designed to reach the greatest number of people must by necessity be inexpensive per capita in implementation. Evaluation o f physical activity interventions suggest that behaviors develop in complex social and ecological contexts, and that to be cost effective, interventions should be designed to take advantage of specific contexts (Roux, Pratt et al. 2008) in which participants discover their own reasons to engage. Therefore, individuals in the targeted population for Denver B cycle will be largely responsible for developing their own motivation to initiate and continue participation in the intervention. Denver B cycle differs from many previous physical activity interventions in several ways. Participants may or may not be aware of any health benefit to using the system. Upon subscribing to Denver B cycle, users are not provided health information or asked to be involved in a health study. Encouragement for motivation to participate is not delivered through coaching, counsel ing, or individually tailored goals for weight loss or physical activity. Instead, Denver B cycle is simply framed as an inexpensive and convenient alternative to a car for short urban trips. The framing strategy sets Denver B cycle as a lifestyle interven tion. Users of the system, whether consciously or unconsciously aware of the objectives of the intervention, procure their own motivation to self select to initiate action and to self regulate continuance of active transportation behavior through use of De nver B cycle. As opposed to dependence on external health information or counseling to lead to action, individual participants devise their own contextually meaningful reasoning for using Denver B

PAGE 57

42 cycle, finding a place for it within their lives. Lifestyl e interventions are designed to fit within the daily routine of targeted populations, often using physical activity to serve purposes other than that of overt pursuit of fitness (Dunn 2009) Lifestyle interventions for physical activity can encompass everythi ng from using the stairs instead of an elevator to commuting by bicycle. A study of the Active For Life Initiative, a lifestyle physical activity intervention, revealed that physical activity behaviors incorporated into lifestyle are more easily maintained (Wilco x, Dowda et al. 2009) The study also suggests that lifestyle based interventions are a promising conduit through which to deliver evidence based interventions at a community level, an objective that is logistically or economically elusive with individu ally tailored physical activity interventions. A key motivator for initial and continued use of Denver B cycle is the perception of benefit from using the system. A cost benefit study of shifting from car use to active transportation revealed benefits ou tweigh costs from an economic standpoint (Rabl and Nazelle 2012) Costs of car ownership versus bicycle ownership, car parking versus bicycle parking, and perceptions of time budgeting play into decision making. Active transportation can be more convenient than car transportation, and perceived convenience contribut es to adherence of behaviors (Lewis, Marcus et al. 2002; Dunn 2009) Leveraging favorable aspects of active transportation against unfavorable aspects of car transportation serves to attract participants and reinforce benefits to keep people motivate d. Motivation for initial and continued use is central to the design of Denver B

PAGE 58

43 cycle. The system has been designed to facilitate motivation through increased convenience, accessibility and awareness. The majority of Denver B cycle stations are at or near popular destinations in central downtown, and in areas that have high concentrations of residents near downtown, and at major transit stops. In addition, Denver B cycle is promoted through media, primarily through news outlets who covered the launch and o peration of the system. Bicycle advocacy and recreational groups in the community also quickly adopted Denver B cycle into existing events and activities, indirectly promoting the system. Through a combination of visible physical presence and media attenti on, people who spend time in or near central downtown Denver are likely have been exposed to Denver B cycle. Denver B cycle also provides tools to support the self regulation of users. The system website features personalized user pages to help users monit or their progress and track totals. The tools include estimates of miles traveled, calories burned and dollars saved through use of Denver B cycle, per trip and aggregated. Personalized web pages of the sort used by Denver B cycle have been linked to self regulation skills and adherence to changed behavior (Sternfeld, Block et al. 2009) Users must have access to a computer connected to the internet in order to gain the full range of self monitoring tools as a benefit of using the system. It is likely that use of Denver B cycle will fit the lifestyle of some users, though specifics of tha t fit are unknown. Identifying some of the reasoning for behavioral motivation and action of users can aid the understanding of pathways and barriers to use. Knowing how people use the system can assist evaluation of the effectiveness with which the interv ention has been implemented, and provide clues for future application and

PAGE 59

44 improvement to reach a larger share of the population. Gaps in these areas will be addressed using a quantitative/qualitative mixed methods approach. Motivation To Use Public Bicycle Sharing As a health intervention to induce active transportation behavior, Denver B cycle relies on SCT constructs of modeled behavior and social interaction to expose members of the population to the intervention, and to spread use from individual to ind ividual. As an encapsulating theory, DI informs the examination of how the use of Denver B cycle penetrates the population through word of mouth, shared experiences of participants, and the observation of participant behavior by the pool of those who are n ot yet participants. Within the population, individuals classifiable in diffusion theory as innovators are at the vanguard of adoption of new behaviors, followed closely by early adopters, who take on a new behavior as a result of the influence of innovators (Rogers 2003) In its initial year of operation, many Denver B cycle users are likely to fit the profile of innovators or early adopters; risk takers who are willing to try new things. Although the innovat ors and early adopters of a new behavior reside at the narrow front end of the s shaped adoption curve of behavioral uptake (Henrich 2001) the reasoning for initial motivation and adoption of a new behavior by an innovator forms the kernel of reasoning for transfer of that new behavior to others. Identifying what compels early users to convince their peers to try something new is important in understanding how the intervention is diffused. Comm onalities among the beliefs and opinions of innovators and early adopters of Denver B cycle may signal the advent of modified social norms more accepting of bicycle transportation. In a society heavily dependent on cars for transportation, any pathway to b road or meaningful change

PAGE 60

45 will be paved by changes to social norms (Lucas 2009) Diffusion of behaviors that contribute to modification of social norms is key to growing and sustaining the intervention. Discovery of if and how use of Denver B cycle is diffused within the population can begin to provide a glimpse into the momentum of the intervention as use spreads within the population. However, in terms of how diffusion occurs with regard to public bicycle sharing as an intervention, little is known. Through primarily qualitative methods, the present research explores the motivat ional reasoning behind the behaviors of innovators and e arly adopters of Denver B cycle to determine if and how diffusion to the broader population is occurring. Because many users in the initial year will be innovators and early adopters, members of this group will likely be at the fore of the adoption curve, actively promoting the intervention. Some of the initial users will be popular opinion leaders, with the power to influence their peers. These early users will not only be participants in the interven tion, they will comprise a cadre of socially connected leaders who play an important role in the diffusion of key concepts and uptake of active transportation behaviors. Summary Of Literature Review Evidence established in the literature provides an overv iew of how, over the course of the past century, the population of United States has become increasingly dependent on car use to satisfy transportation needs. Infrastructure, policies and social norms have adapted to support cars at the expense of active t ransportation modes. During the same time, physical activity levels among the population have declined, to the present point at which the average person does not meet minimum recommendations. The

PAGE 61

46 intensity of car dependence has coincided with a rise in the incidence of o besity which affects a majority of the population. A substantial body of literature has detailed the link between inadequate physical activity and obesity, and associated chronic health conditions, many of which are expensive or difficult to treat. Past interventions to affect physical activity have often used individually tailored strategies concentrated primarily on homogenous subgroups within populations. However, these strategies are less well suited to counter a population scale proble m. Instead, a broad engagement of the general population is more appropriate for addressing the obesity problem which, directly or indirectly, affects everyone. A small but growing body of literature suggests that elements of the built environment may be used to deliver an intervention to affect physical activity. Other, related literature points to the effectiveness of lifestyle interventions for active transportation as readily supportive of maintainable changes to behavior. A public bicycle sharing sys tem, such as Denver B cycle, is a lifestyle intervention for physical activity applied through infrastructure. In such an intervention, individuals develop their own motivational reasoning to participate. By positioning shared bikes as a convenient active transportation alternative, it is possible to shift car use to transportation bicycling, resulting in a net increase of physical activity. However, as a recently developed concept, little is known about public bicycle sharing in application, and many speci fic gaps in knowledge exist. Theories of behavior change including SCT, DI and complex systems theory have helped to identify gaps in knowledge and to inform the development of methods of the present research by framing the complex environment in which Den ver B cycle is installed. Each of these theories

PAGE 62

47 contributes to informing the present research as to the social contexts in which transportation decisions are made. These theories facilitate the detection and exploration of the processes by which behavior among individuals achieves motivation to action, is displayed and observed in social context, and is adopted by and/or transmitted to other individuals. Interconnections between physical and social elements in the surrounding environment support the operat ion of Denver B cycle as a multilevel intervention. Substantial gaps addressed by this research include: Understanding in more depth how a health intervention applied through the built environment is supported by existing infrastructure Identifying partic ipation in the intervention among demographic groups of the general population Understanding the impact of environmental behavioral influences on users of Denver B cycle Assessing the impact of Denver B cycle on the active transportation behavior of the po pulation Understanding the motivational reasoning for participation in the intervention Examining the diffusion of Denver B cycle as an intervention over time

PAGE 63

48 CHAPTER III METHODS OF DATA COLLECTION AND EVALUATION Study Design A mixed m ethods research design was used for this study, with initial emphasis on quantitative data collection and subsequent qualita tive data collection to inform quantitative findings. Quantitative data were collected from multiple sources, including the Denver B cycle system database, geographic information system (GIS) layers, a survey of Denver B cycle users, and from population re sources, such as the U.S. Census bureau, the Colorado Department of Public Health and Environment, and the Centers for Disease Control and Prevention. Qualitative data were collected through site observations and semi structured in depth interviews of sele ct Denver B cycle participants. Use of mixed methods strengthens analytical power and affords cross validation of findings (Neuman 2003) Initial find ings are used to inductively generate hypotheses addressed through data analysis. Qualitative data serve to enrich understanding and to apply contextual meaning to findings derived from quantitative data. This is especially important in helping to determin e why users exhibit particular behaviors. Answering questions as to why something occurs is essential when considering behaviors and processes that occur in complex adaptive systems because of the vast array of possible causal pathways.

PAGE 64

49 Brief Glossary Of Terms Specific To This R esearch B cycle : B cycle is an equipment vendor for public bicycle sharing systems. B cycle is a consortium originally comprised of Trek Bicycle Company, Humana Healthcare and Crispin Porter+Bogusky, an advertising firm. Denver B cycle : A separate entity from B cycle, Denver B cycle is the name of the public bicycle sharing system in Denver Denver Bike Sharing : A non profit organization that owns and operates Denver B cycle. Study Setting The study setting of the present research is the service area of the Denver B cycle public bicycle sharing system, located in Denver, Colorado, shown in Figure III. 1. The 40 Denver B cycle stations located in the central business district of downtown Denver were the primary focus of this study. O utlying groupings of Denver B cycle stations in the Cherry Creek and Denver University neighborhoods, each encapsulating five stations, were not principal to the investigation. These smaller groupings are not well connected with the downtown major group of stations, and exhibit different operational characteristics than stations in the downtown group. GIS and other geographic data collection concentrated on the City and County of Denver and the operational area of Denver B cycle.

PAGE 65

50 Figure III. 1 Map of Denver B cycle Stations in the Central Downtown Area, within the City and County of Denver (2010)

PAGE 66

51 Data S ources A number of data sources were used in the present research. Table III. 1 shows data sources for specific aims of the study. Table I I I. 1 Data Sources Associated with Study Aims And Hypotheses Study Aims Data Sources Aim 1 : Examine the physical characteristics and human activity in the built environment in which Denver B cycle stations are situated Station site observational data Aim 1.1 : Identify features that distinguish low and high performing sites. Denver B cycle system usage data set Station site observational data Aim 1.2 : Investigate the integration of Denver B cycle within the exi sting urban transportation system Hypothesis 1. 2: Denver B cycle stations having greater integration with transportation infrastructure, as indicated by proximity to public transportation stops, on and off street bicycle facilities, and other Denver B cycle stations in the network will experience greater numbers of checkouts. City of Denver geographic information system (GIS) data Regional Transportation District (RTD) GIS data Aim 2 : Describe the characteristics of Denver B cycle users. Denver B cy cle system usage data set City of Denver geographic information system (GIS) data Aim 2.1 : Determine demographic characteristics of Denver B cycle users. Hypothesis 2.1 : Denver B cycle users will be more likely to be male, Caucasian, more highly educated, and have a higher household income than the general population Denver B cycle user survey dataset 2010 U.S. Census Denver County demographic data 2010 U.S. Census American Community Survey Denver County data 2010 Colorado Behavioral Risk Factor Surveilla nce System (BRFSS) dataset

PAGE 67

52 Aim 2.2 : Investigate factors influencing Denver B cycle use Hypothesis 2.2: The number of Denver B cycle checkouts by annual members will be related to lifestyle factors indicating ability to replace car use with shared bike u se Denver B cycle user survey dataset Aim 2.3 : Determine impacts of Denver B cycle on shifts toward active transportation. Hypothesis 2.3 : Denver B cycle annual members will shift mode choice away from car use and toward active transportation via shared bicycles Denver B cycle user survey dataset Aim 2.4 : Identify active transportation benefits for Denver B cycle annual members Hypothesis 2.4 : Denver B cycle annual members will exhibit a net increase in quantity of active transportation Denver B cycle system usage data set Denver B cycle user survey dataset Aim 3 : Provide an in depth description of Denver B cycle and its impacts as a public health intervention Semi structured in depth interviews of select participants Denver B cycle user surv ey dataset

PAGE 68

53 Data Variables Each data source used in the present research included numerous variables, shown in Table III. 2. Table III. 2 Data Variables Used in this Research, Arranged by Source Data Sources Variables Station site observational data (Collected by PI) See Appendix A for site observation data collection instrument Collected during four one hour periods at each station site: Bicycle count Presence of graffiti, litter and vandalism Notes on observed activi ties Description of area around site Behavior mapping of pedestrian and bicycle routes through site area Denver B cycle system usage dataset (Collected automatically through Denver B cycle system software. De identified dataset provided by Denver Bike Sharing) Collected during subscription: Date of purchase Type of subscription (24 hour kiosk, 24 hour online, 7 day, 30 day, annual) De identified addresses of annual members Collected for each checkout session: User identification number Date and time of checkout Date and time of check in Duration of checkout Station of checkout Station of check in Derived from Denver B cycle system usage dataset (system wide): Total checkouts, by station Total subscriptions by type GPS data from a subset of shared bikes e quipped with GPS units GPS speed and motion logs

PAGE 69

54 City of Denver geographic information system (GIS) data (Retrieved from DenverGIS online repository) Denver County GIS layers: Base geographic features Streets and highways Street names and address locations Bicycle facilities Denver B cycle station locations Layers created using ESRI ArcGIS 9.0 software Network distance to Denver B cycle stations at 150, 300 and 500 meters Network density of Denver B cycle stations at 150, 300 and 500 meters Geo cod ed addresses of Denver B cycle annual members Regional Transportation District (RTD) GIS data (Retrieved from RTD online repository) RTD GIS layers: Transit stops: o Bus stops o Light rail stations

PAGE 70

55 Denver B cycle user survey dataset (Collected through online SurveyGizmo survey via link sent from Denver Bike Sharing to Denver B cycle users) See Appendix B for Denver B cycle User Survey instrument Demographic data: Gender Age group Home ZIP Code Ethnicity Race Educational attainment Household income Health data: Self assessed health status Self reported weight Self reported height Self assessed poor physical health in past 30 days Self assessed poor mental health in past 30 days Self assessed days disrupted by poor physical or mental health in past 30 days Bicycle ownership and skill level data: Bicycle ownership Self assessed bicycle skill level Bicycle transportation data: Bicycle transportation activity Number of Denver B cycle checkouts per week Types of Denver B cycle subscription s purchased Active transportation behavior: Shifts from car use to Denver B cycle, bicycling and walking Active transportation activity in previous 30 days Transit related data: Multimodal bicycle/transit use Multimodal B cycle/transit use Transit use depe ndent on bicycle access Bicycle safety data: Helmet use on private bicycles and Denver B cycle bikes Other transportation data: Car ownership/access Commute mode choice 2010 Colorado Behavioral Risk Factor Surveillance System (BRFSS) dataset Health status Poor physical health days in past 30 days Poor mental health days in past 30 days

PAGE 71

56 2010 U.S. Census Denver County demographic data Population total Gender Ethnicity Race 2010 U.S. Census American Community Survey Denver County data Population by age g roups Educational attainment Household income Semi structured in depth interviews of select participants Transcribed and coded text from interviews Geographic D ata Overview GIS data were retrieved from online repositories maintained by DenverGIS, an agency of the City and County of Denver, and the Regional Transportation District (RTD), the local transit authority. The PI collected geographic observational data at Denver B cy cle station sites, outlined in Table III. 2. Refer to Appendix A for the station observation data collection instrument. Denver B Cycle System Usage Dataset Overview As shown in Table III. 2, variables used for this research are generated and collected thro ugh normal operation of the Denver B cycle system. Denver B cycle has two major subscription categories, annual members and short term users. Annual members must register online through the Denver B cycle website, during which time they are required to pro vide contact data, including an address of record. Short term user types include a 24 hour pass purchased at a kiosk, a 24 hour pass purchased online, a 7 day pass purchased online, and a 30 day pass purchased online. The Denver B cycle system database aut omatically records the subscription data, including the date, time and subscription type of each sale. The system automatically aggregates the number of subscriptions sold by type.

PAGE 72

57 Annual members are given a B card' with an integrated RFID chip that enab les quick, automated checkouts of shared bikes at any station. Short term users must check out bikes through the station kiosk using a credit card. Each time a bicycle is checked out, the system software associates the specific user with a specific bicycle and automatically records the time, date, and station location of checkout and check in, as well as the duration of checkout. These data are collected and aggregated every time a user checks out a shared bike, per individual user, per station and system w ide, maintained with proprietary software and stored on the B cycle system secured data server. Each of the preceding data items were extracted from the B cycle database into a de identified dataset by Denver Bike Sharing, and provided to the PI for the p resent research as comma separated values (CSV) documents. The B cycle branded model of bicycle owned and used by Denver Bike Sharing was designed to have an integral GPS unit. However, during the initial year, most of the bikes did not have operational G PS units on board. However, during September and October 2010, B cycle equipped a test group of seven shared bikes with GPS units that logged route, direction and speed data during checkout. These test bikes were put to service in the fleet of shared bikes and randomly logged GPS data. These GPS data were aggregated and provided to the PI by B cycle. Denver B Cycle User Survey Dataset Overview In September 2010, an online survey was administered to Denver B cycle users who had registered contact information through Denver Bike Sharing. Data variables are shown in Table III. 2, and the complete survey instrument is included in Appendix B. A link t o the survey conducted online through Survey Gizmo was distributed via the Denver

PAGE 73

58 Bike Sharing email newsletter contact list, amount ing to approximately 2,000 individuals. Valid survey responses totaled 599, for a response rate of 30%. The survey collected information regarding demography and other characteristics of Denver B cycle users. An overview of subsections of the survey follows, with a description of the derivation of survey questions, as well as reasons for inclusion in the survey. Demographic in formation : Denver B cycle User Survey participants made up the demographic sample of Denver B cycle users. Respondents to the survey were largely comprised of annual members, with a smaller portion of short term users who subscribed to Denver B cycle onlin e, or otherwise registered to receive the Denver Bike Sharing monthly email newsletter. Survey participants self selected to participate in the study by completing an online survey of users administered in September 2010. Questions and response categories were adapted from similar questions asked by the U.S. Census Bureau, so that survey data would be comparable with 2010 U.S. Census data and 2010 U.S. Census American Community Survey data, as listed in Table III. 2. Health status and health related quality of life : Questions and response categories were the same as #'"+&-#*,#$%&# <8;8# I&%'H*3)'(#Q *"+#D'0$3)#C2)H&*((',0&# C5"$&.#ZIQDCC[# \2&"$*3,,'*)& # ZK&,$&)"#]3 )#G*"&'"&#K3,$)3(#',-#B)&H&,$*3,# ZKGK[#<88^[ 7# so that survey data would be comparable. Health data elements used included self assessed health status, poor physical health in the past 30 days, and poor mental health in the past 30 days. Data variables a re presented in Table III. 2. Bicycle ownership and skill level : The PI developed questions and response

PAGE 74

59 categories for this section regarding bicycle ownership and self assessed bicycling skill level. Bicycle ownership indicates whether users have the opti on to use their own bikes instead of shared bikes. The ability to competently ride a bicycle in various environments, including in traffic, are necessary and valuable skills for using a public bicycle sharing system. The confidence of participants in their bicycling abilities is an important indicator of competency. Table III. 2 shows data variables for this category. Bicycle transportation use : The PI developed questions and response categories for this section regarding use of bicycles as transportation. Measures of the use of bicycles for transportation are important in the evaluation of the intervention on active transportation behavior. The survey response dataset is not linked to the Denver B cycle system usage dataset for specific individuals. Therefo re, as an indicator of use, survey respondents were asked to report their number of weekly checkouts of Denver B cycle. Data variables associated with this category are presented in Table III. 2. The questions were developed because no appropriate questions could be located in existing survey instruments. Active transportation : Questions and response categories the same as those used in a survey conducted by the New York Department of Health and Mental Hygiene (New York City Department of Health and Mental Hygiene 2009) An understanding of existing active transportation behavior among survey respondents was important in order to identify established tendencies of participan ts toward engaging in active transportation. Data from this section were used to develop a car use replacement modifier to assess whether Denver B cycle

PAGE 75

60 reduced car use in favor of shared bicycles. A listing of data variables for this category appears in T able III. 2. Bicycling to transit : The PI developed questions and response categories for this section regarding joint use of bicycles and transit. Data variables are shown in Table III. 2, and include items assessing whether respondents accessed transit via Denver B cycle or other bicycles, and if a bicycle/transit link enabled transit use among respondents. Literature suggests mutual support between bicycling and transit (Martens 2004) Therefore, data collection from users was important to detect multimodal linkages in the behavior of Denver B cycle users. The q uestions were developed because no appropriate questions could be located in existing survey instruments. Bicycle safety : The PI developed a question and response categories for this section regarding bicycle helmet use. Data variables for this category ar e listed in Table III. 2, and assess whether respondents wear helmets while riding bicycles. The use of helmets to improve safety while bicycling has long been a subject of debate (Thompson, Sleet et al. 2002) However, during observation, helmet usage among Denver B cycle users was observed as being low. Including this section on the survey allows for inquiry into helmet use behavior among Denver B cycle users. The question was developed because no appropriate question could be located in existing survey instruments. Other transportation : The PI developed a question for this section regarding car ownership. Establishing car ownership status among respondents reveals whether Denver B cycle users have access to a car when choosing a transportation mode.

PAGE 76

61 A question regarding commute mode choice was adapted from the 2010 U.S. Census American Community Survey questionnaire (U.S. Census Bureau 2010) in order to identify commute characteristics among Denver B cycle users for comparison with the general population. D ata variables appear in Table III. 2. Select Participants Following the survey, a subset of individuals, referred to as select participants was identified based on survey responses and were recruited for additional data collection. Of 599 completed surveys 317 respondents (52.9%) indicated a willingness to be contacted for in further research. From this group, t wenty two individuals were recruite d through email and telephone Select participants were purposively chosen based on survey responses to weekly n umber of Denver B cycle checkouts. Human Subjects Review The Colorado Multiple Institutions Review Board granted approval to this project as Protocol #10 0690. One amendment to the original protocol was filed and approved to extend the operational dates o f the study. In accordance with the study protocol, informed consent was obtained from all participants selected for interviews as they were enrolled. Signed consent forms were stored in a secure file cabinet. Randomly selected five digit numbers were use d to identify study participants. A secured server was used for storage of digital data. Methods of Data Collection and Evaluation, by Specific Aim In the following, the methods of data collection and analysis for this investigation

PAGE 77

62 are presented, organiz ed according to the Aims of the study. Aim 1 : Examine the physical characteristics and human activity in the built environment in which Denver B cycle stations are situated Aim 1.1 : Identify features that distinguish low and high performing sites. Data collection : Aim 1 and Aim 1.1 were concurrently addressed using station site observational data. Following the first three months of operation, Denver B cycle system usage data for the total checkouts per station were compiled and evaluated to identify hig h and low performing stations based on quantity of checkouts up to that time. The subset of select stations was chosen for closer examination of characteristics of the surrounding built environment and observable behaviors, in order to understand in more d epth some of the factors that may contribute to station performance. The five highest and five lowest performing stations in the downtown group were designated as select stations listed in Table III. 3. Table III. 3 Denver B cy cle Select Stations Station Name Select Station Performance Designation Market Street Station High 16 th and Little Raven High REI High 19 th and Pearl Street High Union Station High Denver Health Low Five Points Low 15 th and Tremont Low 25 th and Lawrence Low Pepsi Center Low Over the course of the 2010 season, station rankings by checkout changed somewhat. For consistency, select stations remained the same as those initially chosen.

PAGE 78

63 Site observations of each select station during observation intervals were conducted on Tuesday, Wednesday or Thursday, so that normal weekday activities could be observed. The four intervals of site observation were, August 25 and 26, September 28 and 30, October 28 and 29 and December 14 and 15. Obser vations were conducted for one hour at each site during each interval, between the hours of 7:00am and 6:00pm. Individual stations were observed at different times during successive intervals so that activities at different times of the day could be record ed. Days with adverse weather were avoided to reduce weather related effects on observed activities. Appendix A contains the instrument for select station observation. A satellite photo of each station location was retrieved from Google Maps and inserted i nto the map area on the observation sheet. During each observation, data items collected included behavior mapping of routes taken by proximate pedestrian and bicycle traffic, trace measures indicating presence of graffiti, litter and vandalism, and a writ ten description of the built environment surrounding each select station site, as well as activities of human presence. Pedestrian and bicycle routes through the vicinity of station sites were recorded during each observation. Behavior maps of observed ad jacent pedestrian and bicycle routes were drawn over the Google Map on the observation sheets. For pedestrian routes, low traffic routes were designated as having fewer than 10 individuals per hour, medium traffic routes had up to 100 individuals per hour, and high traffic routes had more than 100 individuals per hour. For bicycle routes, low traffic routes had fewer than 5 bicyclists per hour, medium traffic routes had up to 20 bicyclists per hour, and high traffic routes had more than 20 bicyclists per ho ur.

PAGE 79

64 Data analysis : Grounded visualization is a qualitative analytical technique that allows for integrative analysis of quantitative GIS data and qualitative spatial data (Knigge and Cope 2009) Descriptive analysis and grounded visualization of behavior maps and other observations of the built environment collected at select station locations were used to examine the interaction of human activity, Denver B cycle operations, transit and bicycling supportive facilities. Particular attention was paid to evaluating the quantity and type of pedestrian traffic near each select station, as well as the nature of any activities observed in the vicinity of station sites. F indings from grounded visualization analysis were compared with GIS data to inform and validate quantitative observations. Aim 1.2 : Investigate the integration of Denver B cycle within the existing urban transportation system Hypothesis 1. 2: Denver B cycle stations having greater integration with transportation infrastructure, as indicated by proximity to public transportation stops, on and off street bicycle facilities, and other Denver B cycle stations in the network will experience greater numbers of checkouts. Data collection : GIS were retrieved from repositories maintained by DenverGIS and RTD as detailed in Table III. 2. Base layers for Denver County, including layers showing streets and highways, street names and address locations, bicycle facil ities, Denver B cycle station locations, and light rail and bus transit stops. Additional layers were created to show network distance to Denver B cycle stations of 150, 300 and 500 meters using ESRI ArcGIS 9.0 software. GIS layers were assembled for analy sis in ESRI ArcGIS 9.0. The data variables B cycle/transit use and $ ransit use dependent on bicycle access from the Denver B cycle user survey dataset were also used.

PAGE 80

65 Data analysis : Geographic network distance tools are useful to describe interconnectivity of access to facilities in context of non motorized transportation (Oliver, Schuurman et al. 2007) A geographic framework can assist in elucidating patterns within dynamic systems of urban transportation (Shaw and Xin 2003) Network distance of the proximity of Denver B cycle stations to other Denver B cycle stations, to bus and train transit stops, and to on and off street bicycle facilities was used to evaluate the level of integration of Den ver B cycle with itself, and within the existing system of public and non motorized transportation infrastructure. Using ESRI ArcGIS 9.0 software, network distances of 150 meters, 300 meters and 500 meters were used to examine the area proximate to Denver B cycle stations. The network distances of 150 and 300 meters are derived as reference distances from the stated goals and emerging practices of station placement within large and established public bicycle sharing systems in Europe and elsewhere. The 150 meter network distance (the maximum distance from a station in the service area given the target of 300 meter distance between stations) is similar to the distance in several large public bicycle sharing systems, and 300 meters is a doubling of that dista nce. Additionally, a network distance of 500 meters was used as a reference distance to compensate for the lower geographic density of Denver and the Denver B cycle system as compared to elsewhere. Counts were made for each Denver B cycle station of the n umber of transit stops, length of bicycle facilities, and the number of other Denver B cycle stations within each of the three network distance zones. An examination of the interconnectedness of Denver B cycle stations with network distance zones of other Denver B cycle stations was used to determine the level of network cohesion within a mutually supportive system.

PAGE 81

66 Multiple regression analysis using Stata 11 software was perform ed to evaluate Hypothesis 1.2. The outcome variable was total checkouts per st ation for the season. Predictor variables for each station included the number of transit stops (light rail and bus) within 150 meters network distance, presence of adjacent bicycle facilities (on or off street lanes, routes or paths), and other Denver B c ycle stations within 300 meter network density (intersecting 150 meter network distance perimeters between stations). The data variables B cycle/transit use and ransit use dependent on bicycle access from the Denver B cycle user survey dataset were descriptively analyzed using Microsoft Excel. Aim 2 : Describe the characteristics of Denver B cycle users. Data collection : To address this Aim, data from the Denver B cycle system usage database were used, described in detail earlier in this chapter. Data variables for subscription included: d ate of purchase and t ype of subscription ( annual, 24 hour kiosk, 24 hour online, 7 day, 30 day ) compiled weekly. Data used for evaluation of checkout activity include d weekly numbers of checkouts per station and total year end checkouts per station. De identified addresses of annual members were extracted from the Denver B cycle system usage database The Denver street address layer was retrieved from the DenverGIS onl ine repository to facilitate geocoding of the addresses of Denver B cycle annual members. Data Analysis : To address Aim 2, the Denver B cycle system usage dataset was imported into Microsoft Excel for evaluation. Descriptive analyses of the following item s were conducted: weekly cumulative Denver B cycle checkouts in central downtown, number of Denver B cycle checkouts per week from stations in central downtown, and

PAGE 82

67 total year end Denver B cycle checkouts per station in central downtown. Stations in the central downtown group were organized geographically into subgroups of neighborhood service areas, as shown in Figure III. 3. Base GIS layers from Denver GIS (DenverGIS 2011) Figure III. 2 Denver B cycle station Groupings by Geographic Service Areas Weekly average checkouts per station delineated by neighborhood subgroup were tabulated from the Denver B cycle system usage dataset and descriptively analyzed using Microsoft Excel. Membership sales by type per week were extracted from the Denver B cycle system usage dataset and descriptively analyzed using Microsoft Excel. The addresses of annual members collected during registration, and contained in the Denver B cycle system usage dataset were extracted using Microsoft Excel and geocoded into a GIS map layer using ESRI ArcGIS 9.0 software. This map layer was

PAGE 83

68 then overlaid with other GIS layers to examine the dispersal of annual users relative to Denver B cycle stations, and analyzed using grounded visuali zation. Aim 2.1 : Determine demographic characteristics of Denver B cycle users. Hypothesis 2.1 : Denver B cycle users will be more likely to be male, Caucasian, more highly educated, and have a higher household income than the general population Data coll ection : The Denver B cycle user survey dataset, described in detail earlier in this chapter, was used to address Aim 2.1. Data variables from this dataset included: g ender, age group, home ZIP Code, ethnicity, race, educational attainment, household income self assessed health status, elf reported weight, elf reported height, elf assessed poor physical health in past 30 days #$" elf assessed poor mental health in past 30 days #$% icycle ownership #$" elf assessed bicycle skill level, active transportation act ivity in previous 30 days, car ownership/access #$ ',$& ommute mode choice The variable BMI was calculated from the variables elf reported weight and elf reported height using Stata 11 software. Population level data were also collected from 2010 U.S. Census Denver County demographic data for gender, ethnicity and race, 2010 U.S. Census American Community Survey Denver County data for age group, educational attainment and household income. Data were also collected from the 2010 Colorado BRFSS Denver County for population weight categories, health status, days of poor physical health in the preceding 30 days, and days of poor mental health in the preceding 30 days Data analysis : To evaluate Hypothesis 2.1, demographic variables from the Denver B cycle user survey dataset were compared to 2010 U.S. Census Denver County demographic data regarding gender, ethnicity and race and to 2010 U.S. Census

PAGE 84

69 American Community Survey for Denver County data regarding age group, educational attainment and hou sehold income using chi square tests in Microsoft Excel. The variable BMI of each survey respondent was calculated using Stata 11 software from data variables self reported height and self reported weight taken from the Denver B cycle user survey dataset, using the formula (Centers for Disease Control and Prevention 2011) : BMI = weight (lb)/[height (in)] 2 x 703 Descriptive statistical analysis was conducted using Stata 11 software to ascertain the mean, minimum, maximum and standard deviation of the calculated BMI scores of Denver B cycle annual members. The BMI scores of Denver B cycle annual members were organized according to normal, overweight and obese weight categori es and the distributions were compared with equivalent data from the 2010 Colorado BRFSS for the general adult population of Denver County using chi square tests in Microsoft Excel. Health self assessment data of annual members taken from the Denver B cyc le user survey dataset, including health status days of poor physical health in the preceding 30 days, and days of poor mental health in the preceding 30 days were compared with equivalent data from the 2010 Colorado BRFSS for the general adult population of Denver County using chi square tests in Microsoft Excel. The data variable active transportation behavior from the Denver B cycle user survey dataset was organized and descriptively analyzed using Stata 11 software. The data variables self assessed bicycling ability and bicycle ownership from the Denver B cycle user survey dataset were organized and descriptively analyzed using Microsoft Excel.

PAGE 85

70 Data on car and bicycle commuting behavior of Denver B cycle users taken from the Denver B cycle user survey dataset were descriptively analyzed for comparison with 2010 U.S. Census American Community Survey data for Denver County, the State of Colorado, and the Unit ed States. Aim 2.2 : Investigate factors influencing Denver B cycle use Hypothesis 2.2: The number of Denver B cycle checkouts by annual members will be related to lifestyle factors indicating ability to replace car use with shared bike use Data collectio n : The Denver B cycle user survey dataset was used to address this Aim. Specifics of this dataset are described in detail earlier in this chapter. Data variables used included number of Denver B cycle checkouts per week, g ender, age group, home ZIP Code, ousehold income, self assessed health status, self reported weight, self reported height, commuting behavior, shift from car use to Denver B cycle and bicycle ownership Some data variables were derived from existing data variables in the Denver B cycle user survey dataset using Stata 11 software. A dichotomous variable proximity was derived from home ZIP Code variable, indicating in service area or not in service area status. Table III. 4 shows ZIP Codes that contain Denver B cycle stations, which are classified as in the Denver B cycle service area. Table III. 4 Denver B cycle In Service Area ZIP Codes 80202 80205 80264 80293 80206 80203 80211 80265 80294 80210 80204 80218 80290 80209 A variable BMI was calculated from self reported height and self reported weight

PAGE 86

71 variables, using the formula cited earlier in this chapter. A dichotomous variable commute via Denver B cycle was derived by extracting responses for commuting in part or in whole by using Denver B cycle from the commuting behavior variable. A dichotomous variable car replacement was derived from the variable shift from car use to Denver B cycle Data analysis : T o evaluate Hypothesis 2.2, ordered logistic regression analysis using Stata 11 software was perform ed to develop a multivariate model to predict the dependent variable number of Denver B cycle checkouts per week Independent variables included in the model were commute via Denver B cycle car replacement proximity bicycle ownership g ender, age group ousehold income, self assessed health status, and BMI. A second model using multiple logistic regression analys is through was perform ed to develop a multivariate model to predict the dependent variable commute via Denver B cycle Independent variables included in the model were car replacement proximity bicycle ownership g ender, age group ousehold income, self assessed health status, and BMI. Prior to applying the regression models, the assumptions of observations being independent and independent variables being linearly related to the logit were tested and met using the omodel command in Stata 11 (UCLA: Academic Technology Services Statistical Consulting Group 2011) Aim 2.3 : Determine impacts of Denver B cycle on shifts toward active transportation. Hypothesis 2.3 : Denver B cycle annual members will shift mode choice away from car

PAGE 87

72 use and toward active transportation via shared bicycles Data collection : The Denver B cycle user survey dataset, -&"0)*1&-#*,#-&$'*(# &')(*&)#*,#$%*"#0%'/$&)7 was used to address Aim 2.3. The data variable used to address Aim 2.3 was shift from car use to Denver B cycle Data in this variable came from a survey question that asked participants to estimate the frequency of trips they make using Denver B cycle that replace car trips. Response categories for the question were Never,' Rarely,' Sometimes,' Most of the time,' and Always.' Responses to this questi on contributed to a model estimating the frequency with which annual members replace car trips with Denver B cycle trips. Because such trips replace car use with bicycle use, any positive response comprises a net increase in active transportation behavior regardless of any existing active transportation behavior of individual users. Data analysis : In evaluation of Hypothesis 2.3, a weighted car trip replacement multiplier (M CTR ) was derived from the shift from car use to Denver B cycle variable, using the f ormula: M CTR = [(n 1 *0.00)+(n 2 *0.25)+(n 3 *0.50)+(n 4 *0.75)+(n 5 *1.00)]/N The car trip replacement multiplier returned a figure estimating the average percentage of Denver B cycle trips made by annual members that replace car trips. Aim 2.4 : Identify active tr ansportation benefits for Denver B cycle annual members Hypothesis 2.4: Denver B cycle annual members will exhibit a net increase in quantity of active transportation G'$'#03((&0$*3, _#F3#'--)&""#!*.#<9>7#$%&#G&,H&)#I ` 050(&#"5"$&.#2"'4&#-'$'"&$# 6'"#2"&-7#-&"0)*1&-#*,#-&$'*(#&')(*&)#*,#$%*"#0%'/$&)9#G'$'#H')*'1(&"#2"&-#*,0(2-&-# ()!*$)+($!,-*$./$&'*&0.1! #',-# ( uration of checkout for each checkout session logged for

PAGE 88

73 each annual member for the year. In addition, GPS data for speed while in motion from the Denver B cycle system usage dataset was used to calculate an average estimated speed during checkout. G'$'#','(5"*" _# For each annual member, totals were calculated from the data variables ()!*$)+($!,-*$./$&'*&0.1! #',-# ( uration of checkout from the Denver B cycle system usage dataset, to obtain total checkouts total minutes of checkout time total weeks of activity (the number of weeks in which a checkout was logged) and total estimated mi les ridden The figure for total estimated miles ridden was derived from the average duration of checkout time and the average estimated speed during checkout, as calculated from GPS data generated by a test group of 10 shared bikes equipped with GPS unit s, described earlier in this chapter. The totals for all individual annual members for total checkouts total minutes of checkout time total weeks of activity and total estimated miles ridden were summed and averaged to obtain average totals for annual members. F%&# car trip replacement multiplier developed in Aim 2.3 6'"#$%&,#'//(*&-#$3#'H&)'4&&+(5#',-#5&') ` &,-#$3$'("#$3# 0'(02('$&#,&$#*,0)&'"&"#*,#'0$*H&#$)',"/3)$'$*3,#1&%'H*3)9# # F%&#'H&)'4&&+(5#,2.1&)#3]#0%&0+32$"#/&)#'0$*H&#',,2'(#.&.1&)#6'"# 0%')$&-#3H&)#$*.&9#F%�%')$#6'"#&O'.*,&-#$3#*-&,$*]5#$)&,-"#3]#2"H&)#$% )"&# 3]#$%&#"&'"3,9#F%*"#*,J2*)5#6'"#$3#-&$&).*,&#*]#'0$*H&#$)',"/3)$'$*3,#1&0'.&#',# &"$'1(*"%&-#1&%'H*3)#'.3 ,4#',,2'(#.&.1&)"9# # Aim 3 : Provide an in depth description of the impacts of Denver B cycle as a public health intervention

PAGE 89

74 G'$'#03((&0$*3, _# Twenty two select participants were recruited, of which 9 were female and 13 were male. All select participants w ere users of Denver B cycle, at differing levels of involvement. Eight were occasional users ( 1 checkout per week) and 14 were regular users of the system ( 2 checkouts per week). The select participants were split evenly by locality status, between in se rvice area and out of service area ZIP Code of home address, as listed in Table III. 4. Of the group, two were of Hispanic ethnicity. Education and household income levels of select participants encompassed the spectrum of survey respondents. The gender, ag e, occupation, locality status, and level of use for each select participant are shown in Table III. 5.

PAGE 90

75 Table III. 5 Characteristics of Select Participants Gender Age Occupation In service area Out of service area Occasiona l User ( 1 checkouts per week) Regular user ( 2 checkouts per week) Male 30 financial analyst X X Female 34 senior metrics consultant X X Female 43 medical researcher, graduate student X X Male 35 environmental planner X X Male 67 retired mechanical engineer X X Male 39 software developer, graduate student X X Female 21 student X X Female 62 lawyer X X Female 23 communications assistant X X Male 40 systems analyst X X Male 44 mechanic X X Male 52 chief financial officer X X Female 27 marketing manager X X Male 41 small business owner X X Male 44 software engineer X X Female 24 accountant X X Male 46 office manager X X Female 33 high school teacher X X Male 48 telecommunications analyst X X Male 33 software engineer X X Male 58 IT manager X X Female 33 economic development consultant X X

PAGE 91

76 Semi structured in depth interviews of select participants were conducted between December 2010 and May 2011. Interviews typically lasted between 30 and 60 minutes. Select participants were given gift cards of $25, either from REI or Whole Foods in appreciation of their time, after the completion of the interview. The question guide for the interviews appears in Appendix C. Interview questions were dev eloped to probe select participants regarding opinions and use of Denver B cycle, physical activity behavior, and bicycling behavior, bicycling comfort and perceptions of safety. Questions were ordered into the following groups: U,$)3-20$*3, _#I)*&]#-&"0)*/ $*3,#3]#"&(]7#/)3O*.*$5#3]#0(3"&"$#G&,H&)#I ` 050(&# "$'$*3,# # G&,H&)#I ` 050(&#*,*$*'(#&O/&)*&,0& #Q&'"3,"#]3)#]*)"$#2"*,4#G&,H&)#I ` 050(&7#',-# *,*$*'(#*./)&""*3,"9 # # !0$*H*$*&"#*,$&4)'$&-#6*$%#G&,H&)#I ` 050(& !" F%&#/2)/3"&"#]3)#6%*0%#$%&#"&(&0$# /')$*0*/',$#2"&" #G& ,H&)#I ` 050(&7#-&"$*,'$*3,"#H*"*$&-7#"&(] ` '""&"".&,$#'"#$3# 6%&$%&)#G&,H&)#I ` 050(&#]*$"#$%&*)#(*]&"$5(&9# # I*050(*,4#1&%'H*3) _# T36#1*050(&#)*-*,4#%'1*$"# %'H&# 0%',4&-# 6%*()#"*,0& # 2"*,4#I ` 050( &7#%36#4&,&)'(#1*050(&#)*-*,4#%'1*$"#%'H�%',4&-7#',-#6%&$%&)# "&(&0$#/')$*0*/',$"#%'H&#*,](2&,0&-#$%&%'H*3)'(#'0$*3,"#3]#3$%&)# *,-*H*-2'("9# # B%5"*0'(#'0$*H*$5a'0$*H&#$)',"/3)$'$*3, I&%'H*3)#)&4')-*,4#/%5"*0'(#'0$*H*$57# 3/*,*3,"#)&4')-*,4#$%"]#1* 050(&"#]3)#$)',"/3)$'$*3,7#3/*,*3,"#)&4')-*,4# $%&#"30*'(#"$'$2"#',-#'00&/$'1*(*$5#3]#1*050(*,4#]3)#$)',"/3)$'$*3,9 #

PAGE 92

77 B&)0&*H&-#J2'(*$5#3]#(*]&a1&,&]*$" _# V/*,*3,"#)&4')-*,4#%36#G&,H&)#I ` 050(&# .'5#']]&0$#J2'(*$5#3]#(*]&7#',-#%&'($%#',-#&03,3.*0#&]]&0$"#'$$)*12 $'1(&#$3# G&,H&)#I ` 050(&9# # X]]&0$"#3,#3$%&)#.3-&" #G&"0)*/$*3,#3]#%36#2"]#G&,H&)#I ` 050(&#.'5#%'H&# ']]&0$&-#$%"]#3$%&)#.3-&"#3]#$)',"/3)$'$*3,9# # B&)0&*H&-# 1*050(*,4# 03.]3)$#',-#"']&$5 _# F%&#B3)$(',-#I2)&'2#3]# F)',"/3)$'$*3,#*,#B3)$(',-7#V)&43,#-&H&(3 /&-#'#0'$&43)*b'$*3,#"5"$&.#]3)# $)',"/3)$'$*3,#1*050(*"$"9#F%&5#-&$&).*,&-#]32)#(&H&("7#*,#-&"0&,-*,4#3)-&)#3]# 03.]3)$#6%*(&#)*-*,4#*,#$)']]*07#*,0(2-*,4#$%&# "!2.+3$)+($/*)24*"" 7#$%&# *+!'1"*($ )+($&.+/,(*+! 7#$%&# ,+!*2*"!*($%1!$&.+&*2+*( 7#',-#$%&# +.$5)6$+.$' .5 #4)32/"# Zc&((&)#<88^[ 9#G&$'*(&-#-&"0)*/$*3,"#3]#&'0%#4)32/#'//&')#*,#!//&,-*O#K7#',-# 6&)&#/)&"&,$&-#$3#/')$*0*/', $"#-2)*,4#*,$&)H*&6"9#F%&5#6&)&#$%&,#'"+&-#$3# *-&,$*]5#*,#6%*0%#0'$&43)5#$%&5#]&($#.3"$#03.]3)$'1(&9# U,$&)H*&6&&"#6&)&#'("3# '"+&-#$3#*-&,$*]5# 03,0&),"#'132$#"']&$5#$%&5#%'H&#&O/&)*&,0&-#6%*("*,4# G&,H&)#I ` 050(&9# # U./)&""*3,"#3]#G&,H&)#I ` 050(& _# V/&,#&,-/*,*3,"#)&4')-*,4#$%&#]2,0$*3,# ',-#2"&]2(,&""#3]#G&,H&)#I ` 050(&9# # G'$'#','(5"*" _#G*4*$'(#)&03)-*,4"# of interviews were transcribed into Microsoft Word. Qualitative Causal Analysis (QCA) is an analytical method used to identify and explore causal conditions and pathways using qualitative data (Ragin 1999) Within the framework of QCA, matrices of response categories are developed from coded interviews, into which groupings of responses wi th similar context or logic are organized

PAGE 93

78 into "fuzzy sets" (Rihoux 2006) Key words, concepts, and themes were identified and inductively coded into categories organized by thematic question groups, outlined in the preceding. Repeated words, phases and themes, situational context, and commonalities among respondents were noted, an interpretive framing guided by a grounded theory approach (Bernard 2011) A summary table of code definitions appears in Appendix D. The findings of this analysis were compared among and between subgroups of select participants. The coded data matrices were then used to develop deterministic functional models an analytic method of QCA to describ e data and to construct groupings of generaliz ed findings (Rohwer 2011) Following QCA, generalized findings were examined using analytic induction. Analytic induction is a method of examination of social phenomena through the development of hypothetical explanations, which are inductively revised and refi ned in an effort to achieve practical certainty of a hypothesis to fit observations (Robinson 1951) In interpretation of the qualitative findings, quantitative findings were revisited to inform and hone the development of hypotheses derived through analytic induction Quantitative and qualitative findings, when used to cro ss validate each other, allow for greater breadth and depth of understanding (Neuman 2003) O nce qualitative analyses shaped hypotheses with sufficient practical certainty, these f indings were used to form an interpretive lens guided by a grounded theory approach through which quantitative findings were examined.

PAGE 94

79 CHAPTER IV RESULTS Introduction The results of this research are organized by study aim. The specific aim addressed is presented at the beginning of each part, followed by associated hypotheses and analyses. Summaries of findings are offer ed at the end of each segment. Aim 1 : Examine the physical characteristics and human activity in the built environment in which Denver B cycle stations are situated Aim 1.1 : Identify features that distinguish low and high performing sites. Characteristics Of The Built Environment At Select Stations As a public health intervention applied through infrastructure, it is important to understand how Denver B cycle has been incorporated into the space and activities of the surrounding environment. In this interv ention, individuals self select to participate, and determine on their own the frequency and continuity of participation. Examination of the encapsulating space around stations can reveal if, or how, the intervention is being accepted and incorporated. Te n stations of the 40 in central downtown were selected for observation, shown in Figure IV. 1. The five most and five least used stations, as ranked following the first three months of operation, were designated select stations Select stations observed in the study represent one quarter of the 40 Denver B cycle stations of central downtown.

PAGE 95

80 Figure IV. 1 Denver B cycle Select Station Locations Features of the built environment act in conjunction to attract human presence to the station sites in varying degrees. A foundation of literature has identified the strong influence of the built environment on physically active behavior (Handy, Boarnet et al. 2002; Lopez Zetina, Lee et al. 2006; Lee, Ewing et al. 2009) but using the built environment to proactively influence behavior through operation of a PBSS is largely unexplored. Observations of select stations revealed characteristics of the built environment, as well as activities at or near the sites of select stations. A description and summary of observations of each select station appears below. The five high use select stations appear first, followed by th e five low use select stations. A summary of findings follows.

PAGE 96

81 Market Street Station Description of the built environment : The Denver B cycle station at Market Street Station is located adjacent to Market Street and the 16 th Street Mall, to the East of th e southern entrance of the bus station. Market Street Station is a major regional bus station in central downtown Denver. In addition to the bus station, several buses stop along the East side of Market Street, across from the station location. Market Stre et is one way northbound, with two travel lanes of nearly constant car traffic and a large amount of bus traffic. The Denver B cycle station is readily visible from the street and nearby pedestrian areas. The 16 th Street Mall is a major center of activity in the downtown area, including employment, entertainment, retail and commerce. It is a pedestrian/transit mall, served by free shuttle buses in both directions at frequent intervals. A stop for the 16 th Street Mall shuttle bus is West of the station site Several sidewalks converge in the entrance area to the bus terminal to the West of the station site. Many employment centers, hotels, restaurants and retail shops are within view of the station site. A designated parking space for a car sharing vehicle is located a few meters from the Denver B cycle station along Market Street. Bicycle riding and parking are illegal on the 16 th Street Mall. Several other Denver B cycle stations are located within a few blocks of the station site. Trace measures : During s ite observations, small scraps of paper and other debris were sometimes seen in the vicinity of the station site. During at least one observation, a maintenance person was seen sweeping and removing litter from the area. No evidence of graffiti or vandalis m at the station was detected during any of the observations.

PAGE 97

82 Behavior mapping of pedestrian and bicycle activity : A tremendous amount of pedestrian activity was continually present at the Market Street Station site, regardless of time of day. During morn ing and evening commute times, pedestrian activity was very high, with hundreds of pedestrians alighting from or entering buses every few minutes. Foot traffic crisscrossed every available space on sidewalks and much of the adjacent street, as pedestrians entered and exited the bus terminal, or walked along the 16 th Street Mall or Market Street. Midday and lunchtime activity in the area was also high, if somewhat lighter than during commute times. Street furniture and public shaded areas attracted numerous people during each observation. Figure IV. 2 shows a map of pedestrian and bicycle tracks observed at the station site. Bicycle activity in the area was fairly active, with between 18 and 44 bicycles observed being ridden during each hour long observation at this site. Nearly all the available bicycle racks for privately owned bikes were occupied during all observation periods. Other privately owned bicycles were parked on signposts, streetlamps and trees in the area around the station site.

PAGE 98

83 Figure IV. 2 Behavior Mapping: Market Street Station Although illegal on the 16 th Street Mall, bicycle activity on the Mall was not uncommon. There were no on street bicycle facilities along Market Street. Many bicyclists rode northbound in the travel lanes of Market Street while a lower quantity rode southbound along the sidewalks. S ome of the bicycle traffic in the area was due to people entering and exiting the bus terminal with bicycles. Bicycle activity into and out of the bus terminal was most prevalent during commute times. During each observation period, a few people were seen to transition between bus and Denver B cycle.

PAGE 99

84 16 th and Little Raven Description of the built environment : The Denver B cycle station at 16 th and Little Raven is located on the East side of Commons Park, near the intersection of Little Raven Street and 1 6 th Street. The station site is close to several high rise residential buildings with ground level restaurants, retail and commerce, and is readily visible from the street and surrounding area. The area is West of the Millennium Bridge, which connects the area to Union Station and central downtown across several freight railroad tracks. Little Raven Street has a light but frequent car traffic presence. To the West of Commons Park are the Platte River and numerous mixed use commercial and residential buildi ngs. Although the land use in the area surrounding the station site is mixed, many of the buildings in the area are primarily dedicated to residential space. Many of the units appear to be high end apartments and condominiums. A major pathway through the p ark is immediately adjacent to the station site. Several other pathways and sidewalks converge near the station site, and a major pedestrian crossing for Little Raven Street is a few meters to the East. Trace measures : During each site observation, litter was observed at the station site, ranging from small bits of paper and cigarette butts to food wrappers and drink containers. Some of the bicycles at the station occasionally had advertisements for entertainment events taped to their baskets. The map boar d of the station had been tagged with pink paint prior to the August observation, but had been removed by the following observation. The pink tag later reappeared on the station's map. Behavior mapping of pedestrian and bicycle activity : Pedestrian activi ty near the

PAGE 100

85 station site remained light but consistent during each of the observation periods, with the majority of foot traffic passing close to the station site. There were many runners and dog walkers on the paths and sidewalks near the station site. Th e street crossings near the station site maintained a nearly constant pedestrian presence. Figure IV. 3 shows a map of pedestrian and bicycle tracks mapped at the station site. At an observation during morning commute time, a few pedestrians who were apparent annual members in possession of a Denver B cycle B card' were seen to check out bikes from the station then ride in the direction of central downtown. Many of these riders wore business attire and placed handbags or briefcases into bike baskets b efore leaving. Each appeared to be quite familiar with the operation of the docks and the bicycles.

PAGE 101

86 Figure IV. 3 Behavior Mapping: 16 th and Little Raven During midday observations, many pedestrians were present in the vici nity of the station site, eating lunch in the park, walking or exercising. Many midday pedestrians were in small social or conversational groups, primarily consisting of women, who wore business attire with athletic or running shoes. Restaurants, a coffee shop and a public fountain located diagonally from the station site on the East side of Little Raven Street were popular congregation points for pedestrians in the area. Bicycle activity at the station site ranged from 13 to 30 bicycles per hour observed b eing ridden. There were few bike racks for private bicycles within view of the station site, at which a handful of bicycles were parked during each of the site observations. The

PAGE 102

87 main routes for bicycles transiting the area appeared to be in both directions along Little Raven Street and through Commons Park. During the September observation, a B card holder with a dog checked out a bike then proceeded to ride while the dog trotted beside him on a leash, making a few laps around the park. After about 15 minut es, the rider returned the bike to the station, checked it in then walked away with the dog. REI Description of the built environment : The Denver B cycle station at REI is located along a section of the Platte River Trail on the East side of the REI build ing, which is the historic former Denver Tramway Company building. The station site is close to Confluence Park, where the two major off street trail systems of the city meet, and where Cherry Creek joins the Platte River. The station is highly visible to users of the trail network, parks and public areas in the vicinity, but is not easily visible from a public street. Trees and other vegetation along the Platte River partially obscure the view of the station from the opposite side of the river. In addition to the two major multiuse trails that converge near the station site, a few additional tributary trails feed into the trail network nearby. The site is not immediately proximate to any street, making the station one of the few in the system to be largely focused on trail users. The closest streets to the site are 15 th Street about 40 meters to the North and Platte Street on the West side of the REI building. Speer Boulevard runs near the site overhead on a large bridge, but is not easily accessible to thi s area. The station abuts the foundation of a prominent outdoor patio structure and

PAGE 103

88 associated Starbucks coffee shop, which is integrated into the ground floor of REI. The Starbucks at REI and its patio are popular and maintain a high level of human prese nce. The surrounding area has several mid to high density residential buildings and exhibits mixed land use. There are several large public spaces nearby, including Centennial Park, Commons Park and River Front Park. Two other Denver B cycle stations are l ocated nearby. Trace measures : Small scraps of paper and food packaging were only occasionally found at the station site. The grounds of the REI building and adjoining sidewalks and trails, including the area encapsulating the station site, appear to be r egularly maintained and kept tidy. No vandalism or graffiti were observed at the station site during any of the observation periods. Behavior mapping of pedestrian and bicycle activity : The Denver B cycle station at REI is in an area of persistent human pr esence with high pedestrian and bicycle activity. The sidewalks and trails near the station site have a nearly constant flow of foot traffic in all directions. The most frequented routes appeared to be the section of the Platte River Trail immediately adja cent to the station site, as well as over the Platte River bicycle and pedestrian bridge. Figure IV. 4 shows pedestrian and bicycle tracks near the station site. Many pedestrians walking past the Denver B cycle station talked about the system, and some kno wledgeably explained to their companions how the system worked. Environmental canvassers frequent an area a few meters from the station, and were occasionally overheard discussing the Denver B cycle station with each other and with passersby. People on the outdoor patio, which overlooks the station site, were also

PAGE 104

89 overheard to comment on Denver B cycle. During the September observation, a middle aged, apparently overweight woman wearing spandex clothing, a hydration pack and a helmet checked out a B cycle b ike and departed toward the southeast along the Cherry Creek Trail. After nearly an hour, the woman returned and checked a B cycle bike back in to the station.

PAGE 105

90 Figure IV. 4 Behavior Mapping: REI. At midday, small groups of people in business attire often walked past the station site toward or away from REI and Starbucks. The outdoor patio deck as well as the indoor seating area of Starbucks was crowded with people having lunch. It was common for gr oups of B cycle riders to arrive together and check their bikes in to the station before going into Starbucks. Occasionally groups of riders of five or more would arrive, filling or nearly filling the empty docks of the station. On one occasion when a larg e group filled the station to capacity, a couple of riders who were not able to dock their bikes discussed going to the next nearest station to dock their bikes and to walk back to REI, then departed. Many bicycles were observed being ridden near the stat ion site, ranging from 29

PAGE 106

91 to 86 per hour during observations. Bicyclists on privately owned bicycles were very numerous in the vicinity of the station site. Many of those riding their own bikes were wearing cycling clothes and riding road bicycles with per formance equipment. Some bicyclists appeared to be training and rode at speed along the trail. Groups of road bike riders often stopped together to buy drinks at Starbucks. There are numerous inverted U style bike racks close to the station site, many of w hich were occupied by private bicycles during each observation. 19 th and Pearl Street Description of the built environment : The Denver B cycle station at 19 th Avenue and Pearl Street is in an area of recently developed medium to high density residential b uildings, with some retail space at the ground level. The station is located along the North wall of a building housing a telecommunications company and directly across from a large residential building. Commercial storefronts line the ground floor of the residential building across the street. During the first couple of months of observation, a construction project was in progress in one of the storefronts on the opposite side of 19 th Avenue, near the intersection with Pearl Street. The station is readily visible from the sidewalks and streets in the area. Car traffic in the area was light along the three eastbound travel lanes of 19 th Avenue during each observation period. Parallel car parking occurs along both sides of 19 th Avenue. The station site is on a block along the South side of 19 th Street, a one way street heading East, bounded by the two way North South streets of Pearl on the West end of the block and Washington on the East end of the block. A bicycle route runs along 19 th Avenue.

PAGE 107

92 Trace measure s : No evidence of litter, graffiti or vandalism was observed at the station site during the observations. Behavior mapping of pedestrian and bicycle activity : Light but fairly constant pedestrian activity was observed on sidewalks near the station site. A yoga center across the street, then later, a pizza restaurant, which had been the retail space under construction previously mentioned, seem ed to be the most notable destinations for pedestrians in the vicinity. Pedestrians appeared to more commonly frequent the North side of 19 th Avenue than the South side of 19 th Avenue adjacent to the station site. Many pedestrians were seen to exit residential buildings i n the area and walk toward central downtown. Very few pedestrians appeared to be dressed for running or exercising, however there were a few dog walkers. A map of pedestrian and bicycle tracks near the station site is illustrated in Figure IV. 5. Most of t he riders who checked out Denver B cycles from this location were B card holders and appeared to be habituated to the use of the station docks and bicycles. Non B card holders who checked out B cycle bikes also appeared to be familiar with the operation of the system. Many of the users observed during commute were dressed in business attire. Upon leaving the station, most B cycle riders departed in the direction of central downtown. Bicycle activity at the station site ranged from 11 to 20 bicycles per hour being ridden during the observations. No bike racks for private bicycles were visible from the station site, although the occasional bike locked to a pole or tree was observed. Bicycle activity was largely concentrated in the travel lanes of the nearby st reets, but several

PAGE 108

93 bicycles were ridden westbound on the sidewalks of 19 th Street toward the direction of central downtown. Many of the Denver B cycle users who checked out bicycles from the station were observed to depart toward central downtown on the si dewalks of 19 th Avenue in this manner. Figure IV. 5 Behavior Mapping: 19 th and Pearl Street. Union Station Description of the built environment : Denver Union Station is a historic railroad depot and a cornerstone of the Western edge of central downtown Denver. Union Station is being redeveloped into a regional transit hub, projected to eventually serve more than a hundred thousand commuters and travelers a day. During 2010, Union Station was the

PAGE 109

94 final stop on the C and D light rail lines and served as the depot for Amtrak, in addition to being the final stop for the 16 th Street Mall shuttle bus service. The Denver B cycle station at Union Station is located on the East site of the main building, which is the side opposite of the light rail stop and the shuttle bus stop. The station site is not visible from either of these transit stops. The station site is visible from Wynkoop Street, though not prominently, as it is set back from the street. Passengers for light rail acce ss the station via a tunnel under the building, but the tunnel entrance is not visible from the station site. Car traffic along Wynkoop Street is light, and the street features bike lanes in both directions. Several large buildings overlook the station si te, most with ground floor retail and commercial or residential space upstairs. There are several restaurants nearby, and a pedestrian entrance to Coors Field is a few blocks to the North. Trace measures : During the observations, no litter of any kind was noted at the station site. No graffiti was present at any of the observations, but in the final observation, a one inch square area of abraded paint exposing metal was noted on the map frame of the station, caused possibly through vandalism or accident. Behavior mapping of pedestrian and bicycle activity : There is constant pedestrian traffic in the area, mostly along the sidewalks of Wynkoop and 17 th Streets. Several runners and a few dog walkers were seen, primarily along Wynkoop Street. Additional foot traffic in the vicinity was to and from the Union Station building, and appeared to be a mix of commuters and travelers. Figure IV. 6 is a map of pedestrian and bicycle tracks observed near the station site. Some of those who looked to be travelers walked to the station to look at the bikes

PAGE 110

95 and read printed material on the kiosk and map. Groups of travelers were sometimes overheard to discuss the Denver B cycle station, some with complete lack of understanding as to what it was, and others who would talk ab out how they had seen something similar in other cities. In September, two travelers talked at great length about the station in a language other than English, then took several photos of the Denver B cycle station and of themselves in front of it. Severa l users of Denver B cycle who checked out or returned bikes to this station were in business attire, and came from the Union Station building or from the direction of central downtown. These users had B cards and appeared to be familiar with the equipment and the system. Figure IV. 6 Behavior Mapping: Union Station.

PAGE 111

96 Bicycle activity in the area ranged from 19 to 31 bicycles per hour during observation periods. The majority of bicycle activity was along the Wynkoop Street bik e lanes, with several bicyclists entering or exiting 17 th Street. Perhaps as many as half of the bicyclists who rode along Wynkoop Street did not stop at the all way stop sign at the intersection of 17 th Street. Delivery vehicles and taxi cabs were often s een double parked and completely blocking the bike lanes along Wynkoop Street. Several privately owned bicycles were parked at nearby bike racks during each observation. Denver Health Description of the built environment : Denver Health is a large health c are facility to the South of central downtown along Speer Boulevard and North of 6 th Avenue. The facility is comprised of a small campus of buildings surrounding parking lots and a courtyard area. The Cherry Creek Trail, a multiuse off street trail, runs p arallel to nearby Speer Boulevard. The station site is located under a covered walkway in the internal courtyard area of the hospital grounds, and is not visible from any public street. The main entrance to the hospital is to the North of the station site and a parking lot and passenger loading area is to the South of the station site. A few manicured grassy areas and benches are nearby. Other than Denver Health and its associated offices and outer buildings, no other businesses, residences or other desti nations are within view from the station site. No on or off street bicycle facilities are within view of the station site. Trace measures : No litter, graffiti or signs of vandalism were noted at the station site during any of the observed periods. The gro unds of the hospital were immaculately kept, including the area around the station site.

PAGE 112

97 Behavior mapping of pedestrian and bicycle activity : There is a tremendous amount of pedestrian activity passing in all directions around the station site, with the gr eatest concentration of activity along the covered walkway housing the Denver B cycle station. Figure IV. 7 shows pedestrian and bicycle tracks in the vicinity of the station site. Figure IV. 7 Behavior Mapping: Denver Health. However it should be noted that the pedestrian activity observed at this site differs greatly from that observed at other stations. The majority of pedestrians in this area were either headed toward cars in the parking lo t or in the passenger loading area, or destined for other buildings on the campus, usually only a few hundred feet away. Many of these people were in pedestrian mode only long enough to get to a car or a building. A notable number of the pedestrians were of minimal ambulatory capacity, or

PAGE 113

98 were assisting others with ambulatory difficulties. Some were in wheelchairs or used walkers, crutches or canes. Either children or elderly people accompanied many of the others who walked by the station. Several of the people who made their way past the station site stopped to examine the bikes and equipment, or to look at the kiosk and map at the station. Many of these people appeared to be curious about Denver B cycle. Several were overheard to speculate about how the system worked, but did not sound as though they had had any experience in actual use. Some of their speculation was quite incorrect, supposing that the bikes were for hospital employees only or that the bikes could be "rented" and kept for days at a time. All of the correct information on how to use the system was printed on the station kiosk, but was apparently either not read or not understood by the people who made the comments. Bicycle riding activity in the area of the station site ranged from 2 to 10 bicycles per hour during the observations. Most of the bicycle activity was made either to or from bike racks. There were many privately owned bicycles parked at inverted U racks near the station site during each observation. Several of the private bikes p resent were the same at each observation, indicating that they might be those of regular commuters to the hospital. Ridden bicycles were most frequently seen on sidewalks or in the parking lot. During the August observation, two people in hospital scrubs and wearing hospital identification tags checked out B cycle bikes using credit cards and proceeded in the direction of the Cherry Creek Trail. In October, a man wearing a suit and tie with a hospital identification tag rode into the station on a B cycle b ike, checked it in and walked toward an outlying hospital building.

PAGE 114

99 Five Points Description of the built environment : Five Points is so named because of the intersection of Welton Street, Washington Street, 27 th Street and 26 th Avenue forming a five pointe d star. The Denver B cycle station at Five Points is located in the southeast sector of the Five Points intersection, positioned into an architectural alcove in the North wall of a multistory building. The adjacent building has ground floor commercial spac e and covered car parking, with residential apartments on floors above. Welton Street is to the West of the station site, and is a busy one way street with two travel lanes headed to the North, along which lies a North South light rail line. A light rail stop is directly to the North of the station site, along the East side of Welton Street. No on or off street bicycling facilities were visible from the station site. Several commercial and retail destinations are in the area, as are restaurants and places of entertainment. A warehouse and offices for Deep Rock Water are located across Welton Street. Several cultural centers, resource centers and government offices are within a short distance of the Five Points intersection. The building adjacent to the st ation site houses a coffee shop, Coffee at the Point, which opened in November 2010. Zona's, a soul food restaurant across 26 th Avenue from the station site appeared to serve mostly African American clientele. As of November 2010, a sign was placed on the front of Zona's announcing that the restaurant was closed after 40 years at that location. Trace measures : Litter was present a t the station site during the first three observation periods. The litter ranged from fast food containers that had been placed on top of the station's sign, to sports drink containers and beer cans, random scraps of paper

PAGE 115

100 and cigarette butts. During the f inal observation, no litter was observed at the station site. In October, a tag in yellow paint was on the map board on the station. By the subsequent observation, this tag had been removed. No graffiti was noted during other observation periods. No vanda lism was noted at the station at any of the observation periods. However, during the October observation a bullet hole was noted in an aluminum window frame of the coffee shop directly behind the station, only a few inches away from one of the station's do cks, and about two feet above the ground. The bullet hole is about half an inch in diameter, and the projectile was embedded in the window frame without penetrating into the interior of the coffee shop. Behavior mapping of pedestrian and bicycle activity : The intersection of Five Points is a very active area for many types of traffic, with a nearly constant human presence. The intersection exhibits a high level of pedestrian activity, primarily along the sidewalks of Welton Street and on the North sidewalk s along 27 th Street and 26 th Avenue. Pedestrian routes along the South side of 26 th Avenue and Washington Street are generally somewhat more lightly traveled. Many pedestrians cross Welton Street near the light rail stop, at or near the crosswalk, but the re are also many instances of jaywalking in nearly any direction across the intersection of Five Points or adjacent sections of nearby streets. The most popular nearby pedestrian destinations were a branch office of USBank located on the North corner of We lton and 27 th Streets, and Zona's restaurant. Later, Coffee at the Point became another popular destination for pedestrians. In September, pedestrians walking near the station site were overheard to be

PAGE 116

101 talking about, "those red bikes," and discussing thei r impression of how the Denver B cycle system functioned. In October, two women were overheard chatting about how Denver was a "pilot city" for bike sharing, pausing briefly to look at the bikes before continuing on their way. A map of pedestrian and bicyc le tracks near the location is shown in Figure IV. 8. Figure IV. 8 Behavior Mapping: Five Points. Bicycle riding activity in the area around the station site ranged from 8 to 28 bicycles per hour during periods of observatio n. Most of the bicyclists in the vicinity appeared to be riding for utility purposes, with some carrying sacks of groceries and others with backpacks or book bags. Some groups of bicyclists had children riding bikes among them. Few bicyclists that passed n ear the station site wore helmets. No bicycle

PAGE 117

102 racks or bike parking fixtures of any sort were within view from the station site. Bicycles locked to sign posts or other fixed objects were only occasionally observed. Much of the bicycle traffic was from the East on 26 th Avenue, heading westbound across Welton Street onto 27 th Street. Several bicyclists were seen riding in the opposite direction, eastbound from 27 th Street to 26 th Avenue, even though 27 th Street is a one way westbound street. Some bicyclists were also seen riding on sidewalks in the vicinity, and occasionally crossing Welton Street while the traffic light was red. In August, two B card holders approached the station from the light rail stop, checked out B cycle bikes, and departed to the Sout h along Washington Street. In September, the station was empty of bicycles and remained empty during the observation period. 15 th and Tremont Description of the built environment : The Denver B cycle station located at 15 th Street and Tremont Place is about 15 meters from the southwest corner of the intersection. The station is installed in an architectural alcove of a large parking garage, a structure that dominates much of the block on which the station is installed. Trem ont Place is a one way southbound street with three travel lanes and metered car parking next to the sidewalk along both sides, while 15 th Street has four travel lanes and metered car parking along the North side of the street. Both streets near the statio n site have heavy car traffic, and 15 th Street also has heavy bus traffic and many stops along the right hand lane. The station site is dominated by car parking, with some commercial and retail activity. A car mechanic shop and accompanying surface parkin g is located across

PAGE 118

103 Tremont Place from the station site. Across 15 th Street to the North of the station site, there is a surface parking lot on the West side of Tremont Place and the back side of the Sheraton Hotel on the East side. The Denver Pavilions sh opping center and the 16 th Street Mall are a little more than one block to the North of the station site. The Webb Municipal Building, a large office building housing several hundred employees of the city and the center of many city services is located one block to the East. A bicycle lane along the right side of Tremont Place ended about one block North of the station, but was extended to continue past the station site at some point in November 2010. There are no on street bicycle facilities on 15 th Stree t. The nearest off street bicycle facility is the Cherry Creek Trail, several blocks to the southwest. Several other Denver B cycle stations are within a few blocks of this station site, all of which are more heavily used. The closest is about a block away at the Denver Pavilions section of the 16 th Street Mall, at 1550 Glenarm Place. Trace measures : No litter was noted at the station site during the periods of observation. No graffiti was seen at the station site. In September, a one inch long scratch in the red paint of the kiosk sign was noted; minor damage, which may not be ruled out as possible vandalism. Behavior mapping of pedestrian and bicycle activity : The intersection of 15 th Street and Tremont Place has a constant human presence during the day, and is host to moderate but continuous pedestrian activity, as seen in Figure IV. 9. Foot traffi c was observed as steady streams along all sidewalks in the area, with the highest concentration along 15 th Street as compared to Tremont Place. Pedestrians passing through the intersection amounted to several hundred per hour during each of the observatio ns. The

PAGE 119

104 areas with the least quantity of pedestrian traffic are the sidewalks on Tremont Place, South of the intersection. Coincidentally, these sidewalks are the only in the area from which the station site is readily visible. The location of the station partially hidden behind a section of the parking garage, is not visible to eastbound pedestrians on 15 th Street, or to pedestrians on Tremont Place North of the intersection. The pedestrians closest to the station site were often seen entering or exiting the parking garage by a door from which the Denver B cycle station is not visible. Much of the pedestrian traffic appeared to be heading to the East and the North; directions against traffic if a bicycle were to be ridden from the station site on the stre et in a legal and safe manner. Figure IV. 9 Behavior Mapping: 15 th and Tremont Street.

PAGE 120

105 The bicycle riding activity observed in the vicinity of the station site ranged between 31 and 52 bicycles per hour during the periods o f observation. The most frequented bicycle routes were westward along the travel lanes of 15 th Street and eastward along the sidewalks of 15 th Street. A lower quantity of bicycle activity was seen along Tremont Place. Several Denver B cycle riders were se en along 15 th Street, many in business attire and a few in suits and ties or in skirts and high heeled shoes. Some bicyclists on privately owned bikes included people who appeared to be commuting, carrying bags or briefcases, in addition to those who appea red to be bicycle couriers. Very few bicyclists in this area were wearing spandex clothing or appeared to be riding specifically for the purpose of recreation or exercise. 25 th and Lawrence Description of the built environment : The Denver B cycle station a t the junction of 25 th Street and Lawrence Street is located at the northeast corner of the intersection. Lawrence Street is two way with a single travel lane in each direction and parallel car parking on both curbs, and 25 th Street is two way West of the intersection and one way eastbound to the East of the intersection. Traffic is light on both streets. A bus stop is immediately adjacent to the station site, off the northbound lane of Lawrence Street. Sidewalks border both sides of the streets in the area A Denver B cycle station at Walnut and Broadway is about two blocks away to the West, and is visible in the distance from the station site. Lawrence Street, adjacent to the station site, is a bicycle route. The entire block on which the station site is located is vacant and fenced off, awaiting construction of a large housing development. Another block to the northeast is

PAGE 121

106 also vacant. The school and grounds of Crofton Community Elementary is on the block to the southeast of the station site. Low density residential houses and buildings are dominant on the blocks across Lawrence Street to the West. Many of the houses appear decades old and some are dilapidated. The neighborhood seems to have been under redevelopment until recently. A few lots in the area have been developed into high end townhomes, but most of these appear vacant and some have boarded up windows and doors on the ground floor. No commercial or retail offices are visible from the site. The block on which the station is located has a tall fen ce to keep trespassers off the bare land of the future construction site, which contributes to a prevailing atmosphere of exposure and isolation. Trace measures : A large quantity of litter and trash was present at or near the station site during each perio d of observation. Beer bottles, fast food containers, newspapers, scraps of plastic and other random trash items were at the station site. Discarded boards, building materials, ragged clothing, as well as glass and plastic bottles, cans, food containers an d other random trash were visible across the fence on the vacant lot. Litter of various types was also seen in the street gutters near the site. Graffiti was not noted specifically at the station site, but graffiti was present in the vicinity, with tags an d other markings visible on the fronts of vacant townhomes and on poles, signs, benches and other outdoor fixtures and buildings in the area. Tags are prevalent on dumpsters and other objects in alleyways, as well as on fences and the backsides of building s all around the site. No evidence of vandalism was noted at the station site during the observations. However, this station was the only one in the system to encounter a significant episode of

PAGE 122

107 mass vandalism. On a night in June 2010, the tires of all the bicycles docked at the station were slashed. A security camera was installed above the station kiosk following the incident. There have been no reported acts of vandalism at the station since then. Behavior mapping of pedestrian and bicycle activity : Durin g the observation periods at the 25 th and Lawrence Denver B cycle station, very little bicycle or pedestrian activity was recorded, as shown in Figure IV. 10. The combination of low density residences, two vacant blocks and a nearby school that was either c losed for the summer or populated by children under the minimum age to use Denver B cycle contributed to a very low human presence in the vicinity. Fewer than 10 to 15 pedestrians per hour passed near the station site. The sidewalks along Lawrence Street w ere the most prevalent nearby pedestrian routes, although very sparsely used. Figure IV. 10 Behavior Mapping: 25 th and Lawrence Street.

PAGE 123

108 Bicycle riding activity near the station site ranged from 6 to 12 bicycles per hour during the observation periods. Many of the bikes being ridden were heading North along Lawrence Street. Most of the bicycle riders seemed to be commuters, carrying backpacks or bags, and not wearing spandex riding gear. In September, a B card holder appro ached the station on foot from the South, checked out a bike and rode to the East on 25 th Street. Although 25 th Street East of Lawrence Street is one way Eastbound, bicyclists were seen intermittently going both directions along the one way section. No par ked bicycles in the area were noted during observation, and no bike racks were visible from the station site. Observation revealed a couple of items of note regarding public safety in the area of the station site. First, although traffic on the nearby str eets is light, a few cars were greatly exceeding the speed limit along Lawrence Street. In August, two consecutive cars were seen driving North on Lawrence at perhaps 60 miles per hour, and both ran through stop signs along Lawrence to the North of the sta tion site. During other observation periods, cars were also occasionally seen to be traveling at similarly high rates of speed. A second item of concern for public safety was the presence of stray animals. In October, two large and aggressive appearing dogs, both without collars, were seen randomly roaming near the station site and investigating trash in nearby alleys. Animals such as these may pose a threat to pedestrians and bicyclists in the area. Pepsi Center Description of the built environment : The Pepsi Center is a large sporting events arena, notable as the home of the Denver Nuggets basketball team and the Colorado Avalanche hockey team. The Denver B cycle station at the Pepsi Center is located near

PAGE 124

109 the South entrance of the arena at the edge of vast surface parking lots to the South and West. The nearest public street is Chopper Circle, to the southeast of the station. No on or off street bicycle facilities are close to the station site. Twelve inverted U bike racks are located immediately adjac ent to the station site. Several more bike racks are located in other areas on the grounds of the Pepsi Center. The station site is surrounded by broad stretches of concrete and asphalt paved areas. Sidewalks and walkways from many directions converge at the arena. Several privately owned surface parking lots to serve visitors to the arena are the dominant land features of the area. An area to the North of the station site is landscaped with rocks, grass, bushes, trees and flowers, encapsulating team monum ents and informational signs. Other than the Pepsi Center, there are few other buildings or destinations in the near vicinity. A restaurant and bar is located on the far side of Chopper Circle from the station site. A light rail stop serving the Pepsi Cen ter and Elitch Gardens is approximately 150 meters to the North. Very few residential spaces exist anywhere near the station site, with the closest being more than 300 meters to the East. The Auraria Higher Education Campus is South of Auraria Parkway, mor e than 200 meters distant. Trace measures : No evidence of litter, graffiti or vandalism was noted at the station site or in the surrounding area during any of the observation periods. Behavior mapping of pedestrian and bicycle activity : During workday bu siness hours, there is very little human presence at or near the station site, as shown in Figure IV. 11. Administrative offices are housed in the Pepsi Center, but much of the activity appears to be concentrated on the North side of the enormous structure, far from the station site. In the area, fewer than 50 pedestrians per hour were seen during

PAGE 125

110 observations. Much of the limited pedestrian activity is concentrated on the South entrance to the arena and along the sidewalk parallel to Chopper Circle. The do ors of the arena are locked and require a card key to enter. Occasionally, people waiting at the doors would wander over to the station to look at the bikes, kiosk and map. A few pedestrians were seen walking toward the light rail station to the North. Figure IV. 11 Behavior Mapping: Pepsi Center. Bicycle riding activity during the observations periods ranged from 1 to 10 bicycles per hour. Some bicyclists seen near the station wore spandex riding clothes. Other riders wore street clothes and carried backpacks, bags or panniers and appeared to be commuters. In September, two young adult males rode past the station on fixed gear

PAGE 126

111 bicycles then practiced trick riding in the parking lot to the South of the station for more than 3 0 minutes. The plentiful inverted U style bike racks were only sparsely populated with private bikes at any time. Analysis of Characteristics of the Built Environment at Select S tations The built environment at each of the select stations examined in the p receding section exhibits some unique properties. In spite of this select stations in both the high and low performing groups shared certain characteristics, highlighted in Table IV. 1. Table IV. 1 Characteristics of High and L ow Use Denver B cycle Select Stations Pedestrian presence in proximity P edestrian and bike routes in proximity Predominant land use in proximity Visibility of station in proximity Visual incivilities (trash, vandalism) High use select stations Consistent or high Fair to excellent Mixed use Good to Excellent Minimal Low use select stations Intermittent or low Fair to poor Minimal mixed use/ homogenous Marginal to poor Minimal to high A major characteristic observed at high performing stations is nearly constant human presence. High performing stations are in areas of high pedestrian traffic, often with multiple high traffic ingress and egress route by foot or bike. High performing stations are also readily visible to pedestrians, and are in loc ations with popular mixed use destinations and/or concentrated residences or places of employment nearby. REI, 16 th and Little Raven, and 19 th and Pearl are stations that fit this description. Additionally, some high performing stations have strong integra tion with busy public transportation facilities such as those at Market Street Station and Union Station.

PAGE 127

112 Conversely, some low performing stations are nearly devoid of pedestrian traffic, having a smaller pool of pedestrians from which from which to recru it users. The Denver B cycle stations at Pepsi Center and 25 th and Lawrence are examples of having a very limited observed pedestrian traffic. On the other hand, some of the low performing stations actually had a considerable amount of foot traffic nearby, but the nature of the foot traffic in these locations differs in purpose or immediate destination from those observed at high performing stations. The majority of pedestrians near the Denver Health station are only momentary pedestrians, as many are going between nearby buildings or are on their way to parked cars. Visibility is also an issue with some low use stations. The station at Denver Health, as an example, is not visible from a public street from which pedestrians with more distant destinations ma y be drawn. Poor visibility of the station is also a problem at 15 th and Tremont and Five Points stations, both of which are hidden in architectural features of buildings, and are only visible by pedestrians traveling in certain places on adjacent streets. High use stations enjoy apparent strong community support, in terms of both physical infrastructure and social engagement. Some of the most highly used stations have seemingly become a focal point within the built environment, drawing users as a result o f convenient placement adjacent to desired destinations, all while serving substantial populations. The 16 th and Little Raven and 19 th and Pearl stations are in areas of concentrated residences within short commuting distances of central downtown. The inte gration of some stations with major transit stops support users commuting from outside the Denver B cycle service area. Many high use stations are positioned in such a

PAGE 128

113 way as to appeal to regular users who have grown to depend on the bikes to serve their d aily transportation needs. The low use stations exhibit traits that are almost entirely the opposite of the high use stations. Most are poorly integrated within the surrounding area or are in locations of homogenous land use. Most have low or intermittent pedestrian presence. In some cases, the station itself is not readily visible to passing pedestrian traffic, or is not suited to serving the needs of the majority of proximate pedestrians. Some of the lowest performing stations are faced with a huge imped iment; a nearly complete lack of human presence. In summary, those stations that were most highly used simultaneously served a variety of destinations and a sizeable population, and were well integrated into the locale of the built environment. The low us e stations were most commonly in areas with low human presence or fairly homogenous land use in the immediate vicinity. These findings underscore the importance of well considered station placement with regard to reaching and recruiting potential users. I n terms of human presence, high use stations, by and large, have much more observable human presence than most low use stations. However, some low use stations actually have considerable human presence, but the nature of the activities observed being under taken by the majority of people at low use sites was not conducive of the use of shared bikes. Human presence seems to indeed be an important factor contributing to station performance, but the type of activities in which the proximate population is engage d, as well as the basic visibility of a given station appears to be least as important. As a public health intervention, the effectiveness of reaching the greatest quantity of the

PAGE 129

114 targeted population and encouraging continued use among those individuals is greatly influenced by convenient placement and high visibility of stations in places where pedestrians spend time.

PAGE 130

115 Aim 1.2 : Investigate the integration of Denver B cycle within the existing urban transportation system Denver B cycle is dependent on pro ximate access to and from transit stops to attract multimodal users. On and off street bicycle facilities are also important to supply users with safe routes for ingress and egress of stations and to provide a supportive network of routes to reach destinat ions. For Denver B cycle to serve as a public health intervention, cooperative interaction of components within the system itself and between the system and other infrastructural elements is critical. The degree of mutually supportive coordination of Denve r B cycle within the transportation network is used to evaluate the following hypothesis. Hypothesis 1. 2: Denver B cycle stations having greater integration with transportation infrastructure, as indicated by proximity to public transportation stops, on an d off street bicycle facilities, and other Denver B cycle stations in the network will experience greater numbers of checkouts. Network distances of 150 meters, 300 meters and 500 meters were used to examine the area proximate to stations, as discussed in the previous chapter. A map showing network distances to stations in the central downtown area is shown in Figure IV. 12.

PAGE 131

116 Base GIS layers from Denver GIS (DenverGIS 2011) Figure IV. 12 Network Distances to Denver B cycle S tations. Station Network Density of Denver B cycle To achieve a network density of 300 meters between stations, the perimeter of 150 meter network distance boundaries of individual stations would touch each other. Only 13 of the 40 stations in central dow ntown have at least one other station within 300 meters. If the reference distance is expanded to a network density of 600 meters between stations, meaning the 300 meter network distance perimeters of adjacent stations touch, 33 of 40 stations meet this cr iterion. Expanded to a network density of 1,000 meters, in

PAGE 132

117 which the 500 meter network distance perimeters of adjacent station touch, all 40 stations meet this condition. Proximity To Transit Five light rail lines and numerous bus routes service downtown Denver. Regional Transportation District (RTD) stations and stops are shown in Figure IV. 13. Of the 40 Denver B cycle stations in central downtown, 12.5% have a light rail station within 150 meters, 17.5% within 300 meters, and 30% within 500 meters. Of th e 16 light rail stations in central downtown, 5 have a Denver B cycle station within 150 meters (31.3%), 9 within 300 meters (56.3%), and 12 within 500 meters (75%).

PAGE 133

118 Base GIS layers from Denver GIS (DenverGIS 2011) RTD GIS layers from RTD ( Regional Transportation District 2011) Figure IV. 13 Location of Denver B cycle Stations Relative to Transit Stops and Stations in Central Downtown. In central downtown, Denver B cycle stations with bus stops nearby number 33 within 150 meters (82.5%), 39 within 300 meters (97.5%), and all 40 have a bus stop within 500 meters (100%). Because Denver B cycle stations are collocated with the major bus stations in central downtown at Market Street Station and Civic Center Statio n, most of the bus routes in the area have access to Denver B cycle stations.

PAGE 134

119 Proximity To Bicycle Facilities On street bicycle facilities, such as bike lanes, bike routes, shared use lane markers (sharrows) signal to bicyclists, including Denver B cycle users, areas that are designed to support bicycle use. A GIS data layer from DenverGIS, as shown in Figure IV. 1 4 depicts the location of on and off street bicycle facilities in relation to the Denver B cycle stations in the central downtown area. Base GIS layers from Denver GIS (DenverGIS 2011) Figure IV. 14 Denver On and Off Str eet Bicycle Facilities. All of the Denver B cycle stations in the central downtown service area have some form of bicycle facilities within 300 meters network distance, and many have

PAGE 135

120 bicycle facilities directly adjacent to station sites. In the network di stance zones to Denver B cycle stations in central downtown, there are 15.47 miles of bicycle facilities within 150 meters, 29.51 miles within 300 meters and 41.94 miles within 500 meters. Total on and off street bicycle facilities in central downtown Denv er measure 60.58 miles. Multiple regression analysis was conducted to determine the best linear combination of transit stops (within 150 meters), bicycle facilities adjacent, and network density (other stations within 300 meters) for predicting total chec kouts per station. The assumptions of normally distributed errors, homoskedasticity, multicollinearity, and linearity were checked and met. The means, standard deviations, and intercorrelations are shown in Table IV. 2. Table IV. 2 Means, Standard Deviations and Intercorrelations for Total Checkouts and Predictor Variables (N=32) Variable M SD 1 2 3 Total Checkouts 2,891.44 1,280.38 .16 .07 .21 Predictor Variable 1. Transit stops 3.29 4.15 .02 .47** 2. Bicycle facilities .69 .47 .41* 3. Network density .66 .94 p <.05; ** p <.01 The combination of variables did not significantly predict total checkouts, F (3,28) = 1.51, p =0.23. Two variables (transit stops and network density) trended toward significance in the model at the level of p <.1, but the overall model was not significant. The adjusted R squared value was .05, indicating only 5% of the variance in total checkouts w as explained by the model. The beta weights, shown in Table IV. 3 suggest that transit stops and network density contributed most to the model, while bicycle

PAGE 136

121 facilities did not. Removing the bicycle facilities variable improved the model somewhat, but not e nough to render it statistically significant, and insufficient to reject the null hypothesis. Table IV. 3 Simultaneous Multiple Regression Summary for Transit Stops, Bicycle Facilities, and Network Density Predicting Total Check outs (N=32) Variable B SEB # Transit stops 110.07 62.67 .36* Bicycle facilities 250.93 533.93 .09 Network density 565.48 303.61 .41* Constant 2,735.73 434.73 Note. R 2 = .14; F (3,28) = 1.51, p =0.23 p <.1 An examination using Cook's D to identify unusual and influential data revealed that the stations at 15 th & Cleveland and Market Street Station exceeded the threshold of d>4/n, indicating that these stations may influence the outcome of the model more than other stations. However, as these are bot h high performing stations with strong transit connections, and do not vary substantially from other proximate stations, they were not removed. Summary Of Denver B Cycle Within The Urban Transportation System Although collocation of Denver B cycle stations at transit stops was not a significant predictor of the total number of checkouts at stations, connections between shared bikes and transit do contribute to checkout activity. In the Denver B cycle User Surve y, 26.4% of annual members reported using shared bikes to access either light rail or bus stops. A cooperative relationship between shared bikes and transit is important. Bus and light rail ridership benefits from bicycles as a feeder mode, as 20.6% of ann ual members indicated that a bicycle connection to transit was essential to their ability to use transit.

PAGE 137

122 There are apparent relationships between existing transportation infrastructure and total checkouts at Denver B cycle stations; however, the evidence does not support the acceptance of Hypothesis 1.2 The findings suggest that proximity of supportive elements of physical infrastructure likely have some impact on total checkouts. However, infrastructural elements alone are insufficient to predict the qua ntity of station use. The observations collected during investigation of Aim 1 and Aim 1.1 point to the quantity, presence, and activities of pedestrians as being possibly more important in predicting individual station use.

PAGE 138

123 Aim 2 : Describe the characteri stics of Denver B cycle users. Description Of System Wide Patterns Of Use With the goal of describing the characteristics of Denver B cycle users, the present research begins with a description at the macro level before moving focus toward individual level characteristics. The beginning of this section highlights broad descriptive data items over the course of the initial Denver B cycle operating season. These items include checkout activity for stations in central downtown, as well as by individual station and by geographic neighborhood service areas. Membership characteristics include descriptive data of membership sales by subscription type and the geocoded home locations of annual members. Data presented in this section are from the Denver B cycle system usage dataset containing data from both annual members and short term users, unless otherwise noted. Checkout Activity As shown in Figure IV. 15, the number of Denver B cycle checkouts from stations in central downtown increased as the season progressed. A fter opening in April, the quantity of checkouts rose slowly at first. For the bulk of the season, from early June until early October, the cumulative number of checkouts rose consistently from week to week. Checkouts waned near the end of the season as op erations came to a close in December. When operations ended for the year, 79,701 total checkouts had been recorded at stations in central downtown.

PAGE 139

124 Figure IV. 15 Cumulative Denver B cycle Checkouts in Central Downtown, 2010 The weekly rate of Denver B cycle checkouts in central downtown changed across the operating season. During full weeks of operation, the highest weekly frequency of checkouts in central downtown (4,074) occurred during the week ending September 11, and the lowest (628) during the week ending November 27. The average number of checkouts in central downtown per week was 2,344, excluding partial weeks at each end of the season. Figure IV. 16 shows weekly total checkouts in central downtown. 8# ;87888# <87888# =87888# >87888# ?87888# @87888# A87888# d87888# ^87888# F3$'(#K%&0+32$"#*,#K&,$)'(#G36,$36,#

PAGE 140

125 Figure IV. 16 Number of Denver B cycle Checkouts Per Week from Stations in Central Downtown. Of the 40 Denver B cycle stations in central downtown, eight opened a few months or more after the system opened, thus only operated for part of the season. The remaining 32 stations in central downtown operated the duration of the season, from late April until early December. Table IV. 4 shows the year end totals of checkouts at each station in central downtown. 8# ?88# ;7888# ;7?88# <7888# <7?88# =7888# =7?88# >7888# >7?88# K%&0+32$"#*,#K&,$)'(#G36,$36,7#15#E&&+#

PAGE 141

126 Table IV. 4 2010 Denver B cycle Checkouts Per Station in Central Downtown Rank Station name Checkouts Rank Station name Checkouts 1 Market Street Station 5,580 21 15th & Delgany 2,378 2 REI 4,985 22 11th & Broadway 2,339 3 16th & Little Raven 4,833 23 Park Ave West & Tremont 2,126 4 19th & Pearl 4,734 24 Webb Building 1,962 5 16th & Boulder 4,322 25 28th & Larimer 1,912 6 13th & Pearl 4,285 26 19th & Wynkoop 1,912 7 22nd & Market 4,088 27  7th & Grant 1,639 8 Union Station 3,960 28 15th & Cleveland 1,539 9 *1350 Larimer 3,639 29 Denver Health 1,330 10 14th & Champa 3,539 30 Five Points 1,081 11 17th & Curtis 3,243 31 25th & Lawrence 942 12 16th & Platte 3,231 32 15th & Tremont 862 13 14th & Welton 3,150 33 Pepsi Center 815 14 16th & Broadway 3,097 34  12th & Sherman 576 15 1550 Glenarm 3,013 35 1450 Wazee 247 16 9th & Downing 3,008 36 2045 Franklin 144 17 *Denver Public Library 2,993 37 18th & California 143 18 17th & Larimer 2,630 38 4th & Walnut 123 19 19th & Market 2,619 39 16th & Sherman 84 20 Broadway & Walnut 2,379 40 14th & Elati 23 Stations installed within one week of system opening.   Stations installed within four months of system opening. Stations installed less than two months before the end of the season. The Denver B cycle system performed differently across geographic neighborhood service areas. In high density residential areas where car use is less practical, Denver B cycle appears to be attractive to users. The number of stations in each geographic ser vice area varies, from ten in the core downtown area to two in the Auraria area. As shown in Figure IV. 17, some geographic areas recorded more checkouts than others.

PAGE 142

127 Figure IV. 17 Weekly Average Checkouts Per Station by Nei ghborhood Service Areas. The Platte Valley/Lower Highlands area had the greatest number of checkouts, and is among the densest residential areas in central downtown. Conversely, Auraria is nearly devoid of residences and had the fewest checkouts. Auraria had the fewest stations, which may explain some of its lower checkout performance. However, Platte Valley/Lower Highlands had only four stations, equal to half that of Lodo, the second highest performing area, which also contains high density residence. 8# <8# >8# @8# d8# ;88# ;<8# ;>8# ;@8# ;d8# <88# ?a;a;8# @a;a;8# Aa;a;8# da;a;8# ^a;a;8# ;8a;a;8# ;;a;a;8# ;
PAGE 143

128 Membership Characteristics The quantity of annual memberships purchased and activated increased over the course of the season, tapering in the final months of the year. Short term memberships also increased and then tapered near the end of the year, but d isplayed more volatility. Figure IV. 18 shows how membership activation by type occurred over time. Total membership peaked during the week ending September 11. Annual memberships and total memberships increased until a plateau in late September, followed b y a decline in overall new membership. Figure IV. 18 Activated Denver B cycle Memberships by Type, Per Week. The addresses registered by Denver B cycle annual members were used to geocode a GIS layer, shown in Figure IV. 19. Many of the annual members are concentrated near Denver B cycle stations, notably in the Platte Valley/Lower Highlands 8# ?88# ;888# ;?88# <888# a<>a;8# ?a<>a;8# @a<>a;8# Aa<>a;8# da<>a;8# ^a<>a;8# ;8a<>a;8# ;;a<>a;8# F3$'(#!0$*H'$&-#N&.1&)"%*/"# !0$*H'$&-#C%3)$`$&).#N&.1&)"%*/"# !0$*H'$&-#!,,2'(#N&.1&)"%*/"#

PAGE 144

129 and Capitol Hill neighborhoods, both areas of high density residence. This clustering of annual members in high density neighborhoods within close range of central downtown, when viewed in conjunction with Figure IV. 17, suggests that annual members are using Denver B cycle to access central downtown. The people in neighborhoods equipped with Denver B cycle stations and who work downtown are well positioned to take part in the intervention. Base GIS layers from Denver GIS (DenverGIS 2011) Figure IV. 19 Geocoded Addresses of Denver B cycle Annual Members in Central Downtown Denver.

PAGE 145

130 Aim 2.1 : Determine de mographic characteristics of Denver B cycle users. The overall patterns of use of Denver B cycle described in Aim 2 suggest possible variations of engagement among demographic groups. To understand the impact of the intervention, it is crucial to identify the portion of the population who use Denver B cycle, and how it compares with the general population. Understanding which groups are participating in the intervention will assist in discovering groups that are underrepresented or absent. Review of literat ure in the previous chapter revealed that the makeup of bicycle commuters differs from the general population. During this initial year of operation of what is likely a very new concept to most, it is likely that the majority of the Denver B cycle populati on of users exhibit differences from the general population, as reflected in the following hypothesis: Hypothesis 2.1 : Denver B cycle users will be more likely to be male, Caucasian, more highly educated, and have a higher household income than the genera l population Demographic data summary Results in this section are from the Denver B cycle user survey dataset, and are focused on annual members unless otherwise noted. Demographic data from the survey include gender, age group, ethnicity, race, education al attainment, household income, and home ZIP Code. Health specific data from the Denver B cycle User Survey include self assessed health status, reported poor physical and mental health days in the past 30 days, and BMI weight classification derived from self reported weight and height. Additional items collected through the Denver B cycle User Survey include,

PAGE 146

131 active transportation behavior, self assessed bicycling ability, and commuter mode choice. These items, while not specifically demographic, are important when considering how Denver B cycle annual members may exhibit differences as compared to the general population. A summary of primary demographic findings from the Denver B cycle user survey dataset is presented in Table IV. 5.

PAGE 147

132 Table IV. 5 Demography and Characteristics of Denver B cycle Users Annual (n) Short term (n) Total (N) Annual (%) Short term (%) Total (%) Gender Male 253 109 362 62.6% 55.9% 60.4% Female 151 86 237 37.4% 44.1% 39.6% Total 404 195 599 100.0% 100.0% 100.0% Age group 18 to 25 20 21 41 5.0% 10.8% 6.9% 25 to 44 257 116 373 63.8% 59.5% 62.4% 45 to 64 108 55 163 26.8% 28.2% 27.3% 65 or over 18 3 21 4.5% 1.5% 3.5% Total 403 195 598 100.0% 100.0% 100.0% Ethnicity Non Hispanic 383 179 562 95.8% 92.3% 94.6% Hispanic 17 15 32 4.3% 7.7% 5.4% Total 400 194 594 100.0% 100.0% 100.0% Race White 362 167 529 90.0% 86.1% 88.8% Non White 40 27 67 10.0% 13.9% 11.2% Total 402 194 596 100.0% 100.0% 100.0% Education Less than BA 44 31 75 10.9% 15.9% 12.5% BA or higher 359 164 523 89.1% 84.1% 87.5% Total 403 195 598 100.0% 100.0% 100.0% Household Income <$75k 139 98 237 35.7% 51.0% 40.8% $75k or more 250 94 344 64.3% 49.0% 59.2% Total 389 192 581 100.0% 100.0% 100.0% Home ZIP Code In service area 278 94 372 68.8% 48.2% 62.1% Out of service area 126 101 227 31.2% 51.8% 37.9% Total 404 195 599 100.0% 100.0% 100.0% Participation in of Denver B cycle appears to be overrepresented among certain groups. All users, and especially annual members, who are more likely to habitually use Denver B cycle, skew toward being male, supporting similar findings of gender differences in bicycle commuting behavior (Garrard, Rose et al. 2007; Flusche 2009)

PAGE 148

133 Users of all types tend to concentrate on the middle of the age spectrum, predominantly in the range of 25 to 44. Annual members are even more highl y concentrated in this age group than short term users. This population differs in age from that traditionally targeted by most physical activity interventions reported in the literature. With regard to race and ethnicity, users of all types, and especial ly annual members, tended to be White and non Hispanic. Users also tended to have more education and higher household income than what might be expected compared to the general population. Higher education and income is most notable among annual members. A majority of short term users had ZIP Codes of home address outside of the Denver B cycle service area. This is logical, as tourists or occasional visitors who live outside of the service area have less direct access to the system and therefore fewer comp elling reasons to become annual members. Nevertheless nearly half of short term users reported a home ZIP Code within the Denver B cycle service area. Some of these short term users living within the service area likely tried the system to determine if it suited their needs, just as 35.6% of annual members having ZIP Codes within the service reported doing with short term subscriptions prior to buying annual membership. As might be expected, a strong majority (68.8%) of annual members reported home ZIP Cod es within the service area. Still, it is notable that nearly one third of annual members reported home ZIP codes outside of the Denver B cycle service area. Denver B cycle apparently plays a role in the commuting behavior of some annual members living outs ide service area, as 37.0% of this group reported using transit for all or part of their commute, and 15% reported commuting via transit in conjunction with Denver B cycle.

PAGE 149

134 Evaluating Gender, Ethnicity and Race Comparison of demographic characteristics o f Denver B cycle annual members versus the population of Denver reveals differences between the groups. Table IV. 6 shows a summary of demographic characteristics of survey participants compared to U.S. Census 2010 demographic data for Denver County (U.S. Census Bureau 2011) Table IV. 6 Gender, Ethnicity and Race of Denver B cycle Annual Members and Denver County Population (U.S. Census 2010) N % N % Denver B cycle Annual Members Denver County population Gender 404 100.0 Gender 600,158 100.0 Male 253 62.6 Male 300,089 50.0 Female 151 37.4 Female 300,069 50.0 Ethnicity 400 100.0 Ethnicity 600,158 100.0 Not Hispanic or Latino 383 95.8 Not Hispanic or Latino 409,193 68.2 Hispanic or Latino 17 4.3 Hispanic or Latino 190,965 31.8 Race 402 100.0 Race 600,158 100.0 White/Caucasian 362 90.0 White/Caucasian 413,696 68.9 Non Caucasian 40 10.0 Non Caucasian 186,462 31.1 Difference between Denver B cycle survey participants' and 2010 U.S. Census Denver County population' was significant p <.001. Gender : A chi square test detected a statistically significant difference with regard to gender. Denver B cycle annual members hip favored inclusion of males, at e < Z;[#f#
PAGE 150

135 Hispanic or Latinos, at e < Z;[#f#;>898^g# 7 h988;9#F%*"#.&',"#$%'$#'.3,4#G&,H&)#I ` 050(&# ',,2'(#.&.1&)"7#T*" /',*0"#',-#L'$*,3"#'),-&))&/)&"&,$&-7#',-#')&#$%&)&]3)&# ,3$#'"#&,4'4&-#15#$%&#*,$&)H&,$*3,#'"#')&#,3, ` T*"/',*0"7#"2//3)$*,4#T5/3$%&"*"#<9;9# Race : Because of exceedingly small sample sizes in some non Caucasian groups among Denver B cycle annual members non Caucasian groups were aggregated as a single group. A chi square test detected a statistically significant difference favoring inclusion of White/Caucasians among Denver B cycle annual members, at e < Z;[#f#d=9@dg# 7 h988;9#F%*"#]*,-*,4#.&',"#$%'$#,3, ` K' 20'"*',"#'),-&))&/)&"&,$&-#'.3,4# G&,H&)#I ` 050(&#',,2'(#.&.1&)"7#$%2"#*,-*0'$*,4#,3, ` K'20'"*',"#')&#,3$# &J2*H'(&,$(5#']]&0$&-#15#$%&#*,$&)H&,$*3,7#6%*0%#"2//3)$"#T5/3$%&"*"#<9;9# # Evaluating Age Group, Education al Attainment and Household Income Other dem ographic categories among Denver B cycle annual members, including age group, educational attainment, and household income were compared with American Community Survey data for Denver County. Table IV. 7 shows Denver B cycle annual members compared to figur es from the 2010 U.S. Census American Community Survey 1 Year Estimates (U.S. Census Bureau 2011; U.S. Census Bureau 2011; U.S. Census Bureau 2011) Chi square tests were used to analyze differences between Denver B cycle annual members and the Denver adult population.

PAGE 151

136 Table IV. 7 Age Group, Educational Attainment and Household Income of Denver B cycle Annual Members and Denver County Population (2010 U.S. Census American Community Survey 1 Year Estimates) Denver B cycle Annual Members N % 2010 U.S. Census American Community Survey N % Age Group* 403 100.0 18 and over 474,795 100.0 18 to 24 20 5.0 18 to 24 62,522 13.2 25 to 44 257 63.8 25 to 44 214,373 45.2 45 to 64 108 26.8 45 to 64 135,221 28.5 65 or over 18 4.5 65 or over 62,679 13.2 Educational attainment* 403 100.0 Educational attainment (age 18 and older) 474,795 100.0 High school graduate or higher 403 100 High school graduate or higher 397,003 83.6 Bachelor's degree or higher 359 89.1 Bachelor's degree or higher 182,224 38.4 Household income* 389 100.0 Household income 262,093 100.0 Less than $25,000 16 4.1 Less than $25,000 76,446 29.2 $25,000 to $34,999 13 3.3 $25,000 to $34,999 30,092 11.5 $35,000 to $49,999 40 10.3 $35,000 to $49,999 34,915 13.3 $50,000 to $74,999 70 18.0 $50,000 to $74,999 41,923 16.0 $75,000 to $99,999 55 14.1 $75,000 to $99,999 27,047 10.3 $100,000 to $149,999 89 22.9 $100,000 to $149,999 26,705 10.2 $150,000 or more 106 27.2 $150,000 or more 24,965 9.5 Difference between Denver B cycle survey participants' and 2010 U.S. Census American Community Survey was significant p <.001. Age Group : A chi square test detected a statistically significant difference between Denver B cycle annual members and the 2010 U.S. Census American Community Survey 1 Year Estimates with regard to age group, e < Z=[#f#A?9<8g# 7 h988; The 25 to 44 and 65 or over age groups contributed most to e < This means that the

PAGE 152

137 annual membership in Denver B cycle is not equally distributed across age groups, supporting Hypothesis 2.1. People aged 25 to 44 are much more likely to become annual mem bers than are other age groups, which are underrepresented in the intervention. Educational Attainment : Denver B cycle annual members differed significantly from the 2010 U.S. Census American Community Survey 1 Year Estimates with regard to educational attainment, e < Z;[#f#d@9>>g# 7 h988; Denver B cycle annual members favored inclusion of a more highly educated population 9#F%*"#.&',"#$%'$#.3)&# &-20'$&-#*,-*H*-2'("#' )&#.3)&#(*+&(5#$3#1&03.&#',,2'(#.&.1&)"#$%',#(&""#&-20'$&-# *,-*H*-2'("7#6%3#'),-&))&/)&"&,$&-#'.3,4#2"&)"9#F%*"#]*,-*,4#"2//3)$"# '00&/$',0]#T5/3$%&"*"#<9;9 Household Income : A statistically significant difference between Denver B cycle annual members and the 2010 U.S. Census American Community Survey 1 Year Estimates was detected with regard to household income, e < Z@[#f#=8>9=
PAGE 153

138 which is essential to participation in the intervention, and bicycle commuting habits, related to active transportation behaviors, were collected. Health Stat us : Results comparing Denver B cycle annual members to Denver County data from the Colorado Behavioral Risk Factor Surveillance System (BRFSS) dataset (Colorado Department of Public Health and Environment 2011 ) are shown in Table IV. 8. Chi square tests were used to analyze differences between Denver B cycle annual members and the Denver adult population. Table IV. 8 Health Characteristics of Denver B cycle Annual Members Versus the Denver County Adult Population Denver B cycle annual members Denver County population (Colorado BRFSS)   Health status* n=402 Excellent, very good, good 96.3% 82.0% Fair, poor, don't know 3.7% 18.0% Poor physical health days in past 30 days* n=401 0 days 71.8% 65.7% 1 to 7 days 25.7% 21.9% 8 or more days 2.5% 12.4% Poor mental health days in past 30 days* n=397 0 days 53.9% 63.4% 1 to 7 days 38.8% 20.2% 8 or more days 7.3% 16.4% Difference between Denver B cycle annual members' and 2010 Denver County population' was significant p <.001.   2010 Denver County, Colorado BRFSS (Colorado Department of Public Health and Environment 2011) A statistically significant difference was found between Denver B cycle annual members and the Denver County population in self assessed health status, e < Z;[#f#?;9==g#

PAGE 154

139 7 h988; More than 96% of annual members considered themselves to be of excellent, very good or good health, surpassing figures for the general Denver population. Regarding the number of days of poor physical health in the preceding 30 days, a statistically sign ificant difference was revealed between Denver B cycle annual members and the Denver County population, e < Z=[#f#=>9;@g# 7 h988; Denver B cycle annual members reported fewer days of poor physical health than the Denver County general adult population. Simil arly, a statistically significant difference between Denver B cycle annual members and the Denver County population in the number of days of poor mental health in the preceding 30 days, e < Z=[#f#A89==g# 7 h988; Denver B cycle annual members reported more day s of poor mental health in the middle range of 1 to 7 days' than were reported among the general adult population of Denver County. At the high end category of 8 or more days,' the figures among Denver B cycle users are lower than Denver County figures. Considering days of poor physical or mental health in the preceding 30 days, survey respondents had an edge over the general population with regard to physical health days, but reported experiencing poor mental health days somewhat more frequently than th e general population. Overall, Denver B cycle users appear to be more optimistic about their self assessed health status, and report somewhat better physical and mental health than the Denver County general adult population. BMI and Weight Classification : BMI calculated from self reported height and weight is used as an indicator of general physical fitness. A BMI figure was derived for each survey respondent and averaged among annual members, shown in Table IV. 9.

PAGE 155

140 The mean BMI of Denver B cycle annual membe rs is slightly below 25, the threshold score between normal and overweight classification. Table IV. 9 Calculated BMI Figures from Self Reported Height and Weight of Denver B cycle Annual Member s. n 401 Calculated BMI Mean 24.9 Minimum 17.2 Maximum 41.3 Standard Deviation 4.1 The percentages of overweight and obese individuals as classified according to BMI among Denver B cycle users, versus those of the general adult population of Denver County are shown in Table IV. 10. Denver County figures are taken from the 2010 Behavioral Risk Factor Surveillance System (BRFSS) (Centers for Disease Control and Preve ntion 2011) Table IV. 10 Percent within Weight Class Groups as Determined by BMI, Denver B cycle Annual Members Versus Denver County Population. Neither overweight nor obese (BMI <25) Overweight (BMI 30 25) Obese (BMI 30) Denver B cycle Annual members 58.1% 31.4% 10.5% Denver County population   45.7% 36.6% 17.7%   2010 Denver County, Colorado BRFSS (Cen ters for Disease Control and Prevention 2011) A chi square test detected a statistically significant difference between Denver B cycle annual members and the Denver County population with regard to BMI weight class group, e < Z<[#f#;^9@>g# 7 h988; The BMI <25 group and BMI i 30 groups contributed most to e < This means that Denver B cycle annual members include a smaller percentage of obese individuals, and a greater proportion of those who are of

PAGE 156

141 normal BMI than the general adult population of Denver Count y. Much of the difference among the weight categories occurs at the ends of the spectrum, as the group classified as overweight does not differ as substantially between the populations. When interpreting these results, it is important to remember that the demographic makeup of Denver B cycle users differs significantly from that of the general population. Keeping demographic differences between groups in mind, this finding still suggests that participation in Denver B cycle appeals not only to those withi n the population who are presently of normal BMI, but, perhaps more importantly, to those who are currently overweight and at greater risk of progressing toward obesity. In evaluation of Denver B cycle as a preventive health intervention to counter obesity a critical measure of success is the ability to engage the portion of the population at greatest risk of obesity before actually becoming obese. By this measure, Denver B cycle appears to succeed. Active Transportation Behavior : Car dependence and participation in active transportation activities are factors that affect prospects for maintaining individual health. Survey respondents were asked about their engagement in active transportation behavior. Table IV. 11 shows that Denve r B cycle users are accustomed to traveling to destinations 10 or more blocks distant under their own power.

PAGE 157

142 Table IV. 11 Active Transportation B ehavior of Denver B cycle Users. Short term Annual During the past 30 days have you walked or biked ten or more blocks to get to routine destinations, such as work, school or running errands? n 185 397 No 22.2% 6.8% Yes 77.8% 93.2% Denver B cycle annual members are more likely to engage in active transportation than short term members. A comparable figure for the active transportation behavior of the population of Denver or the state of Colorado is not available. However, the survey question was derived from a study of physical activity conducted by the New York City Department of Health, which found that 68% of respondents walked or biked to routine destinations (New York City Department of Health and Mental Hygiene 2009) New York City differs in many ways from Denver, and is known as a city with ubiquitous transit, comparatively low car ownership, and walkable, high density, mixed use developments. The find ing that Denver B cycle users participate in active transportation to a greater degree than citizens of New York City may indicate that the intentional installation of Denver B cycle stations in relatively high density, mixed use areas has successfully tar geted populations who are most able to fit use of shared bicycles into their lifestyles. Bicycling Ability : The ability to confidently ride a bicycle in mixed traffic is an essential skill for people who intend to safely ride a bicycle for transportation. Survey respondents were asked to assess their bicycling ability. Results are shown in Figure IV. 20.

PAGE 158

143 Figure IV. 20 Self Assessed Bicycling Ability of Denver B cycle Users. For inexperienced bicyclists who may be more accust omed to recreational, off street riding, contending with motor vehicles in traffic can pose a challenge. The majority of Denver B cycle users identified as having moderate or experienced bicycling skills, indicating they had at least some confidence in the ir ability to ride in traffic. Comparable figures do not exist regarding bicycling ability among the adult population in Denver. Although it is likely that not all users were completely comfortable riding in traffic or totally adherent to the rules of the road, no serious injuries were reported among Denver B cycle users during 2010. Bicycle ownership : The majority of Denver B cycle users have access to personally owned bikes. As shown in Figure IV. 2 1, about four out of five users of Denver B cycle own at least one other bike. This is notable, as it suggests that, although they have the option to use their own bicycles, users instead choose to check out shared bikes. This indicates that users may have preferences for shared bikes over personal bikes, at least for the purposes for which the shared bikes are being used. ;89@:# ^9>:# >;9d:# >@9=:# >A9@:# >>9=:# 898:# ;898:# <898:# =898:# >898:# ?898:# C%3)$`$&).#Y"&)"# !,,2'(#N&.1&)"# I&4*,,&)# N3-&)'$&# XO/&)*&,0&-#

PAGE 159

144 Figure IV. 21 Bicycle Ownership among Denver B cycle Users. Commuting Mode Share : As shown in Figure IV. 22, the use of a car as a commute mode among Denver B cycle users is considerably lower than that of the general population. Only 18.1% of Denver B cycle annual members and 31.0% of short term users reported commuting by car. By comparison to the rest o f the city, the state and nation, general population commuters are more than four times as likely to be car commuters as Denver B cycle annual members. Ad9^:# d;9?:# <;9;:# ;d9?:# 8:# ;8:# <8:# =8:# >8:# ?8:# @8:# A8:# d8:# ^8:# ;88:# !,,2'(#.&.1&)"# C%3)$`$&).#2"&)"# V6,#'#1*+&# G3#,3$#36,#'#1*+&#

PAGE 160

145 Figure IV. 22 Car and Bicycle Commuting Mode Share: Denver B cycle User s, Denver County, Colorado and the United States (U.S. Census Bureau 2011) C onversely, Denver B cycle users are considerably more likely to commute via bicycle than the general population. Although annual members surpass short term users, both in terms of high bicycle and low car commuting, short term users are still much differen t in these behaviors than the Denver County and larger populations. Although bicycle commuting in Denver has been on the rise for the past several years, and increased from 1.8% to 2.2% between 2009 and 2010 (U.S. Census Bureau 2010; U.S. Census Bureau 2011) bicycle commuter mode share among Denver B cycle users is more than ten times greater. G&,H&)#I` 050(&#!,,2'(# N&.1&)"# G&,H&)#I` 050(&#C%3)$` $&).#Y"&)"# G&,H&)# K32,$5## K3(3)'-3# Y,*$&-#C$'$&"# K')# ;d9;:# =;98:# d89>:# d?9?:# d@9=:# I*050(&# =<9@:# 898:# ?898:# @898:# A898:# d898:# ^898:# ;8898:#

PAGE 161

146 Summary: Denver B cycle Users Versus the General Population The key message from results in this section is that during the first year, Denver B cycle attracted a very select group of users; a finding that has clear implications as a public health intervention. In terms of demography, Denver B cycle annual members are different than the general population in a wide range of categories. Regarding gender, race and ethnicity, Denver B cycle annual members were found to significantly differ from the Denver population. Likewise, significant findings emerged regarding age group, educational attainment and household income between Denver B cycle annual members and the general population. Demographically, Denver B cycle annual members more frequently tend to be male, non Hispanic, Caucasian, aged 25 to 44, more highly educat ed, and with a higher household income as compared to the general population, supporting acceptance of Hypothesis 2.1. However, regardless of inequity of representation, all demographic subgroups exhibited at least some representation among the body of Den ver B cycle annual members. In the subject of health, Denver B cycle annual members also differed from the general Denver population, although less dramatically than by demography. Denver B cycle annual members maintain a more positive assessment of their health than the general population, yet findings are somewhat mixed regarding self assessed days of poor physical or mental health. Annual members are, on average, just within the normal range for calculated BMI, although the mean is close to the cutoff p oint for being overweight. As compared to the Denver general adult population, Denver B cycle annual members exhibit a higher percentage within normal body weight classification, and a lower

PAGE 162

147 percentage categorized as obese. However, the percentages classif ied as overweight do not differ as substantially between the populations. During this initial year, the group of people who chose to use Denver B cycle was by and large accustomed to engaging in active transportation. They are also confident in their bicy cling skills. About nine in ten consider themselves to be moderate or experienced bicyclists. Denver B cycle users are also considerably less likely to commute by car, while much more likely to commute via bicycle as compared to the general population. In all, the evidence strongly suggests that Denver B cycle users indeed differ from the general population. Although Denver B cycle users do not typify members of the general population, it should be noted that many of the characteristics of Denver B cycle u sers are consistent with those at the vanguard of adopting new behaviors; individuals who carry influence over the behavior of others within a population and who are deemed as innovators or early adopters (Rogers 2003) The above findings confirm that certain groups are better engaged in the intervention, while other groups have not been adequately reached. With this in mind, inroads to underrepresented groups exist, as all demographic su bgroups are recorded as having at least some level of engagement. Moreover, these less active subgroups, now identified, may potentially be reached as the system expands and/or as participation becomes more socially normalized. Further study is necessary t o identify and mitigate barriers to use and strategies for expanded engagement.

PAGE 163

148 Aim 2.2 : Investigate factors influencing Denver B cycle use The findings of Aim 2.1 pointed to the demography and characteristics of people most likely to participate in Den ver B cycle. Investigation of the factors that influence actual use of the system within this group is the next step. Multiple factors likely contribute to fluctuations in the use of a public bicycle sharing system. In depth studies of these factors have y et to be well represented in the literature and are at present only minimally understood. Examination of factors that influence use aids in the evaluation of public bicycle sharing as a population scale public health intervention by contributing to the und erstanding of motivational reasoning. In the previous section it was determined that differences in the demography of Denver B cycle annual members are statistically significant as compared to the Denver adult population, possibly influencing amount of use The physical ability to participate may also influence use, as signified by health indicators, such as self assessed health status and BMI. However, indicators of the suitability of shared bikes in the lifestyle of individuals, including ability to comm ute via shared bikes, the feasibility of car use replacement, and residential proximity to Denver B cycle stations may supersede demographic variables in importance. A majority of Denver B cycle annual members already own bicycles, so bicycle ownership sta tus is not as likely to influence overall use, but is worthy of consideration. These potentially influential lifestyle factors are weighed in the following hypothesis. Hypothesis 2.2: The number of Denver B cycle checkouts by annual members will be

PAGE 164

149 relate d to lifestyle factors indicating ability to replace car use with shared bike use Multivariate Modeling Of Number Of Checkouts By Denver B Cycle Annual Members Numerous factors potentially influence the number of checkouts logged by Denver B cycle annual members. It is important to understand some of the key factors that exert influence over how often subscribers actually use the system. The Denver B cycle user survey dataset was used to examine influencing factors on the number of checkouts of annual memb ers. A multivariate model was used to identify which factors affect the number of checkouts among Denver B cycle annual members. In the model, number of checkouts is the outcome variable. Predictor variables include items related to utilitarian transportation needs: the ability to use Denver B cycle to satisfy all or part of a commute ( commute via Denver B cycle ), the feasibility to replace car trips via Denver B cycle ( car replacement ) proximate access to stations from residence ( proximity ), an d status of bicycle ownership ( bike ownership ). Other predictor variables include the demographic variables of gender age group and household income as well as the health variables health status and BMI The variables race and ethnicity are heavily weighted toward Caucasian and non Hispanic, so were not included in the model. Ordered logistic regression was conducted to assess whether the predictor variables listed above are associated with the number of checkouts by Denver B cycle user s. The assumptions of observations being independent and independent variables being linearly related to the logit were tested and met using the omodel command in Stata (UCLA: Academic T echnology Services Statistical Consulting Group 2011) Because the variable commute via Denver B cycle appeared to have a moderately

PAGE 165

150 high correlation with numbers of checkouts, a check of multicollinearity against the dependent variable was conducted, examining the variance inflation factors (VIFs) using Stata 11. In the model, the VIF score for commute via Denver B cycle was 1.15, below the threshold for multicollinearity of 10.0. Tolerance, which ranges from 0.0 (severe multicollinearity) to 1.0 (abse nce of multicollinearity), as derived from 1/VIF for commute via Denver B cycle equaled 0.87 indicating min imal multicollinearity. However, Pearson's correlation coefficient value between number of checkouts and commute via Denver B cycle was 0.58, signifi cant at p <.001. Therefore, commute via Denver B cycle was removed from the model. A separate model was constructed using commute via Denver B cycle as the outcome variable the results of which follow the results and interpretation of the first model When all eight predictor variables are considered together, they significantly predict frequency of checkout among Denver B cycle users, $ 2 ( 8 ) = 100.21 N= 371 p <.001. The results of the model are presented in Table IV. 12. Table IV. 12 Ordered Logistic Regression Predicting Number of Denver B cycle Checkouts by Annual Members Variable Coefficient SE Odds ratio p Car replacement* 1.19 0 .15 3.28 0.000 Proximity 0.28 0.24 1. 32 0. 24 4 Bike ownership 0. 34 0. 27 0. 71 0. 199 Gender   0. 4 6 0.23 0. 63 0. 048 Age group 0.10 0.18 0.91 0.583 Household income   0.12 0.06 0.88 0.031 Health status 0.04 0.14 1. 04 0.751 BMI 0.03 0.03 0.9 7 0.257 Significant at p <.001   Significant at p <.05 Looking at of the odds ratios and p values for the predictor variables provide

PAGE 166

151 insight into some of the most compelling factors that influence the number of checkouts. Car replacement is significant at p <.001. Other significant predictor variables were g ender at p =.048 and h ousehold income at p =.031. The feasibility for annual members to replace car trips with Denver B cycle trips increases the odds of number of checkouts by 3.28 times, substantiating Hypothesis 2.2. This finding suggests that if annual users are able to choose to use Denve r B cycle rather than a car, they have an increased likelihood of doing so. The choice to replace car trips may conceivably result from a realization among annual members that many regular destinations are within reach using Denver B cycle. In light of thi s, Denver B cycle annual members are likely able to not only replace car commuting trips, but also incidental trips made during the day that might have otherwise been made by car. Augmenting the capacity for users to replace car trips with active transport ation through access to Denver B cycle is the central goal of the intervention. This evidence indicates that, as an intervention, Denver B cycle has met a design ambition to serve both as basic, utilitarian transportation and as an alternative to car use. The differences in number of checkouts between genders favored males over females, with women only 0.63 times as likely to generate higher numbers of checkouts as men. This finding is consistent with literature indicating men are more likely than women to use bicycle for transportation (Garrard, Rose et al. 2007) Household income pre dicted the number of checkouts, in a direction favoring annual members at lower income levels. The evidence suggests that annual members at the lower end of the household income scale are more likely to have higher numbers of checkouts than those at upper income levels. This finding departs from the finding

PAGE 167

152 revealed in Aim 2.1 that associates higher income levels with increased likelihood of Denver B cycle membership Taken together, these findings indicate that although individuals with higher household in come may be more likely to join Denver B cycle, individuals with lower household income may be more likely to actually use the shared bikes. A trend toward higher use among annual members with lower income is possible evidence that the cost savings benefit s of using Denver B cycle may be attractive to those with limited resources. This finding suggests that annual members at lower income levels may use Denver B cycle for daily transportation. The perception of financial benefit from use of the system can be considered a lifestyle influence, supportive of Hypothesis 2.2. The variables that do not significantly contribute to the model also reveal interesting findings. Proximity of one's residence to stations, that is, living within the Denver B cycle service a rea, slightly increases the odds ratio of number of checkouts, but is not statistically significant. This is an important finding in that it shows that the beneficial effects of the intervention are not entirely localized to those who live adjacent to stat ions, but that effects extend to populations beyond the geographic service area of the system. The finding that the proximity of residence is not a predictor of number of checkouts is not supportive of Hypothesis 2.2. In further consideration of the influ ence of proximity to stations, it is likely that annual members who are not resident to the Denver B cycle service area, yet who log higher numbers of checkouts probably spend substantial time in the service area for work or other purposes, and therefore m ay be considered as having proximate access. The Denver B cycle User Survey did not include a question about work location, so the influence of work proximity on numbers of checkouts is unknown.

PAGE 168

153 Bicycle ownership status was not a significant predictor of number of checkouts. This is an important finding, in that, as a strong majority of 78.9% of annual members owns a bicy cle, they still choose to use Denver B cycle This finding suggests that annual members may find shared bikes superior to their own bikes for the types of purposes for which they are being used. The accessibility, convenience, lack of maintenance needs, or reduced worry of theft for shared bikes as compared to personally owned bikes are possible factors of influence. The demographic variabl es age group did not significantly predict number of checkouts. In Aim 2.1 it w as found that those aged 25 to 44 had increased likelihood of participation in Denver B cycle. However, the findings of the multivariate model do not reveal any significant asso ciations for this group to be associated with higher numbers of checkouts. Taken together, these findings indicate a more even distribution of use of Denver B cycle across age groups than expected F indings indicate that number of checkouts may not be predicted by an individual s self assessed health status or BMI. Annual members with a variety of present health circumstances and across the range of weight categories are as likely to log higher numbers of checkouts as any other; a valuable discovery in evaluation of Denver B cycle as a preventive health intervention. Commute Via Denver B Cycle To further explore commuting and Denver B cycle another model was constructed, with commute via Denver B cycle as the outcome var iable Predictor variables inclu ded the feasibility to replace car trips via Denver B cycle ( car replacement ) proximate access to stations from residence ( proximity ), and status of

PAGE 169

154 bicycle ownership ( bike ownership ). D emographic variables of gender age group and household income as well as the health variables health status and BMI were also included as predictor variables. Multiple logistic regression was conducted to assess whether the predictor variables listed above are associated with commute via Denver B cycle When all e ight predictor variables are considered together, they significantly predict commute via Denver B cycle $ 2 (8 ) = 38.99 N=280, p <.001. The results of the model are presented in Table IV. 13 Table IV. 13 Multiple Logistic Regres sion Predicting Commute V ia Denver B cycle B y Annual Members. Variable Coefficient SE Odds ratio p Car replacement* 1.14 0.23 3.15 0.000 Proximity 0.20 0.37 0.82 0. 58 5 Bike ownership 0. 33 0.42 0.72 0.42 1 Gender 0.00 0.35 1.00 0.995 Age group 0.21 0.30 0.81 0.484 Household income 0.09 0.09 0.91 0.320 Health status 0.13 0.21 0.62 0.536 BMI   0.10 0.05 0.91 0.047 Significant at p <.001   Significant at p <.05 As in the previous model to predict number of che ckouts, the ability to replace car use was a significant predictor of commuting via Denver B cycle at p <.001 The only other significant predictor variable in the model was BMI, at p =.047. The ability to replace all or part of a car commute with Denver B cycle suggests that annual members who are able to do so have integrated the use of shared bikes into daily routines. Commuting via shared bikes with some reduction in car use imples a degree of lifestyle fit in which the use of Denver B cycle is embraced as an active

PAGE 170

155 transportation behavior supporting acceptance of Hypothesis 2.2. This also shows that the ability to use Denver B cycle to serve a utilitarian transportation purpose is a strongly compelling motivating factor for use. For annual members, having a Denver B cycle station at both ends of a commute or at a transit stop and a destination may greatly affect number of checkouts. This finding supports literature that associates the ease of active transportation integration into regular ro utines with the self efficacy of changed behavior (Berrigan, Troiano et al. 2001) The finding of BMI as a significant predictor variable suggests that there may be a threshold fitness level amo ng annual members to commute via Denver B cycle. However, as determined in the previous model, BMI is not a significant predictor of overall use. Therefore, annual members with higher BMI are not significantly less likely to use Denver B cycle in general, yet differences in BMI contribute significantly to commuting behavior in the use of shared bikes. All other predictor variables were not significant. The lack of significance of proximity of residence suggests that some annual members commute from outside the service area, using Denver B cycle as a connector to other modes. This underscores the importance of developing a redefinition of proximity to include proximity of place of employment, as found during examination of the previous model. Interviews with annual members who commute using combinations of transit, car, walking and Denver B cycle confirmed that some users who live far from the Denver B cycle service area commute, in part, on shared bikes. Details of the interview findings are presented later in this chapter. An other non significant predictor variable was gender. It is important to note that,

PAGE 171

156 although literature suggests a male predominance for transportation bicycling (Dill 2009) women comm uters among Denver B cycle annual members, are equally as likely as men to use shared bikes as part of a commute This finding is substantial, in that women are referred to as an indicator species with regard to bicycle use (Garrard, Rose et al. 2007) When women feel comfortable to use bicycles in a specific setting, it is postulat ed that many other groups of potential transportation bicyclists may also feel comfortable. The apparent gender parity among Denver B cycle annual members who commute via shared bikes suggests conditions of comfort and s afety may be acceptable or impro ving in the Denver B cycle service area, a finding that suggests benefits to all bicyclists or prospective bicyclists who ride in the area. # Summary Of Factors I nfluencing Use The number of checkouts logged by Denver B cycle annual members is influenced by the capacity of the system to fulfill their utilitarian transportation needs. The ability to commute via Denver B cycle or replace car trips with shared bike trips is significantly associated with elevated numbers of checkouts. In addition, there is a tendenc y for annual members toward the lower end of the spectrum of household income to log more trips than at the higher end, suggesting an attraction to use through cost savings over other mode choices. Proximity of residence did not significantly predict the n umber of checkouts, but it is possible that a redefinition of proximity reflecting the proximity of access to stations may alter the findings. In additio n, bike ownership did not predict number of checkouts, suggesting that shared bikes have appeal even to individuals who have the option to use their own bicycle. Taken in total, the findings support acceptance of Hypothesis 2.2.

PAGE 172

157 Among Denver B cycle annual members, the ability to commute in whole or in part via shared bikes is influenced by the ability to r eplace car use, as well as BMI score. Other demographic variables included in the model d o not appear to be significant. Among the non significant variables, it is important to draw attention to the finding that both genders are equally likely to commute u sing Denver B cycle, which differs from the expected gender split among transportation bicyclists. Although evaluation in Aim 2.1 found demographic differences among groups with regard to participation in Denver B cycle, they do not appear to strongly co incide with actual use of the system as indicated by number of checkouts or commuting via Denver B cycle Importalntly, women are as likely as men to use Denver B cycle for commuting. T &'($%#"$'$2"#',-#INU#')&#,3$#/)&-*0$*H]#*,0)&'"&-#2"& 7#'($%324%# (36& )#INU#*"#"*4,*]*0',$(5#'""30*'$&-#6*$%#03..2$*,4 9#F%&"&#]*,-*,4"#*,-*0'$&#$%'$7# *,#4&,&)'(7#$%&#-&.34)'/%5#3]#2"&#*"#.3)&#&H&,(5#-*"$)*12$&-#$%',#.*4%$#1&# &O/&0$&-#6%&,#03,"*-&)*,4#-*]]&)&,0&"#*,#$%&#-&.34)'/%5#3]#/')$*0*/'$*3,7#6%*0%#*"# ',#*./3)$',$#32$03 .&#*,#$%&#&H'(2'$*3,#3]#G&,H&)#I ` 050(&#'"#'#%&'($%#*,$&)H&,$*3,# $')4&$*,4#$%&,&)'(#/3/2('$*3,9 #

PAGE 173

158 Aim 2.3 : Determine impacts of Denver B cycle on shifts toward active transportation. In Aim 2.2, data revealed that active transportation behavior, as indica ted by the ability to use shared bikes to replace car trips, significantly predicted number of checkouts by annual members. A central purpose of this study is to evaluate the effects of Denver B cycle as a public health intervention to affect active transp ortation. A finding of a net increase in the active transportation activities of users would indicate that the intervention has induced positive reactive behavior. The objective of Aim 2.3 is to investigate the extent to which participation in Denver B cyc le shifts travel mode choice away from car use in favor of shared bicycles among annual members, using data collected through the Denver B cycle User Survey. Previous sections have detailed physical and social factors affecting the uptake of increased acti ve transportation behavior among Denver B cycle users. Some of these factors are intentionally designed into the system. Others are the result of emergent combinations of an array of influences and actions as part of a complex adaptive system. The followin g hypothesis posits that, as a result of these various factors, participation in Denver B cycle positively affects active transportation behavior of annual members. Hypothesis 2.3 : Denver B cycle annual members will shift mode choice away from car use and toward active transportation via shared bicycles Some trips on Denver B cycle bikes replaced trips that would otherwise have been made by car. Denver B cycle User Survey data regarding transportation mode shift was used to develop a car trip replacement m ultiplier to estimate the percentage of

PAGE 174

159 Denver B cycle trips that replace car trips. The multiplier was then applied to individual usage data to assess net increase in active transportation as a direct result of Denver B cycle use. Participants in the Denv er B cycle User Survey were asked to estimate the frequency of trips they make using Denver B cycle that replace car trips. Figure IV. 23 shows a summary of responses. Figure IV. 23 Annual Member Responses for Frequency of Denver B cycle Trips that Replace Car Trips. The five point response scale for the question, included Never' (n 1 ), Rarely' (n 2 ), Sometimes' (n 3 ), Most of the time' (n 4 ), and Always' (n 5 ). Two weighted car tri p replacement multiplier s, M CTR Version 1 and M CTR Version 2, were derived from survey responses M CTR Version 1 assumed a midpoint value of .50 and equal value distribution of responses using the formula: M CTR Version 1 = [(n 1 *0.00)+(n 2 *0.25)+(n 3 *0.50)+ (n 4 *0.75)+(n 5 *1.00)]/N As it may be overly optimistic to equate the middle value term "sometimes" with 50% car trip replacement, a more conservative formula was developed for M CTR Version 2 setting the midpoint n 3 value at .30, with values for n 2 (.15) an d n 4 (.65) set at h alfway between adjacent values: <=# @;# <;># A^# ;># 8# ?8# ;88# ;?8# <88# [# !(6'5"#Z,?[#

PAGE 175

160 M CTR Version 2 = [(n 1 *0.00)+(n 2 *0.15)+(n 3 *0.30)+(n 4 *0.65)+(n 5 *1.00)]/N M CTR Version 1 returned a value of 0.500, or an estimate of 50.0% of Denver B cycle trips among annual members replacing car trips. M CTR Version 2 returned a value of 0.355, or an estimated 35.5% of trips among annual members replacing car use. Summary Of The Impact Of Denver B Cycle On Active Transportation On average, the use of Denver B cycle results in a reduction of car use among annual members. Denver B cycle User Survey data revealed through the development of the car trip replacement multiplier, that an estimated 35.5% to 50 .0 % of trips on shared bicycles replace trips that would have otherwise been made by car. This result is e vidence that use of Denver B cycle has resulted in a transportation mode shift away from car use and toward the active transportation behavior of shared bike use, supporting acceptance of Hypothesis 2.3.

PAGE 176

161 Aim 2.4 : Identify active transportation benefits fo r Denver B cycle annual members Active Transportation Benefits The findings of Aim 2.3 confirmed that use of Denver B cycle results in a shift away from car use, toward active transportation via shared bikes. The car trip replacement multiplier indicated that, among annual members, an estimated 35.5% to 50 .0 % of Denver B cycle trips replace car trips. Trips that would have otherwise been made by car but were instead made using Denver B cycle equate to an increase in active transportation behavior regardle ss of any existing level of active transportation of individuals. Therefore, any shift from car use to shared bicycle use via Denver B cycle equals a net increase in the total quantity of active transportation as framed in the following hypothesis. Hypothesis 2.4: Denver B cycle annual members will exhibit a net increase in quantity of active transportation Although most of the bikes in the Denver B cycle system did not have GPS units installed during 2010, B cycle equipped a group of seven test bik es with GPS units and put them into the fleet of shared bikes to randomly collect GPS data, including speed during checkout. The aggregated GPS data logged by test bikes revealed an average speed of 7.7 miles per hour during checkout. The duration of each checkout among annual members averaged 15.74 minutes. An estimated average distance of 2.02 miles per checkout was calculated from the average duration of checkout multiplied by the average speed during checkout. Year end average totals and estimates of De nver B cycle use among annual

PAGE 177

162 members are shown in Table IV. 13. The average frequencies of use per Denver B cycle annual member per week of use are listed in Table IV. 14. Table IV. 14 Year end Average Totals and Estimates of Denver B cycle Use Per Annual Member Average total checkouts Average total minutes of checkout time Estimated average total miles ridden Average total weeks of use 39.1 615.4 79.0 10.2 Table IV. 15 Average Frequencies of Use Per Denver B cycl e Annual Member Per Week of Use. Average weekly checkouts Average weekly minutes of checkout time Estimated average weekly miles ridden 3.8 60.3 7.7 The amount of recommended weekly physical activity for adults totals 150 minutes, or 30 minutes per day, five days per week (Haskell, Lee et al. 2007) However, a lower recommendation for weekly minutes was recently found to be beneficial to even high risk individuals. This second recomme nded standard totals of 90 minutes per week or 15 minutes per day of moderate physical activity (Wen, Wai et al. 2011) While it was not possible to discern the amount of checkout time spent in motion, data from interviews and observation indicate that annual members firml y understand the usage fee policies of the system and are keen to keep the bikes in motion during checkout. Therefore, for evaluation purposes, checkout time is taken as an approximation of time spent in active transportation

PAGE 178

163 As shown in Figure IV. 24, Den ver B cycle annual members' average of 60.3 minutes worth of checkout time per week of use, is equal to 40.2% of the recommended amount of physical activity time. Compared to the alternative weekly recommendation of 90 minutes of physical activity, the ave rage of 60.3 minutes of weekly checkout time amounts to 67% of the recommendation. As a portion of either recommendation level, the contribution of Denver B cycle checkout time toward the total recommendation is sizeable Figure IV. 24 Percentage of Recommended Weekly Physical Activity Met by Average Weekly Total Denver B cycle Minutes of Checkout: 90 Minute Recommendation Versus 150 Minute Recommendations. Perhaps more important than total time of checkout, among annual mem bers, an estimated 35.5% to 50.0% of trips on Denver B cycle bikes replace car trips. The results of application of the car trip replacement multiplier to totals of individual use of Denver B cycle are presented in Tables 4.15 and 4.16, which show year end totals and weekly frequencies of the average net increases in indicators of active transportation behavior attributable to use of Denver B cycle. >89<:# @A98:# ?^9d:# ==98:# 8:# ;8:# <8:# =8:# >8:# ?8:# @8:# A8:# d8:# ^8:# ;88:# ;?8`.*,2$&#)&03..&,-'$*3,# ^8`.*,2$&#)&03..&,-'$*3,# B&)0&,$'4&#.&$#15#$3$'(#6&&+(5#.*,2$&"#3]#G&,H&)#I`050(�%&0+32$# Q&.'*,-&)#

PAGE 179

164 Table IV. 16 Average Net Year end Indicators of Active Transportation Behavior Per Annual Member Attributable to Denver B cycle Us e. Average total checkouts replacing car trips Average total minutes of checkout time replacing car trips Estimated average total miles ridden replacing car miles 13.9 to 19.6 218.5 to 307.7 28.0 to 39. 5 Table IV. 17 Average Indicators of Active Transportation Behavior Per Annual Member Per Week of Use Att ributable to Denver B cycle Use. Average weekly checkouts replacing car trips Average weekly minutes of checkout time re placing car trips Estimated average weekly miles ridden replacing car miles 1.4 to 1.9 21.4 to 30.1 2.7 to 3.9 When the car trip replacement multiplier is applied to the totals of individual use of Denver B cycle annual members, the results indicate a net increase in active transportation behavior, equating to an estimated average net increase in quantity of active transportation of 21.4 to 30.1 minutes of checkout time per week of use. As shown in Figure IV. 25, when considered on its own, the upper end of the estimated average of 30.1 minutes of weekly net active transportation increase per user amounts to 20% of t he 150 minute weekly recommendation of physical activity, or 33% at the 90 minute weekly recommendation level.

PAGE 180

165 Figure IV. 25 Percentage of Recommended Weekly Physical Activity Met by Average Weekly Minutes of Denver B cycle Checkout Time Replacing Car Trips: 90 Minute Recommendation Versus 150 Minute Recommendations. Regardless of comparison with either the 90 minute or the 150 minute recommended weekly physical activity objectives, the key finding is that Denver B cycle us e results in a net increase in active transportation among annual members. This finding is of dual benefit; the use of shared bicycles to replace use of cars simultaneously increases total active transportation time and reduces sedentary time spent in a ca r. The net increase constitutes improvement over any existing level of active transportation of individuals, and signals achievement of the goal of the intervention to increase active transportation behavior within the population, supporting Hypothesis 2.4 Sustaining the increase in activity : The ability to sustain increased active transportation is an indicator of changed behavior. Users of Denver B cycle who became annual members selected to provide themselves with ongoing access to the system, and as a voluntary intervention, each individu al user decided the dosage that he or she would receive, measured as weekly number of checkouts. Figure IV. 26 shows the average weekly number of checkouts per active annual member. <89;:# ==9>:# A^9^:# @@9@:# 8:# ;8:# <8:# =8:# >8:# ?8:# @8:# A8:# d8:# ^8:# ;88:# ;?8`.*,2$&#)&03..&,-'$*3,# ^8`.*,2$&#)&03..&,-'$*3,# B&)0&,$'4&#.&$#15#6&&+(5#.*,2$&"#3]#G&,H&)#I`050(�%&0+32$#$*.&#)&/('0*,4#0')# $)*/"# Q&.'*,-&)#

PAGE 181

166 Figure IV. 26 Average Weekly Number of Checkouts Per Active Denver B cycle Annual Member. From opening day until early September, the average weekly number of checkouts per active annual member grew from approximately 4 checkouts to slightly more than 5 checkouts. The average number of checkouts per active annual user per week was above 4 for 76% of the weeks in the season, suggesting that annual members established and maintained a consistent level of use of the system. The net increase in active transportation with demonstrated adoption and maintenance of active transportation behavior supports acceptance of Hypothesis 2.4. The number of weekly checkouts per active annual member maintained stability until late in the season. During interviews, select participants we re probed to understand reasoning behind the drop in use. Some said that they did not like to ride when the weather became cold, although a few remarked that they were not troubled by colder weather. However, several of the select participants noted that t he earlier darkness that arrives with autumn and is amplified when daylight savings time ends, is a larger factor 8988# ;988# <988# =988# >988# ?988# @988# >a<>a;8# ?ada;8# ?a<a;8# daa;8# !H&)'4&#K%&0+32$"#B&)#!0$*H&#!,,2'(#N&.1&)#

PAGE 182

167 for reduced use of Denver B cycle than cold weather alone. Concerns about riding in the dark seemed to be summed up by a user stating, Now th is time of year that it's getting darker earlier, okay...if it's dark, are people going to be able to see me?" In 2010, daylight savings time ended on November 7, a date after which the weekly number of checkouts in Figure IV. 26 appears to dramatically dec line. Summary Of Active Transportation Benefits For Denver B Cycle Annual Members Denver B cycle annual members logged an estimated weekly 60.3 minutes of checkout time and a net increase in active transportation time of an estimated 21.4 to 30.1 minutes. This activity was sustained, on average, for more than ten weeks. This increase is resultant from a direct replacement of car trips with shared bicycle trips, and therefore a net increase in active transportation behavior, regardless of the prior active tr ansportation behavior. This finding supports that, on average, a percentage of Denver B cycle use concurrently increases time spent in active transportation while reducing time spent in a car, both factors that can lead to reduction in health risks. The co nclusion that Denver B cycle positively affects active transportation behavior supports acceptance of Hypothesis 2.4.

PAGE 183

168 Aim 3 : Provide an in depth description of the impacts of Denver B cycle as a public health intervention The findings of the previous Ai ms suggest that the physical and social elements of the built environment, the demographic makeup of the Denver B cycle user groups, and factors that influence the use of the system all contribute to a shi ft toward active transportation The intervention h as been found to positively affect active transportation behavior among annual members. However, the actions of the participants are an important mechanism within the intervention itself. The intervention counts on users to develop their own motivation to initiate and maintain use of the system, thus increasing their active transportation behavior. The system is also dependent on users to spread use of the system through the population via social interaction. Examination of active transportation behaviors of select participants serves to illuminate motivation for active transportation behavior and to show how their social interactions lead to (,//1",.+$./$1"*$5,!',+$".&,)4$+*!5.20" 9 # Motivation for Active Transportation Behavior As a self applied public health intervention, Denver B cycle is dependent on internally derived motivation of its users to initiate and continue their own behavioral action. Quantitative data in the preceding section revealed that use of Denver B cycle inc reased net active transportation behavior. Interviews of select participants expose d how perceived benefits contributed to motivation toward this behavior. All of the select participants talked about benefits or outcomes that they personally experienced, o f which they attributed, in part or in whole, to their use of Denver B cycle. The two main benefit

PAGE 184

169 constructs revealed through the interviews were health outcomes and economic outcomes Also important for establishing motivation for continued use were /)&! .2"$,+/41*+&,+3$ !'*$,+&.27.2)!,.+$./$%*')8,.2$&')+3* 9 # Health outcomes All participants talked about feeling better after using Denver B cycle, most frequently elaborating on how being outside, being physically active and interacting with other bicyclists w ith whom they felt a connection, contributed to improved mental disposition and energy. In addition to psychological effects, some also experienced physiological effects, including weight loss, improved endurance and increased strength. While their frequen cy of use of the system differed considerably, it was evident that all participants felt as though they benefitted from using Denver B cycle. N&,$'(#0(')*$5#',-#" $)&""#)&(*&] _#!((#"&(&0$#/')$*0*/',$"#'$$)*12$&-#/3"*$*H&# /"50%3(34*0'(#%&'($%#32$03.&"#$3#2"& #3]#G&,H&)#I ` 050(&9#C$)&""#)&(*&]#])3.#63)+#',-# '#6'5#$3#$'+&#'#1)&'+#])3.#$%]]*0&),0&/$"#$%'$#6&)&#)&/&'$&-#$%)324%32$# $%&#*,$&)H*&6"9#V,&#/')$*0*/',$#)&]&))&-#$3#$'+*,4#'#J2*0+#)*-&#'"#1&*,47#j"3)$#3]#(*+&# .5#H&)"*3,#3]#'#0*4')&$$)&'+9k#!#]& 6#/')$*0*/',$"#]302"&-#/)*.')*(5#3,#%36#$'+*,4# '#1)&'+#])3.#63)+#$3#)*-&#'#1*+&#]3)#'#]&6#.*,2$&"#3)#15#)*-*,4#*,#$3#63)+#H*'# G&,H&)#I ` 050(&#)&"2($&-#*,#*./)3H&-#32$(33+#3,#$%&#-'5#',-#*./)3H&-#63)+# /&)]3).',0&9#V,?.',# "'*#3]#%&)#03..2$&#H*'#G&,H&)#I ` 0 50(& # G &]*,*$&(5 7 #UlH&#%'-#'#032/(]#$*.&"#6%&,#1*+*,4#$3#63)+#*,#$%&#.3),*,4#',-# *$m"#,*0&#',-#"2,,5#',-#U#$%324%$#$3#.5"&(]7# jE 367#.5#(*]&#*"#/)&$$5#4)&'$#$%'$# U#0',#1*+&#$3#63)+9 k #!,-#*$l"#43)4&32"#32$#"3#*$#+*,-#3]#/2$"#532 #*,#'#1&$$&)# .33-#&H&)5#.3),*,4n # # N&,$'(#0(')*$5#',-#]302"#'"#'#)&"2($#3]#2"*,4#G&,H&)#I ` 050(&#'//&')&-#

PAGE 185

170 )&/&'$ &-(5#-2)*,4#$%&#*,$&)H*&6"9#!#/')$*0*/',$#-&"0)*1&-# $%&# /)30&""#3]#%36# $%324%$"#43*,4#$%)324%#%*"#%&'-#6%*(&#)*-*,4#'#G&,H&)#I ` 050(*+&#$3#63)+# %&(/&-# %*.#$3#'))*H)4',*b&-#',-#*,#'#])'.]#.*,-#03,-20*H&#$3#&,$&)*,4#$%?)+# &,H*)3,.&,$9# # Ul((#1&#$%*,+*,4#'132$#6%'$#Ul.#43*,4#$3#1&#-3*,4#-2)*,4#$%&#-'5#'$#63)+#',-#U# "/&,-#'#(3$#3]#$*.&#'"#Ul.#/&'( *,4#$%*,+*,4#'132$#.5#-'59#!,-#"3#*$#%&(/"#$3# 3)4',*b&9#U,"$&'-#3]#03.*,4#*,#',-#"&&*,4#'((#$%&#&.'*("#',-#o2"$#1&*,4#$3$'((5# ](2"$&)&7 #%'H*,4#'#]&6#.3.&,$"#'$#$%&4*,,*,4#3]#$%&#-'5#6%&)&#Ul.# &O&)0*"*,4#]3)0&"#532#$3#$%*,+7#1&0'2"Ȕ#-3,l$#o2"$# &O&)0*"&#',-#%'H&#'# 1(',+#.*,-n#F %&#,'$2)'(#$%*,4#$ 3#$%*,+#'132$#*"#6%'$#Ul.#43*,4#$3#-3#6%&,#U# 4&$#*,$3#63)+9#C3#-'5"#')&#'0$2'((5#3)4',*b&-#1&]3)Ȕ#&H&,#4&$#$%&)&7#U# 632(-#"'5#$%'$l"#'#H&)5#1&,&]*0*'(#&]]&0$7#*$#)&-20&"#.5#-'*(5#"$)&""#',-#*,# 4&,&)'(#o2"$#1&*,4#.3)&#%&'($%57#*$#4*H&"#532#'#1&$$&)#+*,#3]#'$$*$2-&#',-# 32$(33+9# # D2,#',-#&,o35.&,$#'"#*,*$*'(#)&'"3,"#]3)#$)5*,4#G&,H&)#I ` 050(&#'("3#"&&.#$3# &,032)'4,$*,2&-#2"&9#X,o35.&,$#*"#',3$%&)#)&02))*,4#03,0&/$#$'(+&-#'132$#15# .3"$#2"&)"9#!#"&,"]#]2,#"&&."#$3#1&#/')$#3]#'#]&&-1'0+#(33/#$%'$#+&&/" #2"&)"# )&$2),*,4#$3#2"&#$%*+&"9#!#/')$*0*/',$#"2..&-#2 /# $%&#)&('$*3,"%*/#1&$6&&,#1&*,4# '0$*H&#',-#%'H*,4#]2,#15#"'5*,47#j U#42&""# $%'$#U#'.#4&$$*,4#"3.&#)&42(')#4&,&)'(# /%5"*0'(#]*$,&""9#U#632(-#"'5#o2"$#)*-*,4#*,#4&,&)'(#.'+&"#.&#%'//5#'"#%&((9#c&$$*,4# '# ,*0&#(*$$(&#)*-&#*,#*"#]2,9 k # Some participants identified connections between psychological and physiological benefits. One participant, who admitted to being overweight and not accustomed to daily physical activi $5 7#12$#6%3#1&0'.&#',#'(.3"$#-'*(5#G&,H&) #I ` 050(&# )*-&)7 # )&](&0$&-#3, # %*"#%&'($%#0%',4&"_ # U#$%*,+#UlH&#]&($#1&$$&)9#U#6'"#4&$$*,4#'#(*$$(& # &O&)0*"&9#!($%324%#3,(5#]*H&#$3# "&H&,#.*,2$&"#&'0%#$)*/7#*]#U#-*-#*$#$6*0&#'#-'57#$%)&)#]32)#-'5"#'#6&&+7#U#]&($# 1&$$&)7#/%5"*0'((5#',-#.&,$'((59#D3)#$%&#P 2(5#>$%#%3(*-'57#.5#,&*4%13)%33-# %'"#'#/')'-&7#',-#.5#(*$$(*)(#"'$# *,#$%'43,n #N5#6*]&#/2((&-#*$#-36,#$%&# %*((#',-#$%&,#6&,$#')32,-#$%),&)#',-#U#%'-#$3#43#2/#$%&#%*((9#U#/2((&-# $%'43,9#!,-#"%&#"'*-7#jU#0',#$&((#532mH&&,#)*-*,4#$%*+& 0'2"&#

PAGE 186

171 532m)&#,3$#&H&,#1)&'$%*,4#%')-9k# n #U#)*-&#%')-7#1&0'2"&#U#6',$#$3#4&$#$%&)&# J2*0+(59#U#-3,m$#6',$#$3#-'6-(&9 # # Although nearly all participants agreed that use of Denver B cycle constituted some degree of physical activity, through their remarks all seemed to emphasize the psychological benefits that they experienced. For most, the references to physical and mental benefits were interrelated and discussion of these joint outcomes were a recurring theme throughout the interviews. One woman, who was an occasional user, was inspired to begin riding her own bicycle more because of her experience with Denver B cycle. Of health impact, she said: # UmH&#,3$*0&-#Ul.#'1(&#$3#4&$#/('0&"#]'"$&)7#)*-&#%*(("#.3)*$%#&'"&7#'"#3/ /3"&-# $3#)&'((5#63)+*,4#'$#*$9# C37#UmH�&)$'*,(5#*./)3H&-#.5#&,-2)',0&7# 6%*0%#%'"# 1&&,#'#433-#1&,&]*$9# !,-#$%&,#.&,$'(#%&'($% ` 6*"&7#U#$%*,+#o2"$#1&*,4#'1 (&#$3# &,o35#$%&#])&"%#'*)#',-#$)5*,4#$3#.'+&#$ %'$#'#/')$#3]#.5#(*]&#(*+&#$%'$9 #Um.#,3$# o2"$#43*,4#$3#'#.&&$*,47#12$#Um.#'0$2'((5#&,o35*,4#1&*,4#32$33)"#6%*(&#Um.#*,# $%&#/)30&""9# S32#+,367#$%'$#%'"#1&&,#'#63,-&)]2(#1&,&]*$9#U$#0&)$'*,(5#%'"# 1&&,#433-#]3)#. 5#.&,$'(#%&'($%9 # # B%5"*3(34*0'(#&]]&0$" _#C3.]#$%&#/')$*0*/',$"#,3$&-#$%'$#$%&5#%'-# &O/&)*&,0&-#6&*4%$#(3""#"*,0"*,4#G&,H&)#I ` 050(&9#N3"$#3]#$%&"&#/')$*0*/',$"#"'*-# $%'$7#'($%324%#$%&*)#2"]#G&,H&)#I ` 050('"#/)31'1(5#,3$#$%&#"3(�'2"]#$%&# 6&*4 %$#(3""7#$%&5#-*-#&J2'$&#*$"#2"&#'"#/')$#3]#',#'-3/$&-#%&'($%*&)#(*]&"$5(&7#3]#6%*0%# "%')&-#1*+&"#/('5&-#'#)3(&9#C3.&#$'(+&-#'132$#%'H*,4#'00&""#$3#%&'($%*&)#]33-# 3/$*3,"#1&0'2"]#G&,H&)#I ` 050(&9#V,&#)&42(')#2"&)#)&/3)$&-#])&J2&,$(5#2"*,4# G&,H&)#I ` 050(&# $3#)*-&#$3#'#4)30&)5#"$3)&#$3#125#%&'($%5#]33-#]3)#%*."&(]#',-#%*"#03 ` 63)+&)"7#6%&)&'"#/)&H*32"(5#$%&5#.*4%$#%'H�%3"&,#$3#&'$#]'"$#]33-#'H'*('1(&# 6*$%*,# 6'(+*,4#-*"$',0&9# T&#"'*-_ # D)3.#'#-*&$')5#"$',-/3*,$7#pG&,H&)#I ` 050 (&q#'($&)&-#6%'$m"#'H'*('1(&#$3#53 29# pN5#03 ` 63)+&)"q#')& #$%&,&]'0$3) "#3]#.&#)*-*,4#I ` 0 5 0(&7#1&0'2"&#*]#U#43#$3# W*,4#C33 /&) "7#$%& ,#U#0',#/*0+#2/#"3.&$%*, 4#]3)#$%&.#'"#6&((9#U$m"#1&$$&)9#U$m"#

PAGE 187

172 %&'($%*&)9 #F%&5#%'H&#'#(3$#1&$$&)#]33-"#]3)#.&9#U#0',#.'+&#*$#$3#W*,4# C33/&)" # ',-#4&$#])2*$7#H&4& $'1(&"7#',-#,*0&#(&',#.&'$"9#V)#*]#U#-3,m$7#$%&,#U#2"2'((5# 3)-&)#N0G3,'(-m"9# # # C&H&)'(#3]#$%3"&#*,$&)H*&6& -#"'*-#$%'$#1&0'2"]#G&,H&)#I ` 050(&7#$%&5#,36# $%324%$#.3)&#'132$#%36#$3#4&$#$3#-&"$*,'$*3,"7#',-#-*-#,3$#'(6'5"#03,0(2-&#$%'$#'# 0')#*"#$%&"$#.3-& #]3)#',5#4*H&,#$)*/9#C3.]#$%&#"&(&0$#/')$*0*/',$"#6%3#%'-# 1&03.&#)&42(')#2"&)"#',-#6%3#6%&)&#*,$&)H*&6&-#]3((36*,4#$%�(3"]#$%&#"5"$&.# *,#G&0&.1&)#"'*-#$%'$#$%&5#&O/&)*&,0&-#"3.&*4%$#4'*,#*,#$%&#*,$&)H&,*,4#.3,$%"# $%'$#032(-#1&#'0032,$&-#]3)7#*,# /')$7#'"#'#)&"2($#3]#,3$#%'H*,4#G&,H&)#I ` 050(&# 'H'*('1(&9 # B&)%'/"#$%&#.3"$#)&.')+'1(&#%&'($%#32$03.&"#])3.#$%&#*,$&)H*&6"#3002))&-# 6*$%#'#032/(&9#X'0%#3]#$63#"&(&0$#/')$*0*/',$"#6&)&#*,-&/&,-&,$(5#*-&,$*]*&-#',-# )&0)2*$&-7#6*$%32$#',5#'6')&,&""#3,#$%&#/')$ #3]#$%&#)&"&')0%&)#$%'$#$%&#$63#6&)&#*,# '#)&('$*3,"%*/#6*$%#&'0%#3$%&)9#I3$%#3]#$%&#*,-*H*-2'("#*,H3(H&-#6&)&#/)&H*32"(5#,3$# %*4%(5#/%5"*0'((5#'0$*H&9#!]$&)#$'(+*,4#6*$%#13$%#3]#$%&.7#$%&5#)&H&'(&-#%36#2"]# G&,H&)#I ` 050(�',#']]&0$#'#%32"&%3(-9# # U,#$%&# 032/(&7#'#.',7#6%3#6*((#1&#)&]&))&-#$3#15#$%&#/"&2-3,5.#I')$7#6%3# 6'"#3H&)6&*4%$#',-#=^#5&')"#3(-7#/)&H*32"(5#%'-#'#]'*)(5#"&-&,$')5#(*]&"$5(&#',-#-*-# ,3$#)*-&#'#1*050(&#]3)#$)',"/3)$'$*3,9#I')$l"#(*H& ` *,#4*)(])*&,-7#6%3#6*((#1&#)&]&))&-#$3# 15#$%&#/"&2-3,5 .#T&*-*7#'4&#>=7#%'"#"*.*(')#%&'($%#0%')'0$&)*"$*0"9#T&*-*#*,$)3-20&-# I')$#$3#G&,H&)#I ` 050(&#"%3)$(5#']$&)#*$#3/&,&-7#',-#$%&5#13$%#1&0'.&#)&42(')#2"&)"9# !]$&)#-&0*-*,4#$3#2"&#G&,H&)#I ` 050(&#'(.3"$#&O0(2"*H&(5#]3)#%*"#$)',"/3)$'$*3,#,&&-"7# I')$#)&/3)$&-#'# (3""#3]#
PAGE 188

173 F%& #032/(&#(*H&"#*,#'#03,-3#,&')#'#G&,H&)#I ` 050(&#"$'$*3,#*,#$%&#L3-3# ,&*4%13)%33-7#',-#&H&,$2'((5#"3(-#$%&*)#3,(5#0')9#F%&5#0*$&-#.2($*/(&#]'0$3)" #$%'$# 2($*.'$&(5#&,'1(&-#$%&.#$3#"&((#$%&*)#0')7#12$#"'*-#%'H*,4#'00&""#$3#G&,H&)#I ` 050(&# 6'"#$%�)20*'(#(*,+#/)3H*-*,4#/&)"3,'(#.31*(*$57#.'+*,4#$%&#$)',"*$*3,#/3""*1(&9# I')$#%'-#$%*"#$3#"'5#'132$#$%&#%&'($%#0%',4&"#%&# &O/&)*&,0&-_ # U#(3"$#(*+&# $6&,$5#]*H &#/32,-"#])3.#$%&#&,-#3]#N'5#2,$*(#(*+&#"'5#C&/$&.1&)#3)# V0$31&)7#',-#U#-*-,l$#)&'((5#&O&)0*"&7#U#o2"$7#532#+,367#6'(+&-#',-#1*+&-# &H&)5 6%&)&9#C3#$%'$#6'"#)&'((5#433-n# C37#U#0',#2,"0*&,$*]*0'((5#"3)$#3]#'((2-&# $3#$%*+*,4#1&*,4#',#*./3)$',$#/')$#3]#6&*4% $#(3""9#F%&)&l"#$%&*4%$#(3" "7# 0')-*3H'"02(')#%'"#*./)3H&-n# C3.&$*.&"#U#4&$#3,#$%*+&7#*$l"#(*+&7#jE367k# 532#+,367#Ul((#43#2/#%*(("#*,#$%*)-#4&')#',-#1&#(*+&7#jF%*"#*"#&'"59k# # U $#%'"#*./)3H&-#.5#6*((*,4,&""#$3#6'(+#',-#1*+&#'("39#L*+&#U#-3,l$#03,"*-&)# *$# "3)$#3]#'#%'""(&9#E%&)&'"#1&]3)&7#4&$$*,4#&O&)0*"'",l$# r #*$#6'"#'(6'5"# "3.&$%*,4#*,#$%&3)5#$%'$#U#6',$&-#$3#-37#12$#*,#/)'0$*0&7#532#+,367#Ul.#$33# $*)&-7#U#-3,l$#]&&(#(*+&#*$9#B)31'1(5#$%&"$#*"#(&4"7#'#(*$$(*$#3]#0')-*3H'"02(')# ',-#$%*((*,4, &""#$3#+&&/#&O&)0*"*,49 # # Y"]#G&,H&)#I ` 050('"#/')$#3]#'#"&)*&"#3]#0%3*0&"#$%'$#%'-#1)3'-#)',4*,4# $)',"]3).'$*3,'(#0%',4&"#3,#$% /(&9#C&((*,4#$%&*)#0')#',-#2"*,4#.2"0(&#/36&)# ]3)#.31*(*$5#6&)�&,$)'(#$3#$%�%',4&"9#L*+&#I')$7#T&*-*#'("3# &O/&)*&,0&-#%&'($%# ']]&0$"#$%'$#"%&#'$$)*12$&-7#*,#/')$7#$3#%&)#2" ]#G&,H&)#I ` 050(&9#C%&#"'*-_ # E&((7#*$m"#%')-#$3#'$$)*12$&#*$#-*)&0$(5#$3#G&,H&)#I ` K50(&7#1&0'2"&#"& H&)'(# $%*,4"#%'//&,&-#'$#3,0&9# S32#+,367#U#"$')$&-#4&$$*,4#%&'($%*&)#1&0'2"]# "2)4&)5#' ,-#)&%'1*(*$'$*3,9#U#%'-#'#%*/#*,o2)59#U$# %'-# .'-'(+*,4#',-#)*-*,4# -*]]*02($#]3)#'132$#$63#5&')"9# C3#$%'$7#',-#'$#$%&#"'.&#$*.+$#)*-#3]#32)# 0')7#',-#I ` K50(&"#0'.&#*,9#C37#5&'%7#Um.#'#(3$#%'//*&)#$3#1&#'1(&#$3#4&$#')32,-# ',-#$3#1&#'0$*H&9#Um.#.3)&# ]*$9#UmH&#(3"$#6&*4%$9#T36#.20%#3]#$%'$#632(-# %'//&,#6*$%32$#I ` K50(&7#U#-3,m$#+,369#I2$#*$m"#'("3#'#.'$$&)#3]#$)5*,4#$3# )&-20 &#$%&#'.32,$#3]#"&-&,$')5#$*.&n # F 3#.&7# pG&,H&)#I ` 050(&q#* "#'#1*4#-&'(# $%'$#)&'((5#']]&0$"#.5#(*]&9 # # I')$#',-#T&*-*#6&) $%#'6')&# 3]#$%&#)*"+"#3]#'#"&-&,$')5#(*]&"$5(&7#',-#13$%# 03,"0*32"(5#-&0*-&-#$3#.3-*]5#$%&*)#(*H&"7#',-#$%&#+&5#*,#-3*,4#"3#6'"#$3#)&-20&# $%&*)#0')#-&/&,-&,0&9#F%&5#2"&#$)',"*$#',-#-&/&,-#3,#12"#"&)H*0&#]3)#'$#(&'"$#/')$#

PAGE 189

174 3]#$%&*)#03..2$&"7#',-#1&0'.&#.&.1&)"#3]#'#0 ')#"%')*,4#"&)H*0&#]3)#300'"*3,'(#2"&9# R&H&)$%&(&""7#13$%#"$)&""&-#$%'$#6*$%32$#%'H*,4#'00&""#$3#G&,H&)#I ` 050(&7#$%&# 0%',4&"#$%'$#$%&5#6&)&#'1(&#$3#.'+ɸ(-#%'H&&,#.20%#.3)&#-*]]*02($9#X'0%# "/3+&#'132$#%36#"%')&-#1*+&"#'((36&-#$%&.#$3#.'*,$'*,#/&)"3,' (#.31*(*$5#6*$%32$# %'H*,4#$3#-&/&,-#3,#'#0')#3)#'-%&)&#$3#$)',"*$#"0%&-2(&"9 # Economic Outcomes F%&#"$3)5#3]#I')$#',-#T&*-*#03,$*,2&"#'"#',#*((2"$)'$*3,#3]#&03,3.*0#1&,&]*$"# ])3.#2"]#G&,H&)#I ` 050(&9#E%&,#$%&5#"3(-#$%&*)#0')7#$% /(&#/2$#$%&#/)30&&-"# ] )3.#$%&#"'(&#*,$3#'#"'H*,4"#'0032,$9#F%&5#03,$*,2&-#$3#)&'(*b,"*-&)'1( "$# "'H*,4"#'""30*'$&-#6*$%#$%$,&)"%*/#',-#2/+&&/#3]#$%�')9# F%&5#1&,&]*$&-#]2)$%&)# 15#)&,$*,4#32$#$%&*)#/')+*,4#"/'0&#$%'$#$%&5#,3#(3,4&)#,&&-&-9#I')$#"'*-#$%*"#'132$# $%&#&03 ,3.*0#*./(*0'$*3,"_ # C&((*,4#$%�')#4'H"#'#,*0&#(*$$(&#,&"$#&447#12$#'("37#6*$%#*$#0'.&#,3# *,"2) ',0&9#U#o2"$#0',0&((&-#$%'$#32$)*4%$#)'$%&)#$%',#+&&/n# C37#$%&)&#*"#$%'$7# ',-#3]#032)"&$#.3,&5#1'0+#])3.#)&,$*,4#32)#/')+*,4#"/'0&#$3#"3.&# ,&*4%13)"9#C 3#5&'%7#*$l"#'0$2'((5#"'H&-#2"#J2*$&#'#1*$#3]#.3,&59#U$l"#.'-"# "3.&#.3,&57#+*,-#3]7#*,'-H&)$&,$(57#12$#&H&,#*]#6&#-*-,l$#"&((#$%�')7#*$#632(-# o2"$#"'H"#.3,&5#15#,3$#%'H*,4#$3#/'5#]3)#4'"#',-#6%'$&H&)#*$#*"#/&)#4'((3,# ,367#U#-3,l$#&H&,#+,369#C3#5&'% 7#& 03,3.*0'((57#*$l"#43$$&,#2"#$3 #(*H&#&'"*(5# 6*$%*,#32)#.&'," 7 #'0$2'((59 # # !($%324%# I')$#',-#T&*-*l" #"$3)5#*"# '$5/*0'(7#*$#-3&"#/)3H*-&#',#&O'./(]#%36# 2"]#G&,H&)#I ` 050(&#)&"2($&-#*,#)&-20&-#&03,3.*0#&O/&,-*$2)&#]3)#0')# .'*,$&,',0&7#',-#"&)H&-#'"#.3$* H'$*3,#$3#2"&#'0$*H&#$)',"/3)$'$*3,9#Y"]#G&,H&)# I ` 050(&#]3)#.3"$#3]#$%$%&)#"&(&0$#/')$*0*/',$"#)&"2($&-#*,#03./')'$*H&(5#.3-&"$# &03,3.*0#1&,&]*$"7#12$#$%&"&#&03,3.*0#1&,&]*$"#03,$)*12$&-#$3#.3$*H'$*3,9#!#]&6# )&/3)$&-#"'H*,4#"3.&#.3,&5#3,#]2&(7#12$#]2 &(#"'H*,4"#6&)&#,3$#'#.'o3)#]'0$3)9#V,#

PAGE 190

175 $%$%&)#%',-7#,&')(5#'((#/')$*0*/',$"#)&/3)$&-#"'H*,4#.3,&5#)&('$&-#$3#$%"$#3]# 0')#/')+*,49# V,&#/')$*0*/',$#"'*-7#j U#-&]*,*$&(5#"'H&#*,#$%'$#Um.#,3$#/'5*,4#]3)# /')+*,49# pB')+*,4q#* "#'#%2,-)&-#',-#"*O$5# p-3((')"q # /&)#.3,$%7#"3999 # &*4%$#-3((')"#'# -'59 k # !,3$%&)#/')$*0*/',$#-&$'*(&-_# # U#632(-#1)*,4#.5#0')#-36,7#',-#$%&,#U# 632(-#%'H&#$3#/'5#]3)#/')+*,49# !,-# ,367# p1&0'2"&#U#2"&#G&,H&)#I ` 050(&q# U#-3,m$#&H&,#$%*,+#$6*0&# '132$#*$9 #U]#U#%'H&# '#.&&$*,47#U#/(',#* $#32$9#U# +,36#6%&)&#Um.#43*,49 #U#+,36#$%'$#U#0',#4&$#$%&)&#*]# *$m"#6*$ %*,#'#$%*)$5 ` .*,2$*+&#)*-&9# U#632(-#)'$%&)#$'+&#'#1*+&# )'$%&)#$%',# 1)*,4#.5#0')#-36,9# C3#*$#%'"#0&)$'*,(5 s .5#/')+*,4#03"$"#%'H&#H*)$2'((5#'((# 12$#-*.*,*"%&-#$3#,3$%*,4#"*,0&#U#"$')$&-#1*+*,49 # # C3.&#/')$*0*/',$" #'$$)*12$,"*-&)'1(&#"'H*,4"#*,#/')+*,4#03"$"#$3#$%"&# 3]#G&,H&)#I ` 050(&9#!#]&6#"'*-#$%'$#$%&5#3]$&,#/')+#,&')#G&,H&)#I ` 050(&#"$'$*3,"# (30'$&-#6%&)&#"$)&&$#/')+*,4#*"#])&&9#F%&5#$%&,#0%&0+#32$#'#1*+&#$3#)*-&#$3#$%&*)# -&"$*,'$*3,9#F%& "&#t/')+#',-#)*-&l#2"&)"#"'H&#"3.&#.3,&57#&,4'4&#*,#/%5"*0'(# '0$*H*$57#',-#&O/&)*&,0&#)*-*,4#'#1*050(&#]3)#$)',"/3)$'$*3,9#V,&#/')$*0*/',$#6%3# /')+"#$%&,#)*-&"#-'*(5#%'-# $%*"#$3#"'5 #'132$#03"$#"'H*,4" # B ')+*,4#6*((#)2,#532# -36,$36, #',56%&)&#])3.# u<88 #$3#u= 88#/&)#.3,$%#',-# $%'$l"#',#&O/&,"&#$%'$7#%'H*,4#$63#+*-"#',-#'#6*]%3#"$'5"#%3.*$%#$%&.7# *"#',#&O/&,"&#$%'$#.5#]'.*(5#)&'((5#0',,3$#12)-&,#)*4%$#,369# U, #3)-&)#$3#4&$# ')32,-#$%'$#$%,(5#$%*,4#U#0',#-3#*"#&""&,$*'((5#]*,-#])&&#/')+*,4#',-#1*+&# ])3.#$% &)& 9#C3#*$l"# )324%(5# u<88 #$3#u=88#'#.3,$%#*,#/')+*,4#',-#$%&,#$%&)&#*"# '((#$%&#$*.&#$%'$#*$#6*((#$'+&#$3#]*4%$#$%&#$)']]*0#$3#4&$#'0$2'((5#'((#$% '5# -36,$36,#',-#$3#$%&#/')+*,4# (3$"9# # # C3.&#/')$*0*/',$"#&O/&)*&,0&-#&03,3.*0#1&,&]*$"#3$%&)#$%',#$%)324%# -& 0)&'"&-#]2&(#3)#/')+*,4#03"$"9#V,&#/')$*0*/',$#$'(+&-#'132$#"'H*,4#.3,&5#15#,3$# %'H*,4#$3#125#'#.3,$%(5#$)',"*$#/'""9#C%&#(*H&"#*,#'#12*(-*,4#6*$%#'#G&,H&)#I ` 050(&# "$'$*3,#o2"$#32$"*-&#',-#03..2$&"#$3#'#"$'$*3,#,&')#%&)#3]]*0&#*,#0&,$)'(#-36,$36,9# B)&H*32" (57#"%&#$33+#$)',"*$ #$3#63)+9# #

PAGE 191

176 X H&,#$%324%#U#(*H&#$63#.*(&"#'6'5#])3.#.5#63)+7#*$l"#uA<988#]3)#'#.3,$%(5# 12"#/'""#H&)"2"#u??#]3)#p',#',,2'(# -*"032,$&-# G&,H&)#I ` 050(&#.&.1&)"%*/q9# C3#$%'$#*"#'#%24&#*./'0$#$3#"3.&3,&m"#12-4&$#$339#S&"7#"3#uA<#$*.&"#@ # p.3,$%" q 7#u >=<7#532#+,367#3H&)#u??9# F%'$l"#'#(3$#3]#.3,&59 # # !,3$%&)#/')$*0*/',$#-&$'*(&-#%36#%*"#*./)3H&-#%&'($%#)&"2($&-#*,#',# &03,3.*0#1&,& ]*$ #]3)#%*"#03./',59#!#(*]&#*,"2)',0&#/3(*05#)&'""&"".&,$#3]#%*"#%&'($%# )&"2($&-#*,#.3)&#]'H3)'1(&#$&)."9#T&#'$$)*12$&-#%* "#2"]#G&,H&)#I ` 050(&#'"#'#]'0$3)# *,#%*"#*./)3H&-#%&'($%#)'$*,49#F%&#/')$*0*/',$#&O/('*," # E&(( 7 #.5#03./',5#%'"#'#(*]&#*,"2)',0&#/3(*05#3,#.&#"3#$%'$#*]#U7#1&0'2"&#U#'.#'# /)*,0*/(&#.&.1&)#%&)&7#*]#U#$&).*,'$& 7 #$%&,#$%&5#,&&-#'#(3$#3]#.3,&5#*,#3)-&)# $3#+&&/#$% "*,&""#)2,,*,47#o2"$#$3#]*((#.5#/3"*$*3, 7 #]*,-#"3.&13-5#6%3#0',# -3#6%'$#U#-39#!,-#"3#$%&#(*]&#*,"2)',0&#/3(*05#1'"*0'((5#)&J2*)&"#$%'$#U#4*H&# 1(33-#',-#-3#'((#$%&"&#$%*,4"9#!,-#U#'0$2'( (5#43$#$%&#%*4%&" $#(&H&(#3]#,3, ` )*"+9#U# 6'"#03,"*-&)&-#$3#1&#(*+&#$%&#(36&"$#/3""*1(&#)*"+#1&0'2"]#.5#%&'($%# (&H&("9#R36#U#-3#/('5#*0&#%30+&5#',-#U#$)5#$3#&'$#6&((#',-#$'+&#H*$'.*,"#',-# "$2]]#(*+&#$%'$9#I2$7#$%&#.3"$#"*4,*]*0',$#-*]]&)&,0&$6&&,#$%&#]*)"$#)'$*,4# ',-#$ %&#"&03,-#6'"#$%&#I ` K50(&7#$%&#'--*$*3,#3]#$%&#I ` K50(&#)*-*,49# # # F%&#"&(&0$#/')$*0*/',$"#6%3#6&)Ĭ'"*3,'(#2"&)"#032(-#,3$#*-&,$*]5#.',5# &03,3.*0#&]]&0$"#$%'$#$%&5#032(-#'$$)*12$&#$3#$%&*)#2"]#G&,H&)#I ` 050(&9#C&H&)'(# "'*-#$%'$#$%&5#%'-#/)31'1(5#"'H&-# '#1*$#3]#.3,&5#*,#/')+*,4#03"$"7#"3.&$*.&"#*,# '""30*'$*3,#6*$%#'$$&,-*,4#"/3)$*,4#&H&,$"#3)#&,$&)$'*,.&,$#6%*(&#-36,$36,9#N',5# 300'"*3,'(#2"&)"#'0+,36(&-4&-#$%'$#*]#$%&5#6&)&#$3#2"&#$%&#"5"$&.#.3)&7#$%&5#632(-# (*+&(5#&,032,$&)#&03,3.*0#1&,&]*$"9#VH&)'((7#$ %3"&#.3"$#])&J2&,$(5#',-#03,"*"$&,$(5# .3$*H'$&-#$3#2"&#G&,H&)#I ` 050(&#)&/3)$&-#$%&#.3"$#1&,&]*$"7#13$%#*,#$&)."#3]#%&'($%# ',-#&03,3.*0"9 # Factors For Incorporating B ehavior C hange Into Lifestyle V,0&#'#2"&)#%'-#$)*&-#G&,H&)#I ` 050(&7#6%&$%&)#$%&5#6&)&#'1(&#$ 3#-&H&(3/# .3$*H'$*3,'(#)&'"3,*,4#]3)#03,$*,2*,4#2"&7#',-#&H&,$2'((5#'#0%',4&-#1&%'H*3)7#6'"#

PAGE 192

177 -&/&,-&,$#3,#.2($*/(&#]'0$3)"9#!#/&)0&/$*3,#3] $* )"*$./$)&&*""$ *"#03..3,#$3#.3"$# )&42(')#2"&)"9#F%&# (*"!,+)!,.+"$"*28*( #')&#'("3#*./3)$',$#$3#.',5#2"&)"7#'"#'# ]'H3)' 1(&#"$'$*3,#(30'$*3,#,&')#'#-&"*)&-#-&"$*,'$*3,"#'((36&-#$%&.#$3#.3)&#)&'-*(5# 0%33"&#I ` 050(&9#F%&# (*7*+()%,4,!6 #3]#G&,H&)#I ` 050('"#'#)&3002))*,4#$%&.&7#',-# )&42(')#2"&)"#)&/&'$&-(5#"$)&""&-#$%&#*./3)$',0]#%'H*,4#'#1*+&#'H'*('1(%&,# ',-#6%&)&#*$#6'" #,&&-&-9#X'0%#3]#$%&",0&/$"#03,$)*12$&-#$36')-#'#"&(] ` '""&"".&,$#3]#%36#6&((#G&,H&)#I ` 050(&# ]*$#$%& $4,/*"!64* #3]#2"&)"9# # X'"]#'00&"" _#V]#$%3"%3#*,$&4)'$&-#G&,H&)#I ` 050(&"$#*,$3#$%&*)#(*H&"7# $%&#(34*0#3]#2"*,4#"%')&-#1*+&"#"&&.&-#$3#]*$#6*$%#$% &*)#,&&-"9#!#2"&)#6%3#%'-# /)&H*32"(5#,3$#)*--&,#'#1*+&#*,#.3)&#$%',#<8#5&')"#]32,-#$%&#"5"$&.#"2*$&-#%*"# .31*(*$5#,&&-"9#T&#"'*-_## # C3#U#)3-&#])3.#$%&#I ` 050(&#"$'$*3,#'$#$%&#(*4%$#)'*(#"$'$*3,#6%&)&#U#4&$#3]]7#$3# $%&#I ` 050(&#"$'$*3,#'132$#'#1(30+#])3.#63)+ 9#F%&)&)&# 6&&+" #$%*"#"2..&)# 6%&)&#U#)3-&#&H&)5#-'59#F%&)&)&#$*.&"#U#632(-#%'H&#'#(2,0%#.&&$*,4#*,# $36,#',-#Um-#6'(+#2/7#Um-#4&$#'#1*+&7#',-#Um-#)*-&#*, # $3#(2,0%#',-#-30+#*$#',-#43# $3#.5#(2,0%#.&&$*,4#',-#Um-#/*0+#2/#',3$%&)#1*+&#',-#Um-#03.'0+9# # # F%*"#2"&)#"&&.&-#"3.&6%'$#"2)/)*"&-#$%'$#$%*+&"#63)+&-#"3#6&((#]3)#%*.7# ',-#$%'$#%ɸ(-#2"&#$%*+&"#"3#])&J2&,$(57#12$#'("3#$'(+&-#'132$#$%&.#1&*,4#$%&# 1&"$#$33(#]3)#$%&#o319#!,3$%&)#)&42(')#2"&)#6%3#%'-#]3)$2,'$&#/('0&.&,$#3]#G&,H&)# I ` 050(&#"$'$*3, "#]3)#%*"#/2)/3"&"7#"'*-_ # U ,#$%&#.3),*,4"#U#6'(+#32$#3]#.5#12*(-*,47#U#6'(+#'132$#?8#]&&$7#/2((#32$#$%&# 0')-#3H&)#$%&#"&,"3) 7 #4)'1#$%*+&9#U#2"2'((5#0%33",&#$%'$l"#43$#$%&#"&'$# $%'$m"#'132$#$%&#%&*4%$#$%'$#U#(*+&9#U#$'+&#$%*+& 9 #U#)*-&#$%&#"%3)$&"$# )32$&#',-# /2$#*$#*,#$%&#"$'$*3,#$%'$l"#)*4%$#32$"*-&#.5#63)+7#"3#*$l"#(*+&#'#;?#.*,2$&# 03..2$&9#!,-#$%&,#U#-3#$%&#"'.&#$%*,4#03.*,4#1'0+#])3.#63)+ 9 # # K3,"&J2&,$(57#&'"]#'00&""#6'"#'("3#',#*""2&#]3)#$%3"%3#6&)Ĭ'"*3,'(# 2"&)"9#!#03..3,#&O0('.'$*3,#'. 3,4#300'"*3,'(#2"&)"#6'"7#jUl-#2"&#*$#.3)&#*]#$%&)&#

PAGE 193

178 6&)&#'#"$'$*3,#,&')157k#3)7#jF%&)&#o2"$#')&,l$#"$'$*3,"#6%&)&#U#,&&-#$%&.9k#N',5#3]# $%Ĭ'"*3,'(#2"&)"#&O/)&""&-#4)&'$#*,$&)&"$#*,#2"*,4#G&,H&)#I ` 050(&#',-#$%324%$# %*4%(5#3]#$%&#"5"$&.#03,0&/$2'((57#12$ #1&0'2"]#/('0&.&,$#3]#"$'$*3,"#$%'$#6&)&# *,03,H&,*&,$#$3#$%&*)#-'*(5#$)',"/3)$'$*3,#,&&-"7#'#%*4%&)#])&J2&,05#3]#2"'"#,3$# (*+&(59# # G&"$*,'$*3,"#"&)H&_#N3$*H'$*3,#$3#03,$*,2)#"2"$'*,#0%',4&"#*,#1&%'H*3)# 0',#1&#"2//3)$&-#15#/3"*$*3,*,4#"%')&-#1*+&" #6%&)&#/&3/(',$#$3#439#F%3"&# *,$&)H*&6&-#.&,$*3,&-#'#H')*&$5#3]#-&"$*,'$*3,"#])&J2&,$&-#,&')#G&,H&)#I ` 050(&# "$'$*3,"9#C3."&)"# &"$'1(*"%&-#)32$*,&#H*"*$"#$3#'$$)'0$*3," #1'"&-#3,#"$'$*3,# (30'$*3," 9#!"#3,"&)#"'*-7#jE&#"$')$&-#/*0+*,4#)&"$'2)',$"#1'"& -#3,#$%&.#1&*,4# 0(3"&#$3#'#I ` K50(&#"$'$*3,9k# # N',5#2"&)"#*-&,$*]*&-#"/&0*]*0 #(30'$*3," #3 )#03,0&,$)'$*3,"#3]#'0$*H*$5#'"# /('0&"#$%&5#(*+&#$3#43#6%&,#3,#'#I ` 050(&7#*,0(2-*,4#03]]&&#"%3/"7#)&"$'2)',$"#',-# 1')"9#U,$&)H*&6&&"#&'4&)(5#"244&"$&-#,&6#"$'$*3,#(30'$ *3,"#$%'$#632(-#1&$$&)#"&)H&# $%&*)#,&&-"9#jR&')#.5#%32"&k#6'"#$%..3,#"244&"$*3,#.'-#'((#300'"*3,'(# 2"&)"9# # G&/&,-'1*(*$5 _#G&/&,-'1*(*$5#6'"#'#$3/*0#3]$&,#/)&"&,$#*,#)&"/3,"&"9#!$#$%&# /3*,$#6%&,#$%&#*,$&)H*&6"#6&),-20$&-7#G&,H&)#I ` 050(&#%'-#&"$' 1(*"%&-#'# )&/2$'$*3,#'.3,4#.',5#3]#$%3"&#*,$&)H*&6&-#'"#1&*,4#'#03,H&,*&,$#',-#]'*)(5# -&/&,-'1('5#$3#4&$#')32,-#$36,9#V,"&)#"'*-_ # U#+*,-#3]#-&/&,-#3,#*$9#U$l "#+*,-#3]#(*+&# ( *4%$# ) '*( 7 #6%*0%#U#)&'((5#-&/&,-#3,9#U$m"# '(6'5"#3,#$*.&7#',-#*$m"#'(6'5"#43* ,4#$3#03.&#',-#2"2'((5#$%&)&m"#43*,4#$3#1&# )33.9#!,-#"3#U#0'.& #$3#-&/&,-#3,#$%&#I ` K50(&7#$33 n # *$m"#43*,4#$3#1&#)&'-5# 6%&,&H&)#U#4&$#$3#$%&#"$'$*3,9#Um.#,3$#43*,4#$3#%'H&#$3#6'*$#]3)#$%&#,&O$#I ` K50(&#$3#03.Ÿ # #

PAGE 194

179 C3.%3#2"&-#$%*+&"#]3)#03..2$*,4#3)#]3)#12"*,&""#"$)&""&-#$%�)*$*0'(# *./3)$',0&#$%&5#/('0,#-&/&,-'1*(*$59#!#2"&)#.'+&"#$%&#/3*,$_# # U#)&'((5#,&&-#$3#%'H&#'#1*+&#$%&)%&,#U#,&&-#*$9#V$%&)6*"&7#UmH+$#$3#6'*$#'# (3,4#$*.&#]3)#'#12"#$)',"]&)#3)#6' (+#'132$#<8#.*,2$&"9#U]#Um.#$)5*,4#$3#4&$#$3# 63)+#3,#$*.&7#U#-3,m$#6',$#$3#-3#&*$%&)9 # # N',5#%'-#&,032,$&)&-#"$'$*3,"#$%'$#6&)&#&*$%&)#03./(&$&(5#]2((#3)#&./$5#'$# (&'"$#3,0&#-2)*,4#$%&*)#2"]#$%&#"5"$&.9#F%3"%3#-&/&,-&-#3,#G&,H&)#I ` 050(&# ]3)#$%&*)#03.. 2$,-&)"03)&-#$%&#*./3)$',0]#%'H*,4#1*+&"#3)#&./$5#-30+"# 'H'*('1(%&,#,&&-&-9#!#2"&)#"2..&-#2/#$%&#*./3)$',0&_ # E&((7#&./$5#"$'$*3," n #U#+*,-#3]#+,36#6%&)&#.5#1'0+2/#"$'$*3,"#')&9#F%&5m)&# ,3$#3H&)(5#03,H&,*&,$7#*,#$%'$#U#43$#$3#43#*,#$%)3,4#-*)&0$* 3,7#6%*0%#'--"# $*.&#$3#.5 s 6%&,#Um.#$)5*,4#$3#.'+&#$%&#(*4%$ ` )'*(7#532#+,367#U#-3,m$#%'H&#'# (3$#3]#$*.&9#U]#$%*+&m"#,3$#$%&)&7#Um.#/)31'1(5#43*,4#$3#.*""#$%'$#$)'*,#',-#U# %'H&#$3#4&$#$%&#,&O$#3,&9#C3#U#(*+&#%'H*,4#$%*+&#$%&)&9#D2((#"$'$*3,#*"#+*,-#3]# $%&#"'.&#/)31(&.9#F%&,#UmH+$#$3#43#"/&,-#$*.&#"&')0%*,4#]3)#'#"$'$*3,#$3# /2$#$%*+&9 # # V$%&)#03,0&),"#6&)*$%#$%&#"3]$6')&#',-a3)#*,$&)]'0"&-#$3#3/&)'$&#$%&# "5"$&.9#F%&#"5"$&.#&O/&)*&,0&-#"&)*32"#/)31(&."#6*$%#'-.*,*"$)'$*H&#"3]$6')&#',-# $%&#*,$&) ]'0&4*,,*,4#3,#$%&#]*)"$#-'59#N',5#3]#$%&#/)31(&."#6&)&#&H&,$2'((5# )&"3(H&-7#12$#,3$#]3)#"&H&)'(#.3,$%"9#V,"&)#)&/3)$&-_ # UlH'$0%&-#/&3/(&#%'H&#.*"&)'1(&#$*.&"#$)5*,4#$3#125#'#3,& ` -'5# .&.1&)"%*/#])3.#$%&#+*3"+9#U$l"#'6]2(9#U#1324%$# p3,(*,&q #.&.1&)" %*/"#',-# ]32,-#32$#U#032(-,l$#2"&#*$#1&0'2"]#$%|"#*,#$%&#"5"$&.9#D')#',-#'6'5# $%?)"$#$%*,49 # # G&/&,-'1*(*$5#*"#'#0)20*'(#*""2&#]3)#$%&#*,$&)H&,$*3,#*,#&"$'1(*"%.&,$#',-# .'*,$&,',0]#0%',4&-#1&%'H*3)9#U]#'#2"&)#&O/&)*&,0&"#$33#.',5#*,"$',0&"#*,#6 %*0%# $%&#"5"$&.#-3&"#,3$#.&&$#$%&*)#,&&-"7#"a%&#.'5#-&0*-&#$%'$#-&/&,-'1*(*$5#*"# *,"2]]*0*&,$#]3)#)&42(')#2"&9# #

PAGE 195

180 !#,2.1&)#3]#"&(&0$#/')$*0*/',$"#&O/)&""&-#'#-&"*)&#$3#/('5#',#'0$*H&#)3(&# )&/3)$*,4#*""2&"#6*$%#$%&#"5"$&.9#N',5#"&&.&-#$3#)&034,*b&#$%'$7#'"#'# 03..3,#/33(# )&"32)0&7#*$#6'"#*,#$%&*)#1&"$#*,$&)&"$#$3#&,"2)&#$%'$#1*+&"#6&)&#*,#433-#)&/'*)#',-# )&'-5#]3)#2"&9#!#)&42(')#2"&)#6'"#/')$*02(')(5#034,*b',$#3]#%36#%*"#*,H3(H&.&,$# 032(-#']]&0$#$%&#"5"$&._ # U#03,"*-&)#$%&# I ` K50(&#'#/21(*0#$)2"$9#!,-#"3 #1'"*0'((57#*$l"#.5#0*H*0#-2$5#$3# .'*,$'*,#*$#',-#.'+&#"2)&#$%'$#*$l"#,3$#'12"&-#6%&,&H&)#U#0 ',#',-#%&(/#3$%&)# /&3/("&#*$ 9#Ul.#H&)5#03"$#03,"0*32"#"3#$%&)&#*"#$%'$7#12$#'("3#$%'$l"#+*,-#3]#'# 4&,&)'( 7 #532#+,36 7 #$%*"#*"#32)#0*H*(*b'$*3,#',-#6&#')&#"2//3"& -#$3#.'*,$'*,#*$9# # # L*]&"$5(& _#C&(&0$#/')$*0*/',$"#6&)& #'"+&# *] #G&,H&)#I ` 050(&#]*$# $%&*)#(*]&"$5(&9 # !.3,4#)&42(')#2"&)"7#$%&#)&"/3,"&"#6&)H&)6%&(.*,4(5#/3"*$*H&7#6*$%#.',5# *,03)/3)'$*,4#$%?)-"7#jB&)]&0$7k#j!1"3(2$&(57k#',-#jG&]*,*$&(59k#N3"$#3]#$%&" &# )&"/3,"&"#6&)&#&('13)'$&-#6*$%# -&$'*("#3]# %36#$%&#"$'$*3,"#6&)&#o2"$#6%&)&#$%&5# 6&)&#,&&-&-7#3)#'$#(&'"$#,3$#]')#])3.#/('0&"#$%'$#2"&)"#6',$&-#$3#439# N',5#3]#$%3"&# .3"$#/(&'"&-#6*$%#G&,H&)#I ` 050(&#*,#$%&*)#(*]&"$5(..&,$&-#3,#%36#$%*+&"# &,'1(&-#$%& .#$3#-3#$%*,4"#$%'$#$%&5#'()&'-5#-*-7#&O0&/$#$%'$#$%&5#-*-#,3$#%'H&#$3# -&/&,-#3,#'#0')#*,#3)-&)#$3#-3#$%&.9# # V,#$%$%&)#%',-7#300'"*3,'(#2"&)"#%'-#J2*$&#-*]]&)&,$#)&"/3,"&"#6%&,# '"+&-#'132$#G&,H&)#I ` 050(&#',-#$%&*)#(*]&"$5(&9#K3..&,$"#0&,$&)&-#3,#('0+#3] #&'"5# '00&""#$3#G&,H&)#I ` 050(&#"$'$*3,"9#V,Ĭ'"*3,'(#2"&)#"'*-7#j U$# p]*$"q #]3)#/(&'"2)&7#12$# *$l"#,3$#6*$%*,#.5#(*]&"$5(&#]3)#'*(5#03..2$)#&H&,#-'*(5#2"&999#' ,-#.3"$(5#1&0'2"&# U#'.#,3$#0(3"&#$3#'#"$'$*3,9 k#C3.Ĭ'"*3,'(#2"&)"#03..&,$&-#$%'$#*$#6'" ,l$# ,&'),&""#3]#"$'$*3,"#$%'$#+&/$#$%&.#])3.#.3)"&7#12$#'#('0+#3]#',#',,2'(# .&.1&)"%*/#3)#$%&#]2,-"#$3#'0J2*),&9# #

PAGE 196

181 C&(&0$#/')$*0*/',$"#6%3#2"&-#G&,H&)#I ` 050(&#.3"$#6&)&#$%3"&#]3)#6%*0%#$%&# "5"$&.#6'"#6&((#"2*$&-#$3#"'$*"]5#$%&*)#$)',"/3)$'$*3,#,&&-"#0 3,H&,*&,$(5#',-# -&/&,-'1(59#Y($*.'$&(57#*,-*H*-2'("#-&)*H&-#$%&*)#36,#)&'"3,"#'"#$3#6%&$%&)#G&,H&)# I ` 050(&#]*$#$%&*)#,&&-"7#']]&0$*,4#.3$*H'$*3,#]3)#2"&9#F%&#-&H&(3/.&,$#3]#"&(] ` '""&""&-#]'0$3)"#6*$%#*,-*H*-2'(*b&-#.&',*,4#'//&')"#/3"*$*H&(5#'""30*'$&-#6*$ %#$%&# /&)0&/$*3,#3]#6%&$%&)#G&,H&)#I ` 050(&#"2*$"#$%&#(*]&"$5(]#',#*,-*H*-2'(9# # Diffusion of Use within S ocial N etworks As a public health intervention, Denver B cycle relies largely on the influence of existing members to diffuse use of the bicycles thr ough interpersonal communication and modeled behavior. The active transportation behavior of users serves to influence and attract other users. Visibility, from in person experiences and interactions, as well as through media exposure contributed to diffus ing the concept and use of Denver B cycle within the population. S ocial interaction and observed behavior are key concepts in diffusion theory and social cognitive theory. Both of these concepts were intertwined throughout the interview data. Effects from social interaction and observed behavior can contribute to social change . Social interaction : All of th e interviewees exhibited some characteristics of innovators or early adopters, as described in classic Diffusion of Innovations theory (Rogers 2003) Many talked about enjoyment of trying new things and that peers often sought them for advice. Several reported first using Denver B cycle as a result of social interaction with people that they knew. All interviewees also indicated that after using Denver B cycle, they either passively or act ively attempted to recruit others to use the

PAGE 197

182 system, or at times went out of their way to inform complete strangers in how to use the system. One woman said: U#%'H&# #03 ` 63)+&)#6%3#U#03,H*,0&-#$3#2"&#*$#3,0&7#12$#%'"#,3$#*,#H&)5# 433-#"%'/&7#"3#%&#%'-#'# )&'((5#%')-#$*.&9#U#.&',#%&#)&'((5#&,o35&-#*$7#4&$$*,4# 32$#',-#4&$$*,4#"3.&#&O&)0*"&7#12$#%'"#,3$#)&'((5#.3$*H'$&-#$3#&O&)0*",# %*"#36,#6*$%32$#"3.&#/)3./$*,49# # # She went on to say that she also recruited her father to use Denver B cycle, and that he was inspired to buy his own bicycle. A few of th &#"&(&0$#/')$*0*/',$"#6&)&#*,#/)3]&""*3,'(#/3"*$*3,"#$%'$# &,'1(&# $%&.#$3#']]&0$#03./',5#/3(*057#"3.&$*.&"#03,H*,0*,4#(&'-&)"%*/#$3#/'5#]3)#G&,H&)# I ` 050(&#',,2'(#.&.1&)"%*/"#]3)#&./(35&&"9#U,#3, & #"20%#*,"$', 0&7#'#)&42(')#2"&)# )&/3)$&-_ # !,-#"*,0&#Um.#)&"/3,"*1(&#]3)#TQ#%&)&#U#6&,$#$3#$%&#KXV#',-#U#"'*-7#$%*"#*"#'# 4)&'$#TQ#1&,&]*$9#E&#'()&'-5#/'5#'#/21(*0#$)',"/3)$'$*3,#1&,&]*$#]3)#32)# &./(35&&"9#U$m"#'#4)&'$#1&,&]*$9#U$l((#(&$#/&3/(&$#'#(*$$(&#&O&)0*"&7#(&$ #/&3/(&# 6%3#6',$#$3#$'+&#$%&#$)'*,#03..2$)#*]#$%&5#o2"$#6',$#$3#)*-&#*,$3#$36,#$3# %'H&#(2,0%#',-#03.'0+7#"*,0&m)&#+*,-#3]#3,#$% $"+*)$"#3]#$36,9#I ` K50(&#*"#4)&'$#]3)#32)#&./(35&&"9#C3#6&#.'-&#*$#'H'*('1(&#]3)#'((#3]#32)# &./(35&&"#'$#$%*"#(30'$* 3,n#6%&)&#/'5#$%&#.&.1&)"%*/#]&&7#',-#$%&5#/'5# '((#$%&*)#36,#2"&)#]&&"9# # # G*]]2"*3,#1&%'H*3)#*"#,3$#"3(&(5#$%&#-3.'*,#3]#$%&#.3"$#&,$%2"*'"$*0#)&42(')# 2"&)"9#!$$&./$"#$3#*,](2&,0&#$%&%'H*3)#3]#/&&)"#')&#'("3#'//')&,$#'.3,4#$%3"&# 6%3#6&)Ĭ'"*3,'(#2 "&)"9#C&H&)'(#300'"*3,'(#2"&)"#"244&"$&-#$%'$#$%&5#/('5&-#'# )3(&#*,#*,](2&,0*,4#3$%&)"#$3#)*-*050(&"#1&0'2"]#$%&*)#2"]#G&,H&)#I ` 050(&9#V,&# 300'"*3,'(#2"&)#"'*-_ # I&0'2"]#pG&,H&)#I ` 050(&q7#. 5#"*"$&)#',-#U#%'H&#"$')$&-#)*-*,47#03..2$*,4# ])3.#32)# %3.&"#2/#*,#$%&#"212 )1"#-36,#$3#$%�*$5#$34&$%&)9#C3#pG&,H&)#I ` 050(&q#%'" #-&]*,*$&(5#%'-#'#"/*((3H&)#&]]&0$9#U#&,032)'4&#',53,&#',-#&H&)53,&#U# +,36#$3#$'+&#/')$#*,#*$7#',-#Um.#%'//5#$3#"2//3)$#*$#'"#6&((9## # #

PAGE 198

1 83 !,3$%&)#300'"*3,'(#2"&)#.'5#%'H&&,#3,#$%&#)& 0&*H*,4#&,-#3]#"30*'(# *,](2&,0&9#C%&#.&,$*3,"#$%'$#'($%324%#"%&#*"#,3$#'#)&42(')#2"&)7#"%&#*"#,36#.3)&# *,$&)&"$&-#*,#1*050(*,4#]3)#$)',"/3)$'$*3,#1&0'2"]#%&)#])*&,-"9#E%&,#'"+&-#*]# 3$%&)"#%'H&#$)*&-#G&,H&)#I ` 050(&0'2"]#%&)7#"%&#)&"/3,-&-_ # I&0'2"&# U#-3v #R37#.3"$#3]#$%&.#'()&'-5#-39#!,-#"3#1&0'2"&#U#%'H"&-#'#1*+&# '#(*$$(*$#.3)&7#$%&)&#')&#'#032/(&#$%'$#UlH&#$'(+&-#$3#'#(*$$( *$#.3)&# ])&J2&,$(5#'132$#*$n# U#%'H&#'#(3$#3]#])*&,-"#6%3#%'H"&-#I ` K50(&#'#(3$# -2)*,4#$%*"#5&')9#!,-#"3#6&lH&#$'(+&-# .3)&#'132$#*$7#/)31'1(5#.3)&0'2"&# 3]#$%&*)#2"&#$%',#1&0'2"]#.*,&9 # # I&53,-#*, ` /&)"3,#"30*'(#03,$'0$7#"3.]#$%&#"&(&0$#/')$*0*/',$"#)&/3)$&-# 2"*,4#3,(*,&#"30*'(#.&-*'#$3#1)3'-0'"$#$%&*)#2"]#$%&#"5"$&.#3)#$3#*,](2&,0&#$%&*)# /&&)"9#F%&#G&,H&)#I ` 050( &1"*$&#*"#&,'1(&-#6*$%#(*,+"#$3#'((36#2"&)"#$3#2/-'$&# $%&*)#2"&)#"$'$2"#H*'#3,(*,&#"30*'(#"*$&"7#"20%#'"#D'0&I33+#',-#F6*$$&)9#F%&1"*$&# ]&'$2)&"#'#(&'-&)#13')-#"%36*,4#$%&#$3/#2"&)"#)',+&-#15#.*(&"#)*--&,7#$)*/"#.'-&7# &"$*.'$&-#0'(3)*&"# 12),&-#',-#&" $*.'$&-#/32,-"#3]#0')13,#&.*""*3,"#'H3*-&-9#!# )&42(')#2"&)#6%3#$)'0+"#%&)#/)34)&""#H*'#$%&#"*$'24&"#%&)#$3$'("#'4'*,"$#$%'$#3]# %&)#])*&,-"_# # Ul.#'(6'5"#0%&0+*,4# p$%&1"*$&q #1&0'2",]#.5#])*&,-"#*"#3,#$%&#$3/# $%)&)#$3/#]32)7#',-#Ul.#(*+&#'((#)*4%$7#Ul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n#U#%*$#$%&#?88 ` .*(&#.')+#',-#*$#"'*-#U#%'-#43,&#%'(]#6'5#$3#$%&#.33, 9 #U# 6'"#(*+&7#jE36#$%'$#*"#033(9k # U#(*+&#$3#/3"$#.5#-'$'#3,#$%&#(&'-&)#13')-#',-#"&&#*]#U#0',#4&$#)',+&-# '4'*,"$# "3.&13-5 7 #"3#*$ l "#]2,#$339#F%'$l"#'#%24&#.3$*H'$3)#'0$2'((59#C3.&$*.&"#*]#U#

PAGE 199

184 "&&#Ul.#0(3"&#$3#4&$$*,4#)',+&-#U#6*((#43#32$#',-#$'+&#'#J2*0+#)*-&#o2"$#$3#4&$# 3,#$%&)&9# # # F%&#N'&"#U,0*-&,$ _# C3."&)"# 6&)&# *,](2&,0&-#$3# 2"& #G&,H&)#I ` 050(&# $%)324%#3$%& )#"30*'(#/%&,3.&,' 9# F%&#.3"$#,3$'1(஺))&,0&#"$&."#])3.#',# *,0*-&,$#6*$%#G',#N'&"7# $%&#<8;8#Q&/21(*0',#0',-*-'$&#]3)#c3H&),3)#3]#K3(3)'-39# N'&"#'((&4&-#$%'$#G&,H&)#I ` 050('"#/')$#3]#'#/(3$ # $3#0&-&#(30'(#03,$)3(#$3#$%&# Y,*$&-#R'$*3,"7#*,#"3.,&O/('*, &-# 6'5 9#U,#',#!242"$#?7#<8;8#G&,H&)#B3"$#')$*0(&# ZV"%&)#<8;8[ 7#)&4')-*,4#N'53)#T*0+&,(33/&)l"#"2//3)$#3]#G&,H&)#I ` 050(&#N'&"#6'"# J2 3$&-#'"#"'5*,4 # F%&"& # ')&,m$#o2"$#6').7#]2bb5#*-&'"#])3.#$%&#.'53)9 # F%&"&#')&#H&)5# "/&0*]*0# "$)'$&4*&"#$%'$#')& # -*0$'$&-#$3#2"#15#$%*"#Y,*$&-#R'$*3,"#/)34)'. # $%'$#.'53)"# %'H&#"*4,&-#3,#$39 # # F%&#!""30*'$&-#B)&""#/*0+&-#2/#$%)*4*,'(#')$*0(&7#',-#6*$%*,#'# ]&6#-'5"#$%&# N'&"#"$3)5#%'-#1&03.&#$%&#]302"#3]#,'$*3,'(#',-#*,$&),'$*3,'(#'$$&,$*3,9#F%&#N'&"# *,0*-&,$#)&"2($&-#*,#.&-*'#"'$2)'$*3,#3]#%*"#)&.')+"#$3#'#/3*,$#$%'$#*$#6'"#2,(*+&(5# .',5#'-2($"#*,#$%&#G&,H&)#')&'#632(-#,3$#%'H&#'$#(&'"$#"3.&#]'.*(*')*$5#6*$% # G&,H&)#I ` 050(&9#C3.]#$%&#"&(&0$#/')$*0*/',$"#$'(+&-#o3+*,4(5#'132$#%36#$%&5#6&)&# t"30*'(*"$"l#]3)#2"*,4#G&,H&)#I ` 050(&7#3)#$%'$#$%&5#%'-#1&&,#$%&#"21o&0$#3]#'$$&,$*3,# ])3.#])*&,-"#3)#]'.*(5#]3)#$%&*)#/')$*0*/'$*3,#*,#'#tY9R9#03,"/*)'059l#!#]&6#&H&,#,3$ &-# $%'$#&H&,#])*&,-"#',-#)&('$*H&"#*,#-*"$',$#0*$*&"#3)#"$'$&"#6&)&#'6')]#G&,H&)#I ` 050(&0'2"]#N'&"9 # N'&"l # "$'$&.&,$# 1&0'.)3'-(5# /21(*"%&-#3,#D'0&I33+7#F6*$$&)# ',-#3$%&)# 3,(*,& # ',-#$)'-*$*3,'(# .&-*'7# )&0&*H*,4 #,'$*3,'(#',-#*,$&),'$*3,'(#'$$&,$*3 ,9#V ]#$%*"7# 3 ,"&)#"'*-7#jE&((7#U#"2//3"&#U#%'H&#$3#"'5#$%'$#G',#N'&"#.'-&#.&#-3#*$9#E%&,#%&#

PAGE 200

185 "'*-#*$#6'"#/')$#3]#$%&#YR#03,"/*)'05#U#$%324%$7#3+'57#UlH+$#$3#o3*,#,369k#C&H&)'(#3]# $%3"&#*,$&)H*&6&-#.&,$*3,&-#$%&#N'&"#*,0*-&,$# '"#%36 #$%&5#(&'),&-#'132$# G&,H&)# I ` 050(&7#1&4',#2"*,4#$%*+&"7#3)#*,0)&'"&-#$%&*)#2"'4&9#V,"&)# -&0(')&_ # G',#N'&" n #-&0*-&-#$%'$#I ` K50(&"#6&)&#/')$#3]#$%('0+#%&(*03/$&)#0)36-7#"3# p.5#6*]&q#-&0*-&-#"%&#o2"$#%'-#$3#-3#$%'$7#1&"*-&"#p$%*+&"q#6&)&#)&-9#UlH&# /)31'1(5#)*--&,#3H &)#&*4%$5#.*(&"#*,#$%)&&&+"#3,#I ` K50(&"9# # # C30*'(#*,$&)'0$*3,"#3]#2"&)"7#1&$6&&,#&'0%#3$%&)#'"#6&((#'"#6*$%*,#$%&,&)'(# /3/2('$*3,#6&).,*/)&"&,$#$%)324%32$#$%&#*,$&)H*&6"9#U$#*"#&H*-&,$#$%'$#.',5# /&3/(%3#1&03."&)"#3]#G&,H&)#I ` 050(&#')&#&,$%2" *'"$*0#&,324%#'132$#$%&*)# &O/&)*&,0&"#$%'$#$%&5#')&#.3$*H'$&-#,3$#3,(5#*,#$%&*)#36,#'0$*3,"7#12$#'0$*H&(5# '$$&./$#$3#)&0)2*$#3$%&)"#$3#'0$*3,9#F%&5#"&)H&#$3#/)3/'4'$&#$%"]#G&,H&)#I ` 050(&#',-#$3#*,0)&'"&#"30*'(#/)&H'(&,0]#'0$*H&#$)',"/3)$'$*3,9 # V1" &)H&-#1&%'H*3) _#V,'5#*,#6%*0%#G&,H&)#I ` 050(&7#'"#'#/21(*0#%&'($%# *,$&)H&,$*3,7#*"#-*]]2"&-#6*$%*,#$%&#/3/2('$*3,#*"#$%)324%#31"&)H'$*3,#3]#$%&# 1&%'H*3)"#3]#2"&)"#15#"/&0$'$3)"9# The bikes of Denver B cycle are uniformly red and iconic in appearance, thus people on the bikes are readily identifiable as users of the system. The stations are also readily visible by pedestrians. Visibility of the bicycles and stations plays a role in drawing attention to the system. Several select participants noted that mos t riders on B cycles "don't look like bikers." Most #6&')#3)-*,')5#0(3$%*,4#*,"$&'-#3]#"/',-&O#050(*,4#4&') 7#.'+*,4#$%&# '0$*H*$5#'//&')#.3)&#'00&""*1(& 9#V, & #2"&)#"'*_#j U#-3#,3$#6&')#"/&0*'(#1*050(*,4# 0(3$%&"#3,#'#I ` 050(&9#U#$%*,+#$%'$ l"#'#1*4#1&,&]*$#3]#$%*050(&n# U l.#6&')*,4#.5# 12"*,&""#0(3$%&"9k # C&H&)'(#3]#$%&#/')$*0*/',$"#$%324%$#3]#$%&."&(H&"#'"#&O&./(')"#$3# 3$%&)"9#C3.&#$'(+&-#'132$#1&*,4#%'//5#'132$#2"*,4#$%&#"5"$&.#',-#%36#3$%&)"#0',#

PAGE 201

186 "&&#$%'$#*$#*"#"*./(&#',-#&'"5#$3#2"&9#V,&# /')$*0*/',$#$'(+&-#'132$#$%&#/&3/(%3# /)31'1(5#6*$,&""&-#/')$#3]#%*"#03..2$&#H*'#G&,H&)#I ` 050(&7#"'5*,47#j U$m"#"%36*,4# /&3/(&#$%'$#532#)&'((5#0',#1*+&#$3#63)+7#*]#532#6',$#$39#S32#-3,m$#,&&-#$3#1&#'])'*-# $39 k # !#]&6#/')$*0*/',$"#"'*-#$%'$#$%&5#]&($#'"#$%3 24%#$%&#/21(*0#$)&'$&-#$%&.#.3)&# 4&,*'((5#6%&,#)*-*,4#'#I ` 050(&#'"#03./')&-#$3#6%&,#$%&5#6&),#/)*H'$*+&"#',-# 6&')*,4#050(*,4 ` "/&0*]*0#0(3$%*,49#C3.]#$%&#*,$&)H*&6&&"#"&&.&-#$3#1",*b',$# 3]#$%&#*./)&""*3,#$%'$#$%&5#.'+,#3$%&)#/&3/(%&,#$%&5 #')&#"&&,#)*-*,4#'#1*+&7# ',-#$%'$#$%&5#-3,l$#6',$#$3#"&&.#t'(*&,l#*,#'//&')',0&9#V,&#/')$*0*/',$#"'*-_ # U#-3,l$#6',$#$3#6&')#"/&0*'(#"%3&"#3)#',5$%*,4#&("&9#!,-#"37#532#+,367#Ul.# 03,"$',$(5#$%*,+*,4#'132$#6'5"#$3#.'+*050(*,4#.3)&#,3).'(9#S32#+,367# ,3$#6&')# p1*+&q# 0(3$%&"7#,3$#6&')#$33("7#,3$%*,4#(*+&#$%'$9# # # Q*-&)"#3]#G&,H&)#I ` 050(&#')"&)H'1(&#$%)324% 32$#$%&#-36,$36,#"&)H*0&# ')&'#-2)*,4#.20%#3]#$%&#-'59#B&3/(&#$'+&#,3$*0&7#',-#"3.&$*.&"#'"+#$%&#)*-&)"# -*)&0$(5#'132$#$%&#"5"$&.9#N3"$#3]#$%3"&#*,$&)H*&6&-#)&/3)$&-#$%'$#6%&,#$%&5#'),# '#G&,H&)#I ` 050(*+&7#$%&5#')&#)&42(')(5#'"+&-#'132$#$%&#"5"$&.#15#" $)',4&)"#3,#$%&# "$)&&$7#'"#6&((#'"#/&3/(&#$%'$#$%&5#+,369#!, #*,$&)H*&6&&# "'*-#$%'$ #%*"#03 ` 63)+&)"# %'H&# ,3$*0&-#$%'$#%&#)*-&"#I ` 050(&" 9#j F%&5#'((#03..&,$#$%'$#$%&5#"&&#.�)2*"*,4# -36,#$%&#"$)&&$#3,# pG&,H&)#I ` 050(*+&"q9k # V,]#$%&#.3"$#/)3.*,&,$#/&3/(&# *,#$%�*$5#*"#3,]#$%&#.3"$#31"&)H&-# /&3/(,#'#G&,H&)#I ` 050(*+&9#U,#<8;87#N'53)#P3%,#T*0+&,(33/&)#6'"#'# /)3/3,&,$#3]#$%&#"5"$&.#',-#])&J2&,$(5#.'-&#'//&')',0&"#*,#"2//3)$#3]#G&,H&)#I ` 050(&9#T&#)3-&#$%*+&"#'$#"/3)$*,4#&H&,$"7#*,0(2-*,4#'$#'#G&,H& )#R244&$"#1'"+&$1'((# 4'.&#',-#'#G&,H&)#I)3,03"#]33$1'((#4'.&7#',-#6'"#3$%&)6*"]$&,#/%3$34)'/%&-#

PAGE 202

187 *,#'""30*'$*3,#6*$%#$%*+&"9#F%&5#.'53) l"#'""30*'$*3,#6*$%#G&,H&)#I ` 050(&#%'"# %&(/&-#$3#*,0)&'"&# /21(*0#'6')&,&""#3]#$%&#"5"$&.9#V,& #"&(&0$#/')$*0*/',$ #"'* -_ # pF%&#FM#,&6"q#"'*-# "3.&$%*,4#(*+&7#j L 33+#'$#$%&"&#(*$$(&#)&-#1*+&"7#&H&,#32)# .'53)#*"#)*-*,4#$%&.9w#C3#U#"'*-7#6&((7#$%'$m"#033(9#E%'$#*"#$%'$v#U#6',$#$3#(&'),# .3)&9 # # Social change Select participants reported ways in which Denver B cycle had affected their personal active transportation behavior, but they also had opinions as to how the system exerted effects at a broader social level. For a public health intervention to address soc ietal level problems, the ability to make an impact that is capable of changing society is paramount. Through a series of several questions, select participants were asked about how Denver B cycle and/or their use of Denver B cycle had an effect on the com munity. Key *""2&#')&'"#6&)�%',4&"#*,#$%&# 7*2&*7!,.+$./$!2)+"7.2!)!,.+$%,&6&4,+3 7# ")/*!6$ &.+&*2+" 7#',-# 91)4,!6$./$4,/* 9# # B&)0&/$*3,#3]#$)',"/3)$'$*3,#1*050(*,4 _#C&H&)'(#3]#$%&#"&(&0 $#/')$*0*/',$"# $'(+&-#'132$#%367#'$#'#/&)"3,'(#(&H&(7# G&,H&)#I ` 050(&#%'-# %&(/&-#$%&.#$3#)&03,"*-&)# 1*050(& #]3)#$)',"/3)$'$*3,#*,#',#2)1',#"&$$*,49#I&53,-#/&)"3,'(#&O/&)*&,0&"7# .',5# /')$*0*/',$"#'("3 #)&](&0$&-#3,#"&&*,4#3$%&)"#3,#1*+&"# ',#$%&#*./(*0'$*3,"#$%'$#$%&# ,3).'(*b'$*3,#3]#$)',"/3)$'$*3,#1*050(*,4#.*4%$#%'H,#"30*&$5 9 #V,&#/')$*0*/',$# 03,$&./('$& -_ # U#%3/&#$%'$#*$m"#-3,&#$%&#"'.&#$%*,4#$%'$#*$#%'"#-3,&#]3)#.&7#532#+,369##P2"$#$3# )&'((5#%&(/#/&3/(&#$3#)&'(*b&#$%'$#$%&)&#') $%&)#6'5"#3]#4&$$*,4#')32,-9 #F%'$# $%&#/('0&"#$%'$#532#.*4%$#%'H&#$%324%$#6&)&#$33#]')#32$#3]#)&'0 %#3,#]33$7#532# +,36 7#')&#&'"5#$3#'00&""#15#1*+&9# U#"&&7#532#+,367#$32)*"$ ` $5/&"#',-#3$%&)# $5/&"#3]#/&3/(&#$%'$#532#o2"$#3,m$#,3).'((5#"&,#1*050(&"9# U#"& &#$%&.# $33(*,4#')32,-#3,#$%&#I ` 050(&"7#',-#*$m"#63,-&)]2(9# U$m"#4)&'$#$3#"&&#$%'$#$%&5# )&'(*b&#$%'$# *$m"#'00&""*1(&#$3#$%&.#',-# $%&5l )&#$'+*,4#'-H',$'4]#*$9# # #

PAGE 203

188 N',5#/')$*0*/',$"#$'(+&-#'132$#%36#$%&5#3]$&,#&,032,$&)&-#/&3/(%3# 6&)&#*,J2*"*$*H&#'132$#$%*+&"7#',-#"/&02('$&-#$%'$#$%&#/&3/(%3#$'(+&-#6*$%# $%&.#.2"$#%'H&#]&($#$%"&)"#$3#1&#'//)3'0% '1(&9#F%*"#.'5#*,-*0'$&#$%'$#$%3"&# /&3/(%3#6&)&#'"+*,4#$%&# G&,H&)#I ` 050("&)"#J2&"$*3,"# ]&($#"3.&$%*,4#*,# 03..3,#6*$%#$%&.7#3)#/&)0&*H&-#3]#$%&.#'"#,3$#1&*,4#3,#$%&#]')#"*-]#'#"30*'(# 1'))*&)9#C3.&#'("3#"244&"$&-#$%'$#$%&#'//)3'0%'1*(*$5#)&('$&"#$3# $%&#-*]]&)&,0&#*,# 0(3$%*,4#63),#15#"$&)&3$5/*0'(#1*050(*"$"#H&)"2"#"$',-')-#"$)&&$#0(3$%*,4#63),#15# .',5#G&,H&)#I ` 050("&)"9#!#12"*,&"".',#6%3#*"#'#)&42(')#2"&)#)&032,$&-#%367# 6%&,#%&#&,$&)&-#'#.&&$*,4#-)&""&-#3$%&)6*"&#'"#%ɸ(-#,3).'((57#12$#0'))5*,4 #'# 1*+&#%&(.&$7#%'"#.&$#6*$%#02)*3"*$5#',-#'00&/$',0#%*"#'""30*'$&"9#T&#"'*-_ # !0$2'((5# U#6&,$#$3#'#('65&)l"#3]]*0&n# 12$#U#%'-#'#%&(.&$ 9#F %&5#6&)&#+*,-#3]# (*+& 7 # jE %'$l"#$%&#%&(.&$#]3)v k #C3#U#%'-#$3#&O/('*,#.5"&(]#1'"*0'((59#I2$#$%&)&# 6&)& #0&)$'*,(5#,3#3 --#(33+"#3)#',5$%*,4#(*+&#$%'$9#F %&5#6&)&#*./)&""&-#*]# ',5$%*,49# # # N',5#3]#$%&#/')$*0*/',$"#)&0'((&-#.3"$(5#/3"*$*H&#)&"/3,"&"#])3.#3$%&)# /&3/(,#$%&#"$)&&$#6%*(&#)*-*,4#G&,H&)#I ` 050(*+&"9#C&H&)'(#"'*-#$%'$#$%&5# )&.&.1&)&-#&,032,$&)*,4# ".*(&"#',-#(33+"#3]#*,$&)&"$9#N3"$#/')$*0*/',$"#)&/3)$&-# 03,4&,*'(#)&"/3,"&"#])3.#1*050(*"$"#',-#,3, ` 1*050(*"$"#'(*+&9#V,&#/')$*0*/',$#"'*-_ # V ,]#$%&#"*-&#&]]&0$"#$%'$#6&)&#&,o35'1(&#'132$#)*-*,4#$%&.7#6'" #/&3/(&# 632(-#3]$&,#".*(&#'$#2"n# E%*(&) $#$%&)&#$%&5#632(-7#532#+,367# ".*(&#/)31'1(5#1&0'2"]#$%'5# 6& #(33+&-7#/)31'1(5#,3$#1&0'2"&7#jT&5# $%3"&#')&#'#032/(]#/&3/(%3#')&7#532#+,367#*,$&)&"$&-#*,#$%&# &,H*)3,.&,$9k#U$l"#o2"$#.3)&#(*+&7#jE36#$%'$l"#$%*4#)&-#1*+&"7k#+*,-#3]#'# $%*,49# # # C3 .&#/')$*0*/',$"#6&)&#]'.*(*')#6*$%#$%*050(*,4#03..2,*$5#*,#$%�*$5#',-# ,3$&-#$%'$#.',5#3]#$%&#)*-&)"#3]#G&,H&)#I ` 050(&#-*-#,3$#(33+#(*+&#"&'"3,&-#1*050(*"$"9# C3.]#$%&"&#/')$*0*/',$"#&J2'$&-#$%&#&.&)4&,0]#*,])&J2&,$#3)#*,&O/&)*&,0&-#

PAGE 204

189 1*050(*"$"#'"# '#/3"*$*H&#-&H&(3/.&,$#$%'$#"*4,'(&-#(')4&)#0%',4&"#$%'$#.'5#1&# 3002))*,4#*,#$%�*$59#V,&#"&(&0$#/')$*0*/',$#6*$%#"&H&)'(#5&')"#3]#$)',"/3)$'$*3,# 1*050(*,4#&O/&)*&,0"&)H&-_ # UlH&#"&&,# r #*$l"#o2"$#'#/%&,3.&,'(#'.32,$#3]#/&3/(Ȕ#"& $#3,#$%&#I ` K50(& #6%&,#$%&'$%&)#*"#,*0&9#U#$%*,+#6%&,#532#"&&#$%3"&#+*,-#3]#.&.1&)"7# *$l"#*./3""*1(&#$3#,3$#]&&(#$%&)&l"#"3.&#+*,-#3]#0%',4,-&)6 '57#5 32#+,36 v# ,-#U#$%*,+#532#')&#"&&*,4#'#(3$#3]#/&3/( $#$%&)%3#')&#,3$#3,#1*+&"# 2"2'((59#!]$&)#$6&(H&')"#3]#1*+& #03..2$*,47#532#"&&#"/&0*]*0#1*+&# 1&%'H*3)"#$%'$#532#/)31'1(5#632(-,l$#"&&#])3.#"3.&3,%3#%'"#"/&,$# '# "*4,*]*0',$#'.32,$#3]#$*.,#'#1*+&9 # # C&H&)'(#/')$*0*/',$"#6%3#%'-#2"&-#$%&#"5"$&.#]3)#.20%#3]#$%&#$*.&#*$#%'-# 1&&,#*,#3/&)'$*3,#,3$&-#%367#3H&)#$*.&7# "&&*,4#j)&-#1*+&"k#6'"#1&03.*,4# *,0)&'"*,4(5#03..3,7#',-#$%'$#*$#"&&.&-#'"#$%324%#$%*+&"#6&)&#j&H&)56%&)&9k# N',5#3]#$%&#/')$*0*/',$"#"&&.&-#$3#%'H&#'#"&,"]#"'$*"]'0$*3,#6%&,#$'(+*,4#'132$# $%)36*,4#/)&H'(&,0&#',-#/)3.*,&,0]#G&,H&)#I ` 050(&9#T36&H &)7#,3$#'((# /')$*0* /',$"#6&)�&)$'*,#$%'$#G&,H&)#I ` 050(&#%'-#5&$#4&,&)'$&-#.20%#3]#',#&]]&0$# 3,#"30*&$5#'$#(')4&9#V,&#/')$*0*/',$#03..&,$&-_ # U#-3,l$#$%*,+#*$#%'"#5&$7#1&0'2"&#U#-3,l$#$%*,+#*$l"#%'-#'#1)3'-#&,324%#*./'0$# 5&$9#U#$%*,+#*]#I ` 050(&#"200&""]2(( 5#&O/',-"#$3#$%&#/3*,$#6%&)&#$%&5#4'*,#$%'$# 0)*$*0'(#.'""#3]#/'*-#2"&)"7#$%&,#U#$%*,+#532#032(-#"&&#$%'$#"%*]$9# # # F%&#/')$*0*/',$#6%3#.'-&#$%'$#"$'$&.&,$#6'"#03,0&),&-#$%'$#$%&#"*b]#$%&# G&,H&)#I ` 050(&#"5"$&.#6'"#,3$#5&$#"2]]*0*&,$7#',-#$%'$#*$#"%32(-#1& #&O/',-&-#$3# "&H&)'(#.3)&#,&*4%13)%33-"#',-#"&0$3)"#3]#$%�*$5#1&]3)&#*$#632(-#1&#'1(&#$3#']]&0$# $%&#"30*'(#"/%&)&9# # C']&$5 #03,0&)," _#U,#3)-&)#]3)#G&,H&)#I ` 050("&#$3#1&03.*-&(5#"30*'((5# '00&/$&-7#*$#*"#,&0&""')5#]3)#*$#$3#1,"*-&)&-#,3#.3)&#*,%&)&,$(5#-',4&)32"#$%',# 3$%&)#$)'H&(#.3-&"9#R&')(5#'((#$%&#*,]3).'$*3,'(#.'$&)*'("#/)&"&,$&-#$3#2"&)"#3]#

PAGE 205

190 G&,H&)#I ` 050(&#%*4%(*4%$#$%&#,& &-#$3#03,-20$#$%&."&(H&"#"']&(5#6%*(,#$%*+&"9# X"$'1(*"%*,4#',-#.'*,$'*,*,4#03,-*$*3,"#*,#6%*0%#$%*050(&"#0',#1&#"']&(5#)*--&,#*"# 3]#/')'.32,$#*./3)$',0&#$3#'((#03,,&0$&-#6*$%#$%&#"5"$&.7#',-#*"#*,$&4)'(#$3#',5# 43'(#3]#&(&H'$*,4#$)',"/3)$'$*3,#1*05 0(*,4#$3#'#1)3'-(5#"30*'((5#'00&/$&-#1&%'H*3)9# # F%,0&),#]3)#"']&$5#&O$&,-&-#$3#$%&#"&(&0$#/')$*0*/',$"9#N',5#*-&,$*]*&-# $%&*)#.3"$#'02$,0&),"#6*$%#)&4')-#$3#"']&$5#6%*("*,4#G&,H&)#I ` 050(&#'"# H&%*02(')#$)']]*0#',-#1&*,4#"&&,9#N3"$#)&/3)$&-#$%'$#$ %&5#-*-#,3$#$)2"$#$%&# '$$&,$*H&,&""#3]#0')#-)*H&)"7#&"/&0*'((5#6%&,#-)*H&)"#.'-&#$2),"#$%'$#0)3""&-#*,$3# $%*+&#(',)#6%&,#0')#3002/',$"#3/&,&-#0')#-33)"#*,$3#$%&#/'$%#3]#',#3,03.*,4# 1*050(&9#!#]&6#%'-#03,0&),"#$%'$#"3.�')#-)*H&)"#-*-#,3$#)&034,*b&#$ %&#)*4%$#3]# 1*050(*"$"#$3#1,#$%&#)3'-9#C&H&)'(#$%324%$#$%'$#$%&"&#"']&$5#03,0&),"#'0$&-#'"#'# 1'))*&)#/)&H&,$*,4#.3)&#/&3/(&#])3.#2"*,4#$%&#"5"$&.9#V,&#/')$*0*/',$#"2..&-#2/# $%*"$#3]#$%,0&),"#'132$#"']&$5#.'-#.',5_ # K')" 9#S&'% 7 #/&3/(&#')&#)&'((5#o2"$#31(*H*32"9#R36#U# 63,l$#'H3*-#)3'-" #"*./(5# 1&0'2"&#$%&)&#*"#0')"#3,#$%&.#12$#U#'1"3(2$&(5#$)5#$3#" $*0+#$3#$%*+&#(',&"#3)# "%'))36"n #U#1&03.&#%5/&)#'(&)$#',-#'#(*$$(*$#,&)H32"#1'"*0'((5#12$#532# +,36#$%'$l "#o2"$#/')$#3]#$%& #1*+*,4 9# F%'$#'--" #'#1*4#('5&)#3]#03.]3)$9#T'H*,4# $%*+&#(',&"#&"/&0*'((57#1&0'2"�')"#-3,l$#H*3('$&#$%*+&#(',&"9#R367#$%&# "%'))36" 7#$%&5#')&#&]]&0$*H&7#12$#,3$#'"#&]]&0$*H&9#!,-#$%&,#*$#+*,-#3]#4&$"# $%&.#*,#$%&*)#.*,-"&$#$%'$#$%&5#')&#'6')&#$%'$#1*+&"#')&# ')32,-9#C3#U#$%*,+# $%*+&#(',&"#')&#)&'((5#433-9#F%&5#&(&H'$&#.5#03,]*-&,0&#'#(3$9 # # R&')(5#'((#$%&#/')$*0*/',$"#*-&,$*]*&-#$%&."&(H&"#'"#1&*,4# *+!'1"*($)+($ &.+/,(*+! #1*050(*"$"9#F%*"#*"#'#0('""*]*0'$*3,#]3)#)*-&)"#6%3#')&#.3"$(5#03.]3)$'1(&#*,# $)']]*07#1 2$#&O/&)*&,0&#'#03,]*-&,0…"$#6%&,#1*+&#(',&"#3)#3$%&)#1*050(&#]'0*(*$*&"# ')&#/)&"&,$9#C3.&#/')$*0*/',$"#'0+,36(&-4&-#,3$#'(6'5"#]3((36*,4#$)']]*0#('6"9#!# ]&6#'-.*$$&-#$3#"3.&$*.&"#)*-*,4#$%)324%#)&-#(*4%$"#3)#"$3/#"*4,"#*]#$%&5#-&&.&-#*$#

PAGE 206

191 "']&#$3#-3#"39 #T36&H&)7#"&H&)'(#"'*-#$%'$#$%&5#%'-#1&03.&#.3),"0*&,$*32"#'132$# '-%&)*,4#$3#('6"#"*,0"*,4#G&,H&)#I ` 050(&7#1&0'2"&#$%&5#)&034,*b&-#$%&*)# *./3)$',0&9#!#0'./'*4,#03,$&./3)')5#$3#$%&#*,*$*'(#G&,H&)#I ` 050(&#"&'"3,# 03,-20$&-#15#'-H30'05#4)32/#I*+&G&,H&)# $3#j-)*HȔ)#1*+&#(*+Ȕ#-)*HȔ)# 0')k#"&&."#$3#%'H&#%'-#"3.&#&]]&0$#3,#'#]&6#3]#$%&#/')$*0*/',$"#6%3#)&/&'$&-#$%&# /%)'"&#H&)1'$*.9# # !#J2&"$*3,#3,#%&(.&$#2"'"#*,0(2-&-#*,#$%&#*,$&)H*&6"#1&0'2"]# *,$&)&"$*,4#-'$'#/'$$&),"#$%'$#&.&)4&-#])3.#*,*$*' (#','(5"*"#3]#"2)H&5#-'$'9#C2)H&5# )&"/3,-&,$"#)&/3)$&-#])&J2&,$(5#6&')*,4#%&(.&$"#6%*(&#)*-*,4#$%&*)#36,#1*+&"7#12$# .20%#(&""#])&J2&,$(5#6%*(&#)*-*,4#G&,H&)#I ` 050(*+&"9#Q&"2($"#])3.#$%&#"2)H&5#')&# "%36,#*,# D*42)&#UM9
PAGE 207

192 "'*-#$%&5#'(6'5"#63)&#%&( .&$"#0*$&-#"']&$5#-'$'#3)#/&)"3,'(#&O/&)*&,0&#]3)#$%&*)# )&'"3,*,49#F%3"%3#2"2'((5#-*-#,3$#6&')#%&(.&$"#6%*("*,4#G&,H&)#I ` 050(&# 3]$&,#"'*-#$%'$#$%&5#63)&#%&(.&$"#3,#$%&*)#36,#1*+&"7#12$#-*-#,3$#6*$%#G&,H&)#I ` 050(&0'2"&#$%&5#-*-#,3$#%'H&#'#%&(.&$# 6* $%#$%&.7#3)#$%'$#$%&5#-*-,l$#03,"*-&)#'# %&(.&$#,&0&""')5#1&0'2"]#$%&#"%3)$&)#-*"$',0&"#*,H3(H&-9# # V,&#/')$*0*/',$#"'*#'132$#6&')*,4#'#%&(.&$ 7#j I ` 050(&v#R39#!((#.5#3$%&)# 1*050(&"7#'1"3(2$&(57#'(.3"$#6*$%32$#&O0&/$*3,9k# N',5#3]#$%3"%3#&O%*1*$&-#$%*"# $$*$2-&#'0+,36(&-4&-#$%&#-*""3,',0&#*,#(34*07#12$#"244&"$&-#"*$2'$*3,'(#)&'"3,"7# "20%#'"#,3$#'(6'5"#0'))5*,4#'#%&(.&$#3)#$'+*,4#2,/(',,&-#G&,H&)#I ` 050(&#$)*/"9# N3"$# 3]#$%&#/')$*0*/ ',$"7#6%&$%&)#3)#,3$#$%&5#63)&#'#%&(.&$7#-*-#,3$#03,"*-&)#'#%&(.&$#$3# 1&#'#) &/('0&.&,$#]3)#03..3,#"&,"&7#',-#3]]&)&-#0'2$*3,#',-#'6')&,&""#'"#1&*,4# .3"$#*./3)$',$9 # \2'(*$5#3]#(*]& _# C&(&0$#/')$*0*/',$"#$&,-&-#$3#1&#]'*)(5#&,$%2"*'"$*0#'132$#%36# G&,H&)#I ` 050(&#%'-#']]&0$&-#$%&*)#J2'(*$5#3]#(*]&#/&)"3,'((57#',-#6&)/$*.*"$*0# '132$# %36#G&,H&)#I ` 050(&#.*4%$#']]&0$#$%&#J2'(*$5#3]#(*]&#*,#$%�*$59#C&H&)'(#$%324%$# $%'$#G&,H&)#I ` 050(&#%'-#'()&'-5#03,$)*12$&-#$3#*./)3H&.&,$"#*,#J2'(*$5#3]#(*]&7# 6%*($%&)"#$%324%$#$%'$#'($%324%#$%&#"5"$&.#%'-#/3$&,$*'(7#*$#%'-#,3$#-*)&0$(5# ']]&0$&-#J2'(* $5#3]#(*],#'#6*-&#"0'(&9# # C3.]#$%&#/')$*0*/',$"#$'(+&-#'132$#%36#1&03.*,4#.3)*050(& ` ])*&,-(5# 0',#*./)3H&#$%&#'//&'(#3]#$%�*$57#$3#)&"*-&,$"#',-#,3, ` )&"*-&,$"#'(*+&9#F%&5#$'(+&-# '132$#%36#$%&#/)&"$*4]#G&,H&)#*"#*,0)&'"*,47#*,#/')$#1&0'2"]#/ )3o&0$"#(*+&# G&,H&)#I ` 050(&9#V,&#/')$*0*/',$#"'*-_# # U#$%*,+#*$l"#'()&'-5#*./)3H&-#$%&#J2'(*$5#3]#(*]&9#U#$%*,+#*$l"#.'-"#.3)]#',# 3/&,#0*$5#',-#'#.3)&(03.*,4#0*$59#S32#+,36#/&3/(&#])3.#'6'57#532#

PAGE 208

193 +,367#H*"*$*,4#%&)&#(3H&#$%*+&"#',-#*$#.'+&"#2"#(3 3+#.3)&7#532#+,367# 03".3/3(*$',9#U#$%*,+#*$#.'+&"#2"7#532#+,367#(33+#.3)&#(*+&#',#*,$&),'$*3,'(# 0*$59#U#$%*,+#*,#'#(3$#3]#J2')$&)"#6&l)&#$%&#&,H5#3]#(3$"#',-#(3$"#3]#3$%&)#0*$*&"9#U# $%*,+#6&#(33+#.3)&#(*+&#!."$&)-'.#3)#"3.]#$%$%&)#0*$*&"#(*+&#B')*" 7 # %'H&#'#(3$#3]#1*050(&"#)*-*,4#')32,-#*,#*$9#U#$%*,+#*$l"#)&'((5#*./)3H&-#32)# "$'$2"9 # # N',5#3]#$%&#/')$*0*/',$"#$'(+&-#'132$#%36#)&-20&-#0')#-&/&,-&,0&#*"#'#433-# $%*,4#',-#$%'$#6*$%#"2]]*0*&,$#,2.1&)"#3]#/&3/(&#)*-*,4#1*+&"#*,"$&'-#3]#-)*H*,4#0')"7# (')4&#&] ]&0$"7#"20%#'"#*./)3H&-#'*)#J2'(*$57#.'5#1&#/3""*1(&9#C&H&)'(#/')$*0*/',$"# .&,$*3,&-#$%&#/3""*1*(*$*&"#3]#&,H*)3,.&,$'(#',-#*,-*H*-2'(#%&'($%#1&,&]*$"9#V,&# /')$*0*/',$#$'(+&-#'132$#J2'(*$5#3]#(*]&#*,#$%,$&O$#3]#%&)#&O/&)*&,0&"7#*]#'//(*&-#$3# '#(')4&)#/ 3/2('$*3,_ # S32#+,36#U#632(-,l$#1&#"2)/)*"&-#*]#532#"'6#.'51&#'"#.20%#'"#3,)#$63# /&)0&,$#3]#$%&#/3/2('$*3,#2"*,4#*$#',-#*]#3,)#$63#/&)0&,$#3]#$%&#/&3/(&#')&# $&,#/32,-"#(*4%$&)#1&0'2"&#$%&5#')"*,4#I ` K50(&7#$%'$l"#',#&,3).32"# *./)3H&.&,$#*,#$%&,&) '(#%'//*,&""#',-#(*]&#J2'(*$5#3]#$%�*$59 # # C3.]#$%&#/')$*0*/',$"#6&)&#'#1*$#.3)&#"+&/$*0'(#6*$%#)&4')-#$3#1)3'-# 0%',4&#*,#J2'(*$5#3]#(*]&7#"2 44&"$*,4#$%'$#*]#J2'(*$5#3]#(*]&#%'-#*./)3H&-#*,#/')$# 1&0'2"]#G&,H&)#I ` 050(&7#*$#(*+&(5#']]&0$&-#'#".'((&)#/3/2('$*3,9# C&H&)'(#*-&,$*]*&-# $%&#(*,+#1&$6&&,#$%&#/3$&,$*'(#$3#']]&0$#J2'(*$5#3]#(*]&#',-#/)3O*.'$&#'00&""#$3# G&,H&)#I ` 050(&#"$'$*3,"9#V,&#/')$* 0*/',$#')$*02('$&-#$%&#(*,+_ # U#$%*,+#*$#']]&0$"#$%&#J2'(*$5#3]#(*]&#]3)#/&3/(%3#(*H&#,&')#"$'$*3,"#)*4%$#,36# ',-#')*((*,4#$3#$'+&#'-H',$'4]#*$9##I2$#*,# 3)-&)#$3#)&'((5#%'H&#',# *./'0$# 3,#$%�*$5#'"#'#6%3(&7#U#$%*,+#*$#632(-#,&&-#$3#1&#'#(')4&)#"5" $&.9##S32#+,367# )*4%$#,36#*$#&,03./'""&"#/)31'1(5#'#/)&$$5#".'((#/)3/3)$*3,#3]#$%&# /3/2('$*3,#3]#G&,H&) 9 # # !((#3] #$%&#"&(&0$#/')$*0*/',$"#)&032,$&-#"30*'(#*,$&)'0$*3,"#$%)324%#6%*0%# $ %&5#%'-#1&&,#*,H3(H&-#6*$%#-*]]2"*3,#3]#2"]#G&,H&)#I ` 050(&9#C3.&)&#. 3)&# '0$*H&#/)3/3,&,$"#$%',#3$%&)"9#!#]&6#3]#$%&#.3)&#'0$*H&#/)3/3,&,$"#')&#*,$*.'$&(5#

PAGE 209

194 034,*b',$#3]#%36#$%&*)#'0$*3,"#03,$)*12$&#$3#"%'/*,4#$%&#'0$*3,"#3]#3$%&)"9#U$#*"# &H*-&,$#$%'$#G&,H&)#I ` 050("&)"#')&7#'$#H')5*,4#(&H&("7#*,$&4)'((5#*,H3(H&-#6*$%#$%&# / )3/'4'$*3,#3]#2"]#$%&#"5"$&.#$%)324%#"30*'(#03,,&0$*3,"7#',-#*,#$%&#&O%*1*$*3,# ',-#-*]]2"*3,#3]#'0$*H&#$)',"/3)$'$*3,#1&%'H*3)"9# #

PAGE 210

195 CHAPTER V DISCUSSION AND LIMITATIONS Discussion Interpreting The Findings Data from the present research indicate that Denver B cycle was effective in increasing net active transportation among annual members. Nearly 1,450 individuals benefited from using Denver B cycle as an active transportation alternative to car use. Annual members logged a weekly average of 60.3 minutes of checkout time on shared bikes, of which an estimated 35.5% to 50.0% of this time replaced car use and comprised a net increase over any existing level of active transportation These benefits to annual members were sustained on average for more than 10 weeks, equal to 31% of the time during which Denver B cycle was i n operation during 2010. The behavior observed in annual members is consequential for two interrelated trends: a shift away from car use toward use of shared bikes, and an increase in overall active transportation The data show a quantitative average incr ease in active transportation behavior among annual members as a direct result of decreased car use due to Denver B cycle. Additionally, the effects of increased active transportation behavior carried through to activities among users when not on Denver B cycle bikes. Of those individuals who were interviewed, all said that their total bike use has increased since using Denver B cycle, whether on a shared bike or a personally owned bike. Even those interviewed who were not annual members said that they have increased their use of a

PAGE 211

196 bicycle for transportation, for recreation, or have begun to consider options other than a car when making choices for their daily travel needs. In sum, Denver B cycle annual users displayed a prolonged increase in total active t ransportation resulting from decreased car use. They also show ed evidence of skills to develop their own motivational reasoning and to self regulate their activity to maintain changed behaviors. Applied through infrastructure, the intervention affected a l arge and broad population, through which users acted to further diffuse participation. These findings validate the goals of the intervention to reduce car use and to catalyze active transportation behavior, and it is promising that the effects of the inter vention may have a broader reach The findings of the present research also contribute to the foundation of evidence in support of the effectiveness of a theory based, multilevel ecological approach toward inducing active transportation behavior, as sugges ted in the literature (Sallis, Cervero et al. 2006) Denver B Cycle As A Lifestyle Intervention The quantified increase in active transportation behavior among annual users indicates that Denver B cycle is successful as a lifestyle intervention. Those who were most likely to use Denver B cycle on a regular basis shared some common characteristics, re inforcing the lifestyle connection. Analysis using ordered logistic regression found that the ability to commute via Denver B cycle and the ability to replace car use with shared bike use contributed significantly to predicting the number of checkouts by a nnual members. Each of these items suggests accommodation of Denver B cycle into the lifestyle of users through a capacity to serve recurring, utilitarian transportation needs.

PAGE 212

197 These findings are in line with the Centers for Disease Control and Prevention 's (CDC) guide for strategies to increase physical activity in communities. The CDC guide recommends that interventions for the promotion of physical activity aim for broad reach within a population, incorporate measures to assess effectiveness, and strive for long term sustainability of the effort through physical and social support elements (Centers for Disease Control and Prevention 2010) The CDC guide stresses the importance of an ecological approach in which campaigns are community wide, and incorporate community buy in of the intervention by engaging key stakeholders. The presence of Denver B cycle stations and the modeled behavior of users throughout the service area also serve as what the CDC guide terms p oint of decision prompts,' reminding people of the active transportation alternative that the intervention offers. Qualitative interviews of Denver B cycle users supported the findings of quantitative analysis. Of the annual users who were interviewed, all agreed, to varying degrees, that using Denver B cycle fit into their lifestyles. Those who were most enthusiastic about use of the system as part of their lifestyle enjoyed ready access to the stations as a consequence of geography; most having stations n ear their residence, near their place of employment, and near frequented destinations. This finding corroborate s literature that postulates the promotion of active transportation through an ecological approach using infrastructural elements supportive of d esirable behavioral action (Ogilvie, Bull et al. 2011) The Ogilvie et al. paper outlines the appropriateness of using an ecological approach as a framework in complex intervention settings in which behaviors are targeted through specific elements of the bu ilt environment. The qualitative and quantitative results of the present research substantiate links between infrastructure,

PAGE 213

198 in this case Denver B cycle stations, and behavioral action, in this case uptake of active transportation enabled by shared bikes. A few of those interviewed that most closely exhibited a lifestyle fit in their use of Denver B cycle were able to dramatically reduce their car use in a remarkably short time, and some reported sizeable health benefits. This finding indicate s that such be nefits were achieved, in part, because adherence to adopted active transportation behavior was perceived as being easy, as supported in literature finding similar links between perceived ease of active transportation with behavioral adherence (Berrigan, Troiano et al. 2001) Some of the reported health benefits among users included weight loss, increased fitness, and improved mental outlook. Such benefits suggest that public bicycle sharing has th e capacity to reduce car dependence and to return health benefits when presented as a convenient alternative to car use. The capacity for users to generate their own reasons for motivation and maintenance for active transportation behavior through use of Denver B cycle underscores its function as a lifestyle intervention. P articipants develop ing their own individualized reasons to initiate and continue use supports literature indicating that suggest participant derived individualization is key to achieving sustained behavior change (Black 2005) In evaluation of an intervention to encourage physical activity through the use of step counters, Black found that an intervention that is too rigid to allow for individualization of the level of participation or goal setting might experience problems with participant dropout, or only intermittent adherence to a changed behavior. Bla ck also suggested that individualization provides for a sense of control over action, an element valued by participants.

PAGE 214

199 Many annual members who were interviewed described a variety of reasons why they chose to use the system and continue to do so. They al so transferred their acquired knowledge of the lifestyle suitability of the intervention to their peers, who they consciously or unconsciously identified as having similar needs for which Denver B cycle would also be well suited. In each case, many of the reasons annual members gave for using the system were linked to the capacity of shared bikes to conveniently service lifestyle needs. The reasoning of many Denver B cycle annual members reflect literature that suggests a combination of individual level env ironmental learning, accompanied by biased cultural transmission, in which individuals rely on social learning or cultural transmission to acquire new behaviors, jointly interact to diffuse new behaviors within a population (Henrich 2001) In this case, the use of shared bikes, a new concept for much of the population, is diffused both by individual contact, as well as influence within social groups between peers, acquaintances or even with stran gers possessing similar lifestyle parameters. Convenience of Denver B cycle over other mode choices is a lifestyle indicator, and was a recurring theme among the group of most active users. Annual members having stations well situated to suit their lives often talked about Denver B cycle as being the easiest mode of transportation to use for many trips, even though other options were at their disposal. Those who associated Denver B cycle with a convenient lifestyle fit were often residents of medium to hig h density neighborhoods well served by Denver B cycle stations and on the periphery of the central downtown area; places in which car use is expensive, inconvenient, or overly complicated for many trips. Well connected networks of bicycle supportive infras tructure between locations of origin and destination

PAGE 215

200 are essential to encouraging adults to use bicycle for transportation (Dill 2009) Adequately supported to improve safety and to make access convenient infrastructure supportive of active transportation can encourage even traditionally low participation groups, such as women, to perceive bicycling as a viable active transportation alternative to motorized modes (Dill and Carr 2003) The findings of the present research build on the foundation of literature that identifies the presence and convenience of bicycle supportive infrastructu re as contributing toward positive behavioral action. S tealth" Intervention In an age when the public endures a nearly constant barrage of advertising from multiple sources, many people have become savvy in identifying overt marketing of which they are a target, and are wary of campaigns with any implied intent to change their behavior. In this environment, a health intervention that does not appear to be a health intervention has certain advantages. It is in this sense that Denver B cycle may be conceptu alized as a "stealth" health intervention. Although Denver Bike Sharing provides information regarding environmental, economic and health benefits from using the system on their website, the operators have carefully shaped their efforts to highlight the fu n aspects of participation. As evidence, several of those interviewed had not strongly considered the health impact of riding a shared bicycle over driving a car or using other, less active modes. Instead, some interviewees focused largely on non health r elated benefits, such as improved time availability or productivity at work, or expenses spared in comparison to car use. During questioning some of these individuals appeared to realize, perhaps for the first time, that they had also received health benef its through use of Denver B cycle. A

PAGE 216

201 few did not even initially consider riding a Denver B cycle bike to be classifiable as "exercise," although upon reflection, these individuals did acknowledge that they were indeed exerting energy to propel the bike and themselves. Another indicator that Denver B cycle functions as a stealth health intervention is the concept of fun. Both annual members and short term users that were interviewed reflected the notions of having fun or receiving enjoyment through particip ation. Most indicated that it was easier to find time in their lives for an activity that they considered to be fun, rather than an activity deemed to be merely for health benefit. Fun was a major factor for motivation, present at all stages of behavior ch ange among users, but most associated with preparation just prior to behavioral action, and maintenance of the changed behavior. Without prompting, nearly all of the annual members interviewed mentioned how the concept of fun was a primary motivator to reg ularly ride shared bikes, especially while completing seemingly mundane tasks requiring transportation, such as commuting or errands. The perception of fun through participation in Denver B cycle may be described as an outcome of the power of an individua l participant to control the way in which the intervention is engaged. Because the purposes for using Denver B cycle are individualized, users decide when, where and how they use shared bikes. The flexibility of the intervention to be re imagined by partic ipants to suit their purposes, and to be interpreted as being fun facilitates diffusion of the intervention through informal channels. This supports literature that reveals how informal social interaction to promote adoption of new behavior complements for mal guidance when targeting the behavior of a population (Elder, Lytle et al. 2007) During the process of experimentation in how

PAGE 217

202 Denver B cycle might be used, behaviors developed and cultural transmission of those behaviors occurred. Discovery through learning and experimentation leads to the development of innovative uses for an innovation (Van de Ven, Polley et al. 1999) As evidence, the Tour de B cycle, an unofficial challenge for users to v isit all the Denver B cycle stations in the system by shared bike in a single day, emerged as a result of fun derived experimentation. In the social setting of Denver B cycle users, Tour de B cycle became a challenge that at least 250 people accomplished d uring the initial year. In the larger social sphere of Denver, Tour de B cycle was recognized as the best bicycle challenge of 2011 by local media (Westword 2011) illustrating the breadth of reach of an idea spawned and diffused informally by users of Denver B cycle. As a motivational aid to support behavior change, the concept of fun also extended beyond the actual use of the bikes. Interviewees shared a variety of stories about having fun while using Denver B cycle, whether the experience was attributed to riding the bicycles or occurred as a result of accessing an enjoyable destination on shared bikes. This reinforces the idea of Denver B cycle as a "stealth" h ealth intervention that supports a lifestyle. Interviewees did not tend to focus as much on riding the bikes themselves, that is, the treatment part of the intervention, but instead appeared to think about what they were able to do as a result of riding th e bikes, or of benefits not overtly connected with physical activity. For many interviewees, any cognitive recognition of the benefits of active transportation behavior was simply a byproduct of other pursuits. Denver B cycle's assignation as being fun ap parently has other side effects. As a high profile project that received a lot of notoriety in the media and visibility in the

PAGE 218

203 public sphere during its initial year, Denver B cycle was largely granted the benefit of the doubt, even in the midst of an often cynical public. Evidence of cynical immunity occurred during the Dan Maes incident, during which the Republican candidate for governor groundlessly asserted that Denver B cycle was part of a nefarious international plot. Mr. Maes's comments were met local ly, nationally, and internationally as absurdly ridiculous by commentators across the political spectrum, most of which are not likely to have been specifically advocates of bicycling. The apparent approval and trust of the public in favor of an entreprene urial populist idea public bicycle sharing in the face of a politically motivated attack by a member of the establishment, may indicate rising populist support in the social consciousness (Fieschi and Heywood 2004) Such support, even among non users of Denver B cycle, could be a sign of a growing disenchantmen t of the business as usual approach in the typically car centric urban transportation system. While the Maes incident quickly gained a high profile and subsequently faded from view, it served as a test of the cultural resilience of the core concepts behind the intervention. If Maes' statement had been received differently, it could have disrupted the viability of public bicycle sharing in the U.S., and thus diminished any potential of shared bikes as a health intervention for some time to come. As it stands the outcome of this incident may reflect emerging positive changes to social norms. Whether or not they are active participants, many people who are increasingly aware of the constraints of life in a complex global system may be coalescing around the pra cticality of economically and environmentally sustainable solutions, such as shared bicycles, within the sphere of urban transportation (Mostafa 2011) Such changes may reflect that, as a society, space is being mad e for inclusion of bicycle transportation to fit within the lifestyles of the U.S.

PAGE 219

204 population. Population Reached Through The Intervention Denver B cycle succeeded in positively changing active transportation behavior among annual members and has become a publicly visible platform to showcase active transportation as viable and achievable in everyday life. During the initial year, in addition to 1,446 annual members, more than 30,000 individuals experienced Denver B cycle as short term users. The total num ber of people who used Denver B cycle during 2010 is equal to about 10% of the adult commuter population of the City of Denver, or about 30% of the number of employees in central downtown (Downtown Denver Partnership 2011) an extensive number of people to have personally engaged in the intervention. The impact of Denver B cycle reached many population subgroups, in contrast wi th previous interventions that have focused largely on specific, mostly homogeneous target groups. Intervention efforts to affect physical activity among school aged children (Dishman, Motl et al. 2004) or among older adults (Stralen, Vries et al. 2009) have returned positive results, however, heterogeneous young to middle aged adult populations have not often been the subject of focus. Results from the evaluation of Denver B cycle suggest that a broad swath of young to middle aged adults have participated in the intervention. This finding is confirmation of the possibility of achieving a wide impact by implementing the intervention through infrastructure, as was h inted at in the findings of similar smaller studies, such as the High Point intervention (Krieger, Rabkin et al. 2009) Denver B cycle reaches a broad population of participants in part because it does

PAGE 220

205 not target any specific subgroup, nor does it strictly define the terms of participation. Users are free to determine the level and type of engagement suitable for their own individual purposes. The results of evaluation of the reasons why Denver B cycle annual members continue to participate again substantiate prior research, that individualization of participation is important in supporting self regulation (Black 2005) The ability for participants to control their involvement may be a crucial aspect of an intervention designed to incorporate physical activity into the busy lives of young to middle aged working people. Although all demographic subgroups had some representation among the group of Denver B cycle users, there is certainly much room for improvement, specifically in attracting mor e racial and ethnic minorities, women, younger and older adults, and those at the lower end of the socio economic strata to become annual members. However, it is also important to note that actual use of the system as indicated by numbers of checkouts per individual was much more equally distributed than might be expected, given the demographic profile of users. This finding is important in the evaluation of Denver B cycle in its ability to affect the general population. It suggests that once individuals ha ve become annual members, they are as likely to incorporate the regular use of shared bikes into their lives as much as any other annual member, regardless of individual demographic status. The intervention successfully engaged young and middle aged workin g adults. As discussed in previous chapters, this group is often overlooked as a target of physical activity interventions, even though its members may greatly benefit from a preventive health intervention. Introducing and encouraging preventive health beh aviors to

PAGE 221

206 individuals before their health status deteriorates is a key strategy to mitigate the burgeoning trend toward obesity. Young and middle aged working adults help to shape social norms and influence the behavior of society through positions of lea dership, by making decisions and setting policies. Members of this group wield power to affect the behavioral actions of other people at many levels (Sallis and Owen 2002) Parents who become accustomed to using active transportation can affect the beliefs and behaviors of their children and others around them, by modeling positive behavior and shaping family transportation habits (Babey, Hastert et al. 2009; Pabayo, Ga uvin et al. 2010) This spillover effect was apparent in some of those interviewed, who described changes in their family lives to reduce their car dependence as a result of using Denver B cycle during the day. As was also discovered during interviews, some annual members who are in positions of employment where they are able to shape workplace policies did so to encourage use of Denver B cycle within their organizations. Although the Denver B cycle user group of the initial year of operation did not mi rror the makeup of the general population, many of these pioneering users fit the description of the small but influential segment of the population who are at the vanguard of social change. Many were decisive and confident in their own ability to weigh co sts and benefits as to use of Denver B cycle, apparently concluding a net gain from participation for themselves and for their peers. Many extended extra effort to contribute in the positive development or operation of the system, and actively worked to re cruit others to join them. Broader Impacts on Bicycling Activity In Denver

PAGE 222

207 From a population scale, looking at the possible effects of Denver B cycle on bicycle use in the city as a whole reveals some evidence of broader changes in active transportation b ehavior. Data detailing total bicycle use in Denver are very limited. The most consistent existing data source to track bicycle use in Denver is the U.S. Census Bureau's American Community Survey (ACS), which has tracked bicycle commuter behavior from 2005 through 2010. The percentages of bicycle commuter mode share in Denver over the span of years in which the ACS has been conducted are shown in Figure V. 1. Source: U.S. Census American Community Survey 2005 to 2010 (U.S. Census Bureau 2012) Figure IV. 1 Denver Bicycle Commuter Mode Share, 2005 to 2010. Over the timeframe of the ACS, bicycle commuting in Denver has exhibited a rising trend, with an overall increase from 1.4% to 2.2%, though it is notable that half of the net increase occurred between 2009 and 2010. Tab le V. 1 shows details of the changes between the final two years in the series, the latter year coinciding with the introduction of Denver B cycle. Even though the total population of commuters declined in 2010 as compared to 2009, the percentage of bicycle commuters increased. The total number of 898:# 89?:# ;98:# ;9?:# <98:# <9?:# <88?# <88@# <88A# <88d# <88^# <8;8#

PAGE 223

208 bicycle commuters in Denver rose by an estimated 986 individuals in 2010 over the previous year. Table V. 1 Denver Commuter Population and Estimated Numbers of Bicy cle Commuters, 2009 ve rsus 2010. 2009 ACS 2010 ACS Difference Workers 16 years and over 307,556 296,453 11,103 Commute by bicycle 1.8% 2.2% 0.4% Estimated bicycle commuters 5,536 6,522 986 Source: U.S. Census American Community Survey 200 9 and 2010 (U.S. Census Bureau 2012) It is possible that Denver B cycle contributed to the observed increase in bicycle commuting in the city. Table V. 2 outlines estimates of the possible impact of Denver B cycle commuters on the total increase in bicycle commuting between the 2009 and 2010 ACS figures for Denver. Table V. 2 Estimated Effects of Denver B cycle Commuters on the Increase in Bicycle Commuting in Denver, 2009 to 2010 Denver B cycle annual members (2010) Percent of annual members who commute via Denver B cycle Estimated Denver B cycle commuters Estimated increased number of bicycle commuters between 2009 and 2010 (ACS) Percent increase in bicycle commuters equivalent to Denver B cycle commuters 1,446 13.2% 191 986 19.4% According to the Denver B cycle user survey dataset, 13.2% of Denver B cycle annual members reported commuting using shared bikes. When applied to the 1,446 annual members active during the initial operating season, this equates to an estimated 191 individ uals who commute using Denver B cycle. This estimated number of Denver B cycle commuters is equivalent to 19.4% of the increase in Denver bicycle commuters recorded between 2009 and 2010 by the ACS.

PAGE 224

209 With the data available, it cannot be ascertained whethe r Denver B cycle commuters accounted for any increase in the ACS figures. It is not possible to know whether or how Denver B cycle annual members responded to commute choice questions on the American Community Survey. Likewise, the Denver B cycle User Surv ey did not include questions about changes in commute mode choice. Regardless, based on data collected through interviews and observations of commuter activity at Denver B cycle stations, some people who commuted via Denver B cycle in 2010 switched from pr ior modes other than another bicycle. As a result, users of Denver B cycle contributed to the increase in bicycle commuter mode share of Denver whether or not this activity was captured in the ACS data. Other ways in which Denver B cycle was used : Joint u se of transit and shared bicycles is another possible influence of Denver B cycle on the commuting behavior of Denver. However, such multimodal trips are not recorded in the methodology of the ACS. Of annual users, 6.9% reported using Denver B cycle in con junction with transit as their commute mode. The annual members reporting joint Denver B cycle/transit commuting are not included in the 13.2% who reported Denver B cycle as a sole commute mode. Therefore, a total of 20.1% of Denver B cycle annual members used shared bikes as all or part of their commute. During the interviews, some annual members who are Denver B cycle/transit commuters revealed they are able to replace a car commute only because of access to Denver B cycle. These multimodal commuters sai d that transit alone was not a viable alternative personally, because either transit stops or scheduled connections between routes were not convenient for their needs. This situation reflects another aspect of the

PAGE 225

210 lifestyle suitability of the intervention. Some multimodal commuters are able to reduce their car dependence because joint access to Denver B cycle and transit suits their lifestyles. Other Denver B cycle users exhibited commuting patterns that were not previously anticipated. At some Denver B cyc le stations on the periphery of central downtown where on street parking is free, some people were observed parking a car, then checking out a shared bike before riding toward central downtown. A few annual members reported similar behavior, saying that it saved them money in parking costs, and/or the hassle of locating a parking space near densely occupied destinations. Some short term users reported using Denver B cycle similarly to avoid parking costs to get to event and entertainment venues in central d owntown. Long Term Outcomes \2',$*$'$*H&#',-#J2'(*$'$*H&#-'$'#"%36#$%'$#.',5#G&,H&)#I ` 050(&#',,2'(# .&.1&)"#'//&')#$3#%'H&#'0J2*)&-#3)#"$)&,4$%&,&-#"+*(("#$3#"&(] ` )&42('$&#'0$*H&# $)',"/3)$'$*3,#1&%'H*3)#'"#'#)&"2($#3]#2"*,4#$%&#"5"$&.9#!,,2'(#.&.1&)"#-*"/( '5&-# *,0)&'"&-#,&$# '0$*H&#$)',"/3)$'$*3, #$%'$#6'"#"2"$'*,&-#3H&)#$*.&9#N',5#/')$*0*/',$"# 6%3#6&)&#*,$&)H*&6&-#%'-#,3$#/)&H*32"(5#1&&,#'0$*H&#$)',"/3)$'$*3,#1*050(*"$"7#12$# $3#H')5*,4#-&4)&&"7#'((#'0+,36(&-4&-#$%'$#']$&)#2"*,4#"%')&-#1*+&"#$%&5#6&)&#.20%# .3 )&#(*+&(5#$3#03,$*,2&#$3#-3#"39#!($%324%#G&,H&)#I ` 050(&#"2"/&,-&-#*$"#3/&)'$*3,"# -2)*,4#$%*,$&)#.3,$%"7#'((#*,$&)H*&6&&"#"'*-#$%'$#$%&5#632(-#1'0+#$3#2"&#$%&# "%')&-#1*+&"#6%&,#$%&#"5"$&.#)&3/&,&-9# # C&H&)'(#*,-*H*-2'(#2"&)"#'//&')&-#$3#%'H&#*,*$*'$&-# /3"*$*H&7#(*]& ` '($&)*,4# 0%',4&"#*,#$%&*)#%&'($%#1&%'H*3)"7#3]#6%*0%#$%&*)#/')$*0*/'$*3,#*,#G&,H&)#I ` 050(&#

PAGE 226

211 /('5&-#'#)3(&9#C3.]#$%&#/')$*0*/',$"#6%3#&O%*1*$&-#$%)&'$&"$#*./'0$#-*-#"3#'"#'# )&"2($#3]#)&-20&-#0')#-&/&,-&,0&#',-#*,0)&'"&-# '0$*H&#$)',"/3)$' $*3, #']]3)-&-7#*,# /')$7#15#'00&""#$3#"%')&-#1*+&"9#F%&# '$$'*,.&,$ #3]#$%&"&,&]*$"#6'"#&,'1(&-#15# (*]&"$5(�%',4&"#]3)#6%*0%#*,-*H*-2'("#-&H&(3/&-#$%&*)#36,#.3$*H'$*3,'(#)&'"3,*,47# ',-#]3)#6%*0%#.&',*,4#]3)#0%',4&"#032(-#1&#-&)*H&-#',-#034,*$*H&(5#"2//3) $&-#3,# ',#*,-*H*-2'(*b&-#1'"*"9#F%*"#"$)'$&45#*"#'//)3/)*'$&#]3)#(3,4 ` $&).#.'*,$&,',0]# 1&%'H*3)#0%',4&7#1&0'2"&%'H*3)"#$%'$#')&#'-3/$&-#-2&#$3#(*]&"$5(&#]*$#'//&')#$3#1&# .3)&#)&'-*(5#*,$&4)'$&-#$%',#1&%'H*3)"#$%'$7#])3.#',#*,-*H*-2'(l"#/&)"/&0$*H&7#" &&.# ]3)0&-#3)#3H&)$(5#/)3.3$&# ZG2,,#<88^[ 9#U,03)/3)'$*,4#"2//3)$#]3)#'0$*H&# $)',"/3)$'$*3,#*,$3#/%5"*0'(#',-#"30*'(#&(&.&,$"#3]#$%&#&,H*)3,.&,$#*,#6%*0%#/&3/(&# (*H,$)*12$&"#$3#$%&#(3,4 ` $&).#"2"$'*,'1*(*$5#3]#$%&#&]]&0$"#3]#$%&#*,$&)H&,$*3,7#',-# /)3.3$*,4 #"30*'(#,3)."#]3)#('"$*,4#0%',4&# ZK&,$&)"#]3)#G*"&'"&#K3,$)3(#',-# B)&H&,$*3,#<8;8[ 9 # !"#&H*-&,0]#03,$*,2&-#'0$*H&#$)',"/3)$'$*3,#1&%'H*3)#'.3,4#2"&)"#1&53,-# $%&#-'$'#03((&0$*3,#/&)*3-#3]#$%&#/)&"&,$#)&"&')0%7#$% &#,2.1&)#3]#*,-*H*-2'(# 0%&0+32$"#(344&-#15#G&,H&)#I ` 050("&)"#,&')(5#-321(&-7#])3.#;8<7^d;#$3$'(# 0%&0+32$"#*,#<8;87#$3#<8<7A=;#*,#<8;;# ZG&,H&)#I*+&# C%')*,4#<8;;[ 9#F%*"#-)'.'$*0# )*"&#*,#'0$*H*$5#3002))&-#6*$%#3,(5#$63#'--*$*3,'(#"$'$*3,"#*,#<8;;#'"#03./')&-#$3# <8;89#I5#$%&#&,-#3]#<8;;7#<7@?^#',,2'(#.&.1&)"#%'-#o3*,&-7#'(3,4#6*$%#><7=;d# "%3)$ ` $&).#2"&)"9#F%,$*,2&-#)*"&#*,#$%&#,2.1&)#3]#0%&0+32$"# ',-#$%&#,2.1&)#3]# 2"&)"#')&#*,-*0'$*3,"#$%'$#$%&#-*]]2"*3,#3]#$%&#*,$&)H&,$*3,# /&)"*"$ "#*,#)&'0%*,4#,&6# /')$*0*/',$"#'"#*$#&O/',-"#6*$%*,#$%&#/3/2('$*3,9# # Ongoing Intervention

PAGE 227

212 G&,H&)#I ` 050(,$*,2&"#$3#1&#',#3,43*,4#*,$&)H&,$*3,9#F%�&,$)'(# 31o&0$*H]#$%&#*,$&)H&,$*3,#)&.'*,"#'"#/)3H*-*,4#(')4&#"&4.&,$"#3]#$%&#/3/2('$*3,# 6*$%#',#'($&),'$*H&#$3#0')#2"&#]3)#"%3)$#2)1',#$)*/"9#F3#63)+#$36')-#$%*"#43'(7# G&,H&)#I*+&#C%')*,4#%'"# /(',"#$3#&O/',-#G&,H&)#I ` 050(#',#'--*$*3,'(#:#3]# ',,2'(#.&.1&)"#)&/3)$&-#2"*,4#"%' )&-#1*+&"#$3#'00&""#$)',"*$9#F%&#"5.1*3$*0# )&('$*3,"%*/#1&$6&&,#$)',"/3)$'$*3,#1*050(*,4#',-#$)',"*$#*"#/)&-*0$&-#$3#1]# *,0)&'"*,4#*./3)$',0&#*,#$%&#&.&)4*,4#]302"#3,#.2($*.3-'(#)&/('0&.&,$#3]#0')# -&/&,-&,0&#*,#R3)$%#!.&)*0',#0*$*&"# ZI'0%',` N')(&'27#L')"&,#&$#'(9#<8;;[ 9#F%&# /)&"&,$#)&"&')0%#/)3H*-&"#&./*)*0'(#&H*-&,0&#$3# H&)*]5 #$%*"#)&('$*3,"%*/7#*,-*0'$*,4# $%'$#]3)#03..2$*,4#',-#3$%&)#$)',"/3)$'$*3,#/2)/3"&"7#2"&)"#3]#G&,H&)#I ` 050(&#-3# 3/$#]3)#.2($*.3-'( ` "%')&-#1*+&a$)',"* $#$)*/"#$3#)&/('0&#"3.�')#$)*/"9# # U,#'--*$*3,#$3#2"&)"7#$)',"*$#/)3H*-&)"#%'H&#)&034,*b&-#$%&#H'(2]#$)',"*$# 03,,&0$*3,"#6*$%#"%')&-#1*+&"#',-#%'H&#'("3#&.1)'0&-#G&,H&)#I ` 050(&9#!]$&)# "200&""&"#*,#$%&#*,*$*'(#5&')7#G&,H&)#I ` 050(&#*"#1&*,4#]2)$%&)#*,$&4 )'$&-#6*$%#$)',"*$#

PAGE 228

213 $3#&,032)'4&#"%')&-#1*+&a$)',"*$#.2($*.3-'(#2"&9#QFG#*"#*,#$%&#.*-"$#3]#'#.2($* ` 5&')# /)3o&0$#$3#&O/',-#$%&#(*4%$#)'*(#"5"$&.#$%)324%32$#.20%#3]#$%&#G&,H&)#N&$)3#')&'9# QFG#%'"#*-&,$*]*&-#"&H&)'(#]2$2)&#(*4%$#)'*(#"$'$*3,"#$3#%'H&#G&,H&)#I ` 050(&#"$'$*3,"# 03((30'$&-9# # !"#/')$#3]#'#)&4*3,'(#$)',"/3)$'$*3,#"0%&.&7#/21(*0#1*050(&#"%')*,4#*"#1&*,4# '-3/$&-#'"#/')$#3]#'#1)3'-&)#'//)3'0%#$3#)&-20�')#2"&9#U,#N'5#<8;;7#I32(-&)#I ` 050(/&,&-#*,#,&')15#I32(-&)7#K3(3)'-3#'"#'#".'((&)#"0'(&#"5"$&.#"*. *(')#$3# G&,H&)#I ` 050(&9#D2$2)&#/(',"#')&#]3)#.&.1&)"%*/#*,#$%&#"5"$&."#$3#1�)3""# 03./'$*1(&7#"3#$%'$#03..2$&)"#1&$6&&,#$%&#$63#0*$*&"#.'5#%'H&#$%/$*3,#3]# 03./(&$*,4#'#$)',"*$#$)*/#6*$%#"%')&-#1*050(&"#3,#&*$%&)#&,-9#!#)&"2($#*"#$%'$#'"#1*+&# "%')*,4#&O /',-"#*,#$%&#)&4*3,7#&H&)#4)&'$&)#/3/2('$*3,"#6*((#1&#&O/3"&-#',-#1&4*,#$3# 1&03.&#'002"$3.&-#$3#$%&#*-&'#]'0$3)*,4#'0$*H&#$)',"/3)$'$*3,#*,$3#$%&*)#(*H&"9 # F%&#/)&"&,$#)&"&')0%#03))313)'$&"#(*$&)'$2)&#$%'$#"244&"$"#'#-*"/')*$5#3]# &,4'4&.&,$#*,#/%5"*0'(#'0$* H*$5#'.3,4#(36 ` *,03.&#/3/2('$*3,"# ZB)*,0&7#F)&.1('5# &$#'(9#<8;;[ 9#G&,H&)#I*+&#C%')* ,4#%'H&&,#'6')]#%36#'#)&J2*)&.&,$#3]#%'H*,4#'# 0)&-*$#0')#$3#2"&#$%&#"5"$&.#/3"&"#'#1'))*&)#$3#2"&#'.3,4#$%&#(36&)#*,03.&# /3/2('$*3,9#C&H&)'(#/)3o&0$"#'),-&)6'5#$3#'$$&./$#$3#'$$)'0$#2"&)"#])3.#(36 ` *,03.&#/3/2('$*3,"7#',-#$3#)&-20'))*&)"#$3#2"&9 #G&,H&)#I*+&#C%')*,4#',-#QFG# %'H&#o3*,$(5#)&0&*H&-#'#4)',$#$3#/('0&#G&,H&)#I ` 050(&#"$'$*3,"#,&')#12"#"$3/"#*,#(36 ` *,03.&#,&*4%13)%33-"#,&')#0&,$)'(#-36,$36,9#F%&#G&,H&)#T32"*,4#!2$%3)*$5# ZGT![#%'"#1&&,#033)-*,'$*,4#6*$%#G&,H&)#I*+&#C%')*,4#$3#/('0&#,&6#"$' $*3,"#'$#GT!# /)3/&)$*&"9#GT!#',-#G&,H&)#I*+&#C%')*,4#%'H&#'("3#1&&,#63)+*,4#$3#-&H&(3/#

PAGE 229

214 "$)'$&4*&"#$3#&,'1(&#*,-*H*-2'("#6*$%32$#0)&-*$#0')-"7#3)-*,')*(5#)&J2*)&-#]3)# 03(('$&)'(#-2)*,4#"21"0)*/$*3,#$3#G&,H&)#I ` 050(&9# # B)34)&""#$36')-#*,0(2-*,4#'#4)&'$&)#,2. 1&)#3]#(36 ` *,03.&#*,-*H*-2'("#'"# G&,H&)#I ` 050("&)"#*"#*./3)$',$#*,#)&-20*,4#-*"/')*$*&"#3]#'00&""#$3#'0$*H&# $)',"/3)$'$*3,7#',-#)&-20*,4#-*"/')*$*&"#*,#%&'($%#32$03.&"9 #!"#'#03,$*,2',0]#$%&# /)&"&,$#)&"&')0%7#$%&#/)*0*/'(#*,H&"$*4'$3)#3]#$%&#/)&"&,$#) &"&')0%#%'"#"&02)&-# ]2,-*,4#$3#*-&,$*]5#',-#&O/(3)'))*&)"#$3#2"&#'.3,4#(36&)#*,03.)32/"9#!# 0&,$)'(#31o&0$*H&#*"#$3#*,]3).#"$)'$&4*&"#$3#)&-20&#"30*3&03,3.*0#-*"/')*$5#3]#2"&7#"3# $%'$#2,-&))&/)&"&,$&-#4)32/"#.'5#1&,&]*$#])3.#2"*,4#"%')&-#1*+&"9#F% ,$*,2*,4# *,H&"$*4'$*3,#*"#-&"*4,&-#$3#03((&0$#',-#&H'(2'$&#-'$'#])3.#'#4)32/#3]#/3$&,$*'(#2"&)"# '.3,4#$%&#)&"*-&,$"#3]#$%&#G&,H&)#T32"*,4#!2$%3)*$5#"5"$&.7#"/&0*]*0'((5#*,# )&"*-&,$*'(#')&'"#6*$%#',#'-o'0&,$#G&,H&)#I ` 050(&#"$'$*3,9 # Informing Application : Po licy And Infrastructure U.S. Secretary of Transportation Ray LaHood acknowledged the importance of public bicycle sharing as an emerging tool for urban transportation during a visit to the Biennial of the Americas conference held in Denver, July 2010. Secr etary LaHood, joined by Mayor Hickenlooper and national and international dignitaries, called Denver B cycle, "a model for America," as he rode with the group in central downtown (Moreno 2010) Improving health through the policy support of active transportation is an emerging focus of civic le adership (Nazelle, Nieuwenhuijsen et al. 2011) How Denve r B cycle came to be and the effects that it has had on the population of users and the city at large highlight a number of important factors for application that may benefit practitioners.

PAGE 230

215 Although the City and County of Denver did not provide funding for the development and implementation of Denver B cycle, the GreenPrint Denver initiative of the Mayor's office helped to incubate the idea into a mature concept able to stand on its own. Created as a community effort, city leadership fostered communication between members of the public, local bike advocacy organizations, neighborhood groups, members of the business community, and representatives from city agencies to envision the system. Denver leadership at the level of the Mayor and City Council should be recognized for being able to foresee the value that public bicycle sharing could add to the city. !$#'#1'"*0#(&H&(7#,&')(5#',5#/3(*05#-&0*"*3,#$%'$#']]&0$"#*,])'"$)20$2)&#%'"# -36,"$)&'.#%&'($%#*./'0$"7#1&0'2"&#/3(*05#-&0*"*3,"#$%'$# "%'/& #&(&.&,$"#3]#$%&# 1 2*($#&,H*)3,.&,$#"21"&J2&,$(5#']]&0$#$%&%'H*3)#3]#)&"*-&,$#/3/2('$*3,"9#F%&# %&'($%#03"$"#'""30*'$&-#6*$%#$)',"/3)$'$*3,#/3(*05#')&# *,0)&'"*,4(5 # )&0&*H*,4# .3)& # '$$&,$*3,7#'"#$%&#-*)&0$#(*,+"#1&$6&&,#/3(*0*&"#',-#32$03.&"#')&#J2',$*]*&-# Z!.&)*0',#B21(*0#T&'($%#!""30*'$*3,#<8;8[ 9#K*$*&"#')&#$'+*,4#,3$*0]#%36# *,])'"$)20$2)&#/3(*05#*"#*,&O$)*0'1(5#(*,+&-#$3#%&'($%#/3(*057#',-#')&#(3 3+*,4#$3#'0$*H&# $)',"/3)$'$*3,#'"#'#.&',"#$3#*./)3H&#/21(*0#%&'($%# Z!.&)*0',#B 21(*0#T&'($%# !""30*'$*3,#<88^[ 9#U$#*"#*,0)&'"*,4(5#*./3)$',$#$%'$#*,])'"$)20$2)&#/3(*0*&"#$'+&#*,$3# '0032,$#/3""*1(&#)&"2($',$#%&'($%#&]]&0$"7#6%&$%&)#3)#,3$#"20%#/3(*0*&"#')&#/')$#3]#',# *,$&,$*3,'(#*,$&)H&,$*3,9#F%&#'-H&,$#3]#G&,H&)#I ` 050('"#$%&#)&" 2($#3]#03,"0*32"# &]]3)$"#$3#)&"$)20$2)&#$% *($#&,H*)3,.&,$#$%)324%#$%&#*,$)3-20$*3,#3]#'#,&6# 03./3,&,$#6*$%#',#&O/)&""&-#43'(#$3#*,-20&#-&"*)'1(&#%&'($%#32$03.&"9# #

PAGE 231

216 It is valuable to understand the emergence of Denver's current trend toward support of transportation bicycling before exploring the future trajectory. Denver B cycle started having an impact on the policy and infrastructure of Denver before it was officially launched. G&,H&)#I ` 050(&#%'"#%'-#'#2,*J2&#-&H&(3/.&,$#/)30&""g#$%&# +&),&(#3]#$%&#/(',#1&4',#'"#'#$&./3)')5#/)3o&0$#$%'$#%'"#"*,0&#&H3(H&-#*,$3#(3,4&) ` $&).#/3(*05#',-#*,])'"$)20$2)�%',4&"9# Much of the recent efforts to increase transportation bicycling in Denver, i ncluding the implementation of Denver B cycle itself, can be traced to the lead up to the Denver 2008 Democratic National Convention (DNC). In preparation for the DNC, plans were developed for a temporary bike sharing system for residents and visitors to use during the event, inspired by bike sharing systems in Europe, such as VÂŽlib' in Paris, as well as the work of the GreenPrint Denver initiative through the Mayor's Office. To support the temporary bike sharing system, the city undertook an ambitious cam paign to improve on street bicycle facilities. Many of these improvements focused on street markings downtown, including the installation of the first shared lane "sharrow" markings in the city. The work leading up to the convention soon evolved into a lo nger term strategy to support bicycling in the city. During a speech to mark the opening of the temporary bike sharing system, Mayor John Hickenlooper shaped a vision for how the city might be by challenging Denver to a reach a goal of 10% bicycle commute mode share in ten years (Gerig 2008) The temporary bike sharing system proved to be popular, and following the convention, lessons learned from the experience emerged. Plans for a permanent publi c bicycle sharing system, began to take shape as policies and infrastructure changes

PAGE 232

217 necessary to make a permanent system viable were identified (Duvall 2008) Since then, the City of Denver has undertaken a dramatic upswing in projects to improve bicycle infrastructure and to adopt policies to encourage active transportation. In 2009, inspired by Mayor Hickenloope r's challenge and the desire for improved bicycle infrastructure in the community, the Denver Department of Public Works introduced a position for a dedicated bicycle and pedestrian planner. The position was to coordinate the implementation of a growing nu mber of active transportation projects. Emily Snyder, hired as the city's new planner, paid special attention to on street improvements in areas to be served by Denver B cycle, as well as making improvements to the connectivity of bike lanes and routes of access to the central downtown area. During the same timeframe, Denver began several planning projects to modify the policies and procedures with which city agencies managed traffic. One of the major outcomes of this effort was the Denver Moves plan, desi gned to lead the city toward development of a more sustainable transportation network by strengthening support for lifestyles inclusive of bicycling, walking and transit. Development of the plan included a strong community engagement component, so that res idents had the opportunity to become directly involved with blueprints for the future. In parallel with these efforts, the Denver Strategic Transportation Plan adopted the strategy "to move people, not just cars" (City and County of Denver 2011) evidence of an important realignment away from former, more car centric policy models. Going forward, plans to further support active transportation in the city are long term and e xtensive in scope, as outlined in the Denver Moves plan. Planning for the predecessor of Denver B cycle helped to set in motion changes to policy that continue to

PAGE 233

218 affect infrastructure and, in turn, are incrementally but measurably changing the active tran sportation behavior of the population in the city. With expansion, Denver B cycle will further serve to reinforce social normalization of transportation bicycling, as a more sustainable transportation network in the city is assemble d. Like many cities, De nver has produced plans to address challenges inherent in urban systems. Denver has a climate action plan to outline objectives to reduce petroleum dependence and resultant carbon, particulate and pollutant emissions. Similarly, Denver has plans to improve population health through improvements to livability and quality of life in its neighborhoods, in part, by support of active transportation. Denver B cycle helps to achieve these and other goals set in both the climate action and population health arenas simultaneously, through the reduction of car use. These are compelling reasons that leadership of many other cities have also identified, as evidenced by the mounting interest in public bicycle sharing systems blossoms in the U.S. (Kisner 2011) Since Denver B cycle started, several major cities including New York, San Francisco, Boston, Chicago, Washington, D.C., Portland an d Minneapolis, among others, have put in place public bike sharing systems or have definite plans for the implementation of systems. As each system comes online, the collective knowledge can help to inform the development of subsequent systems. As a pionee r in this group, findings gleaned through study of Denver contribute to the body of knowledge. Theoretical Implications The results of this research support the view that behavioral actions are resultant from complex social and environmental influences. T he effects of these influences appear to be especially acute when individuals weigh lifestyle fit when considering active

PAGE 234

219 transportation among possible mode choices. As an intervention, Denver B cycle was designed to deliberately interject an attractive ac tive transportation option into the social and environmental sphere in which decisions are made. In certain settings and for trips of certain distances or destinations, shared bikes were positioned as being among the most convenient choices, thus influenci ng choice outcomes. At the level of the individual, theoretical constructs embodied in social cognitive theory (SCT) are supported by findings of the present research, especially on the contributions of interactions between the individual and small group l evels as determinants of behavioral action. Previous investigation has found links between the effects of social support and the achievement of self efficacy in the context of physical activity intervention (Rovniak, Anderson et al. 2002) The present research provides further evidence to the value of social interaction in creating an environment in which the use of shared bikes is socially supported at t he individual and small group levels, contributing to positive feedback that reinforces positive behaviors. The SCT construct of reciprocal determinism was also identified among Denver B cycle users, in that they recognized the importance of their roles i n modeling behavior, in reporting problems with the system, and in assisting other users or potential users in the operation of the equipment. Reciprocal determinism aids in the encouragement of maintenance of behavior change by participants mutually suppo rting the continuance of changed behaviors through structural support (Dube and Stanton 2010) From a user standpoint, the smooth functioning of Denver B cycle included the availability of both shared bikes in good repair and open parking docks. Regular user s also had a vested interest in the continued growth and popularity of the system to ensure that it would

PAGE 235

220 continue to function. As evidence, many annual members recounted actively reporting bicycles with mechanical issues, as well as empty or full stations to Denver Bike Sharing, They also informed nascent users of the details of use of the system. Annual members who were regular users clearly understood the advantages of increasing the functional efficiency and use of the system in order to improve their o wn experiences. Some SCT constructs, when expanded to the population level, are visible as elements of diffusion of innovations (DI) theory. In a study of determinants of effective implementation of worksite based health interventions, linkages between ind ividual level cognitive aspects of behavior adoption, and group level behavioral implementation influenced by social interaction have been identified (Weiner, Lewis et al. 2009) These links indicate that behavioral decisions made on an individual level, when observed by other individuals have an impact on the behaviors of those observers, thus contributing to diffusion of an intervention to groups of individuals. In evidence through the pres ent research, the observation of behavior of participants by non participants, and the social interaction within and between individuals and groups in the population, appeared to be major conduits through which knowledge and information supportive of the i ntervention passed. Social interaction between Denver B cycle users, their peers, and the public at large served to diffuse participation in the intervention to a widening population. A gap in knowledge exists between theory based science and practical im plementation. Bridging this gap can assist in informing policy as to the manner in which interventions may be most effectively disseminated within a population. Literature has revealed that findings of evidence based research sometimes do not reach policy makers preventive health practitioners, and the public to inform health policy and to

PAGE 236

221 reach the greatest population at the lowest cost (Green, Ottoson et al. 2009) The present research contribut es to the evidence base of the effectiveness of an applied intervention in reaching a large audience, in part using participants to spread the word. Findings support DI constructs of how the initial uptake of behaviors occur within smaller groups of influe ntial popular opinion leaders and are dispersed to larger groups, eventually leading to uptake of behaviors within the general population. A study of the diffusion of smartphone adoption, a technology product, found that the establishment of a positive rep utation through the experiences of innovators and early adopters bolstered acceptance of uptake by potential adopters (Yoo, Yoon et al. 2010) In the case of Denver B cycle, which building an image as a te chnology empowered product through smartphone and web integrated tools the impact of its reputation among popular opinion leaders also appears to have positively affected promotion of the intervention. Likewise, the diffusion of the use of Denver B cycle seems to have some similarity to the diffusion of behaviors through online social networks, in that individual popular opinion leaders influence clusters of adopters, and reinforcement of behaviors from multiple peers appear to make participants more willi ng to adopt behaviors (Centola 2010) Denver B cycle users had an online presence visible in many so cial media outlets, encouraged by the online social media enabling components of the Denver B cycle website. It is therefore possible that the intervention functioned, in the context of diffusion, as a hybrid intervention, in part online and in part within the built environment at least for some users who take advantage of the smartphone and web integrated tools Just as social influence is a determinant of the diffusion of user generated online content (Bakshy, Karrer et al. 2009) user generated aspects of the use of Denver B cycle, such as

PAGE 237

222 the Tour de B cycle and other socially encouraged use, facilitated dissemination of use within the population. Those committed individuals who became champions of Denver B cycle at an early stage are at the epicenter of apparent population scale change. A committed minority of as few as 10% of a population engaging in an activity can indicate a tipping point in the behavior change of the population as a whole (Xie, Sreenivasan et al. 2011) The present research found that a number of individuals equivalent to 10% of the commuting population of Denve r tried Denver B cycle in the initial year, approximating the tipping point number of the Xie et al. study. It is unlikely, however, that the total number of those exposed to Denver B cycle in the initial year is comprised of only the truly committed and f ervent users, yet it is an indication of a broad general appeal of Denver B cycle. Although the population of Denver who choose to commute on bicycles is on the rise, achieving a consistent 10% commuting share remains some distance away. However, as Denver B cycle becomes more established and the core of committed users grows, diffusion literature suggests that a tipping point of the normalization of a behavior is possible if a 10% adoption rate can be achieved (Rogers 2003) It is in the detection of population scale behavior change that the theory of complex adaptive systems comes into play. The findings of this research suggest that users of Denver B cycle were affected by physical and s ocial elements of the urban system as reflected by their behavioral reactions to stimuli in the built environment, supporting systems theory related ecological approaches to physical activity intervention (Sallis, Cervero et al. 2006) On the other hand, the behavioral reactions of Denver B cy cle users exhibited the power of collective agency, contributing to systemic responses

PAGE 238

223 reflected in modifications to the urban system, as indicated by emerging policy and infrastructural changes supportive of transportation bicycling. In all, Denver B cyc le contributed to a rising tide of bicycle use in the city, in part reflected in the increased ACS bicycle commuting figures. A total increase in the number of bicyclists on the street may also strengthen bicycling activity through the perception of safety in numbers. As with other activities that become more attractive or more socially accepted with an increasing number of participants, users of Denver B cycle who are observed in the public riding shared bikes on the street, on trails and in other public s ettings, may inspire others to ride bicycles, shared or otherwise. Results of the research indicate that multiple levels of theory framed and informed the detection of changes in behavior observable at multiple levels. However, complex adaptive systems th eory encapsulated the contributions of other theoretical perspectives from different levels. A cycle of population scale change begins when the total volume of bicyclists grows, and other bicyclists are tacitly encouraged to join them as social normalizati on adapts. With higher bicycle volume, all street users become more aware of the presence of bicyclists, and more attentive of them in traffic. With a certain level of activity, the parameters of the built environment are pressured to adapt to meet changin g behaviors. More bicyclists create demand for bicycle supportive policy and infrastructure, which leads to more bicyclists creating a virtuous cycle The result is a changed behavior of a population with a sustainable trajectory, a recommended goal of co mmunity based active transportation interventions (Centers for Disease Control and Prevention 2010) The changes to infrastructure, policy, culture, and society within a city necessary to make public bicycle shari ng viable in a city may serve as a catalyst to set in motion more

PAGE 239

224 encouraging conditions for all bicyclists. As a final note on theory, a finding with important implications to any theory of behavior change i s that individuals who took part in the interve ntion self selected to participate, and developed their own motivational reasoning to initiate and maintain changed behavior. Health benefits from participation were achieved without external coaching, counseling or tailoring by an intervention administrat or. This is evidence that an intervention designed with an intent to have an impact on the behavior of thousands of individuals can be effective in inducing sustained behavior change, and can be accomplished without a need to overtly dictate motivational o bjectives, such as weight loss, to the targeted population. The results of the study indicate that, from a standpoint of applied theory, there is room for development of theory descriptive of what may be referred to as stealth interventions, in which memb ers of the targeted population are given control to set their own goals and to derive their own individualized motivation to meet them. The idea of stealth as a guiding framework for health intervention may be informative of efforts to develop theory for b ehavior change at the population level, an area of present applied theory that is noticeably sparse. Opportunities For Future Research Continued and future research is necessary to further elucidate the impact of public bicycle sharing systems on users an d the communities in which they are implemented. There is much yet to be known about many aspects of how public bicycle sharing systems shape and are shaped by the behaviors of individuals, groups and populations. As metropolitan areas become home to an in creasing percentage of the

PAGE 240

225 world population, economic factors and competition for limited resources will oblige community leaders to rethink current policies and adapt to a changing environment. The potential for public bicycle sharing to address issues as sociated with many current and emerging challenges ensures that, as a tool to affect health, economics, petroleum dependence, and climate change, shared bikes will remain a topic of interest for some time. Three major categories of recommended inquiry are presented and discussed in the following. B arriers to use of public bicycle sharing : T he present research identified differences in participation among some demographic subgroups, including minorities, women and those at lower income levels. Further resear ch is necessary to isolate specific barriers to participation among these groups in order to develop strategies for mitigation. This effort may return the greatest benefits within low income populations, as on a per use basis, shared bikes can be among the least expensive public transportation options, and provide access to a greater variety of food sources and employment options. Reducing the barriers of access to shared bikes can work toward reducing overall health disparities within the population. In a ddition to identifying socioeconomic barriers to use, shared bikes can be of value in helping to identify physical barriers to use. A fleet of GPS enabled shared bicycles would be greatly valuable to broadly collect geographic data of bicycle use in a comm unity. Aggregated route mapping can identify how specific infrastructural elements support or act as barriers, to active transportation behavior. Determining if users avoid specific roads or areas, and comparing route tracks of shared bikes with patterns o bserved in other transportation modes could assist in informing planners and policy makers how

PAGE 241

226 to reduce conflicts and improve safety. Expanded measures of the health effects of public bicycle sharing : The results of the present research pointed to positi ve health effects among users, which in part may be attributed to the use of Denver B cycle. However the understanding of these effects would benefit from formalized clinical measures of individuals who use the system. Likewise, more in depth assessment of the mental health effects of participation could be helpful in understanding the impact of shared bikes on work productivity, potentially important in encouraging employer based support for membership. A key item for future research is the refinement of methodology to estimate the reduction of car use that might be attributable to the use of shared bikes. A tool designed to randomly select rides for which a short survey could be administered immediately after returning a shared bike could help to hone in on the details of car replacement by shared bikes. Determining the details of car trips replaced, including distances, destinations, time of day, day of the week and date, as well as trip purpose and membership type would enable the construction of much mo re robust models to calculate the car trip replacement factor of public bike sharing. Measuring effects of public bicycle sharing within urban systems : The present research examined potential effects that Denver B cycle had on the bicycle commuting behavi or within the city. However, available data were not adequate to definitively establish any connection between the introduction of shared bikes and the overall increase in the rate of bicycle commuting in the city. Understanding the effects that public bic ycle sharing has on the total active transportation behavior within a community is of value for future research. The development of tools to regularly track and record all

PAGE 242

227 bicycle transportation in a city could greatly help in determining the impact of the use of shared bikes on total bike use, and therefore identify indirect health benefits attributable to public bicycle sharing. Another item that merits future research is the economic effects of public bicycle sharing on the community it serves. A detaile d examination of economic impacts, not only for individual users, but also for non users, businesses near stations, and for the community as a whole would be of great value for policy makers considering the implementation of a public bike sharing system. Data from the present study revealed that shared bike users choose destinations in part based on access to stations, near which they spend money and contribute to local commerce. As an intervention in which infrastructural elements must be secured with fun ding, sometimes outside of traditional infrastructure funding streams, public bicycle sharing is dependent on business sponsorships, public or private grants, and donations. Identifying the economic impact that bicycle sharing has on communities is of valu e in aiding its expanded implementation as a public health intervention. Related to the item discussed earlier in this section regarding improved methodology for estimating car use replacement, is the need for improved methodology to track and account the climate change mitigation effects of public bicycle sharing systems. The replacement of car use is only one factor in this subject. Others include how multimodal transit trips enabled by access to shared bicycles further reduce car use, how reduced traffic congestion affects wasted fuel due to idling time, how reduced road wear affects resource expenditures on materials, and how lifestyle changes reduce demand for fossil fuels.

PAGE 243

228 Study Limitations The present research was limited in data collection to the i nitial year of execution of an intervention, which presented what was likely an entirely new concept to many people within a large population. The focus of the investigation was on annual members, who are likely not typical of the total group of people who used Denver B cycle at least once during the year, and who are also not likely to typify the general adult population of the city. However, as the study was designed to only examine effects of use among those who had actually used Denver B cycle and not t he general population, the findings should remain valid in reference to the group of focus. Denver B cycle users are a group who self selection to use the system. The Denver B cycle system usage dataset included all usage data from all users, so self selec tion bias was not an issue among the total group of users. However, the Denver B cycle user survey dataset was dependent on responses from a subgroup within the group of all users, also on a self selected basis. It is therefore possible that some findings are subject to self selection bias. Strategies to reduce the effects of this potential for bias have been built into the study through the use of mixed methods and multiple data sources whenever practical. A mixed methods approach helps to mitigate the eff ects of bias by cross validating findings between quantitative and qualitative methods of inquiry (Neuman 2003) Another limitation was that it was not possible to collect both pre and post data on the participants to assess existing levels of active transportation behavior. Setting a baseline of active transportation activity followed by a posttest evaluation of observed difference would have been infor mative in determining relative change observed that

PAGE 244

229 might be attributable to the use of Denver B cycle. Instead the focus was set on evaluating any net increase in active transportation behavior that may be directly attributable to use of shared bikes. Thi s was accomplished by calculating the estimated quantity of car use that was replaced by shared bike use as a net increase over existing levels. This is justifiable, as any outcome in which car use is replaced by active transportation serving the same tran sportation purpose can be nothing other than a net increase over any prior activity level, representing a net increase in active transportation The quantitative portion of this study relies on some data that were not collected for research purposes, and are thus subject to both known and unknown errors in translation to research application. The Denver B cycle system usage dataset was extremely extensive, and produced by software that was undergoing development and modification on an ongoing basis through out the data collection period. Some data items required checking and adjustment in order to be useable for research. As there were no alternatives than to use these data, they were carefully reviewed for any obvious errors. Totals and results derived from this dataset were cross validated with other observations for plausibility. Another limitation of this dataset was that demographic data were not collected for all users during registration. Additionally, some data used in the study were dependent on the self reporting of participants. Self reported data can be biased in a direction favorable to the participant making the report. Such self reporting bias is possible in the Denver B cycle user survey dataset. However, not all measures and activities necessa ry for the present study could be practicably collected through means other than through a survey, in which a certain amount of self reporting bias is unavoidable. Totals and results from the Denver B cycle

PAGE 245

230 user survey dataset were cross validated with oth er observations when possible. The limitation of data collection to the initial year of operation reduced the ability to detect any observable continuity of changed behavior among individual participants after the winter hiatus, during which Denver B cyc le was closed. It is possible that some participants reverted to prior behaviors with less favorable outcomes. The scope of the study did not allow for the evaluation of data from the second year of operation. Conclusions As an intervention to affect acti ve transportation Denver B cycle departs from traditional approaches in a number of ways. Instead of concentrating on a specific target group with shared attributes, it is applied through an element of infrastructure, targeting a broad swath of the genera l population. The central concept of the intervention is that individual participants derive their own individualized motivational reasoning for the initiation and continuity of behavioral action. Therefore, the intervention is motivationally self supporte d, and participants develop tools to self regulate their behaviors, both suggesting viability and independence in the long term maintenance of changed behaviors. It is an intervention in which participants do not merely fill a passive role as the subjects of treatment; rather, they are an integral piece of the operation of the intervention itself. People are given the freedom to determine for themselves the level and manner of involvement that fit within their lifestyle. After the initial year of the intervention, evidence points to success in achieving a sustained net increase in active transportation among annual members. Other evidence indicates the possibility of broader impacts by inducing positive change in the city,

PAGE 246

231 including emerging social normalization of transportation bicycling, incremental modification of the built environment, and increase d active transportation activity within the general population. As a strategy for catalyzing behavior change, public bicycle sharing can be an effective multi level intervention to encourage active transportation at a population scale.

PAGE 247

232 REFERENCES Alliance for Biking & Walking (2012). Bicycling and Walking in the United States: 2012 Benchmarking Report. K. Swanson. Washington, D.C. : 248. American Public Health A ssociation (2009). At the Intersection of Public Health and Transportation: Promoting Healthy Transportation Policy. Washington, D.C., American Public Health Association. American Public Health Association (2010). The Hidden Health Costs of Transportation. Washington, D.C., American Public Health Association. Babey, S. H., T. A. Hastert, et al. (2009). "Sociodemographic, Family, and Environmental Factors Associated with Active Commuting to School among US Adolescents." Journal of Public Health Policy 30 (S1) : S203 S220. Bachand Marleau, J., J. Larsen, et al. (2011). "The much anticipated marriage of cycling and transit: But how will it work?" Transportation Research Record 2247 : 109 117. Bakshy, E., B. Karrer, et al. (2009). Social Inuence and the Diffusion of User Created Content. Proceedings of the 10th ACM conference on Electronic commerce New York, NY. Bandura, A. (1989). "Human agency in social cognitive theory." The American Psychologist 44 (9): 1175 1184. Bassett, D. R., H. R. Wyatt, et al. (2010). "Pe dometer Measured Physical Activity and Health Behaviors in U.S. Adults." Medicine & Science in Sports & Exercise 42 (10): 1819 1825. Beech, B. M., M. L. Fitzgibbon, et al. (2011). "The impact of Socioeconomic Factors and the Built Environment on Childhood a nd Adolescent Obesity." Childhood Obesity 7 (1): 19 24. Bernard, H. R. (2011). Research Methods in Anthropology: Qualitative and Quantitative Approaches Lanham, Maryland, AltaMira Press. Berrigan, D. and R. P. Troiano (2002). "The Association Between Urban Form and Physical Activity in U.S. Adults." Am J Prev Med 23 (2S): 74 79. Berrigan, D., R. P. Troiano, et al. (2001). "Active Transportation Increases Adherence to Activity Recommendations." Am J Prev Med 31 (3): 210 216.

PAGE 248

233 Besser, L. M. and A. L. Dannenberg (2005). "Walking to Public Transit: Steps to Help Meet Physical Activity Recommendations." American Journal of Preventive Medicine 29 (4): 273 280. Black, K. J. (2005). Evaluating the Effectiveness of an Intervention Promoting Walking and Lifestyle Physical Activity in a Primary Care Practice. Health and Behavioral Sciences Denver, Colorado, University of Colorado at Denver. Doctor of Philosophy: 230. Bohle, H. G., B. Etzold, et al. (2009). "Resilience as Agency." IHDP Update 2009 (2): 8 13. Centers for Dise ase Control and Prevention (2005). "Trends in Leisure Time Physical Inactivity by Age, Sex, and Race/Ethnicity United States, 1994 2004." Morbidity and Mortality Weekly Report 54 (39): 991 994. Centers for Disease Control and Prevention (2010). The CDC Gu ide to Strategies for Increasing Physical Activity in the Community. Physical Activity Guide Atlanta, GA, Department of Health and Human Services : 39. Centers for Disease Control and Prevention. (2011). "2010 Denver County, CO: Overweight and Obesity (B MI): Weight classification by Body Mass Index (BMI)." SMART: Selected Metropolitan/Micropolitan Area Risk Trends from http://apps.ncc d.cdc.gov/brfss smart/MMSACtyRiskChart.asp?yr2=2010&CtyCode=9&cat=OB&MMSA=24&q key=4409&grp=0 Centers for Disease Control and Prevention. (2011). "How is BMI calculated and interpreted?", 2011, from http://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html#Interpret ed Centers for Disease Control and Prevention (CDC) (2009). 2010 Behavioral Risk Factor Surveillance System Survey Questionnaire. U.S. D epartment of Health and Human Services. Atlanta, GA, Centers for Disease Control and Prevention. Centola, D. (2010). "The Spread of Behavior in an Online Social Network Experiment Science 329 : 1194 1197. Cervero, R. (1995). "Mixed land uses and commuting : Evidence from the American Housing Survey." Transportation Research Part A 30 (5): 361 377. Cervero, R. and M. Duncan (2003). "Walking, bicycling, and urban landscapes: evidence from the San Francisco Bay Area." Am J Public Health 93 (9): 1478 83. Cervero, R. and C. Radisch (1996). "Travel choices in pedestrian versus automobile oriented neighborhoods Transport Policy 3 (3): 127 141.

PAGE 249

234 City and County of Denver (2011). Denver Moves: Making Bicycle and Multi Use Connections. Denver, CO, City and Count y of Denver : 45. Cohen, D. A., D. Golinelli, et al. (2009). "Effects of Park Improvements on Park Use and Physical Activity." Am J Prev Med 37 (6): 475 480. Colorado Department of Public Health and Environment. (2011). "Colorado Health Information Dataset." Retrieved Nov. 4, 2011, 2011, from http://www.chd.dphe.state.co.us/cohid/topics.aspx?q=Behavioral_Risk_Factors Currie, G. and J. Phung (2008). "Understanding Link s Between Transit Ridership and Gasoline Prices Evidence from the United States and Australia." Transportation Research Record 2063 : 133 142. DeMaio, P. J. (2003). "Smart bikes: Public transportation for the 21st century." Transportation Quarterly 57 (1): 9 11. DeMaio, P. J. and J. Gifford (2004). "Will smart bikes succeed as public transportation in the United States?" Journal of Public Transportation 7 (2): 1 15. Denver Bike Sharing (2011). December Newsletter: 2011 Season Recap. Denver, Colorado. DenverGIS (2011). "Denver GIS Layers." Retrieved April 13, 2011, 2011, from http://www.denvergov.org/Default.aspx?alias=www.denvergov.org/GIS Dill, J. (2009). "Bicycling for Transpo rtation and Health: The Role of Infrastructure." Journal of Public Health Policy 30 (SI): S95 S110. Dill, J. and T. Carr (2003). "Bicycle Commuting and Facilities in Major U.S. Cities: If You Build Them, Commuters Will Use Them." Transportation Research Rec ord 1828 : 116 123. Dishman, R. K., R. W. Motl, et al. (2004). "Self efficacy partially mediates the effect of a school based physical activity intervention among adolescent girls." Prev Med 38 (2004): 628 636. Downtown Denver Partnership (2011). Downtown De nver Commuter Survey 2011. Denver, Colorado. Dube, A. R. and C. A. Stanton (2010). The Social Context of Dietary Behaviors: The Role of Social Relationships and Support on Dietary Fat and Fiber Intake. Modern Dietary Fat Intakes in Disease Promotion Nutri tion and Health : 31 42. Dunn, A. L. (2009). "Effectiveness of Lifestyle Physical Activity Interventions to Reduce Cardiovascular Disease." American Journal of Lifestyle Medicine 3 (1): 11S to 18S.

PAGE 250

235 Dutton, G. R., B. C. Provost, et al. (2008). "A tailored pri nt based physical activity intervention for patients with type 2 diabetes." Preventive Medicine 47 : 409 411. Duvall, A. (2008). Democratic National Convention in Denver Freewheelin Bike Sharing System: Summary Report. Denver, CO, Office of the Mayor, City and County of Denver : 3. Elder, J. P., L. Lytle, et al. (2007). "A description of the social ecological framework used in the trial of activity for adolescent girls (TAAG)." Health Education Research 22 (2): 155 165. Ewing, R., R. C. Brownson, et al. (2006) "Relationship between urban sprawl and weight of United States youth." Am J Prev Med 31 (6): 464 74. Ewing, R. and R. Cervero (2001). "Travel and the Built Environment: A Synthesis." Transportation Research Record 1780 : 87 114. Ewing, R., T. Schmid, et al (2003). "Relationship Between Urban Sprawl and Physical Activity, Obesity, and Morbidity." American Journal of Health Promotion 18 (1): 47 57. Fieschi, C. and P. Heywood (2004). "Trust, cynicism and pop ulist anti politics." Journal of Political Ideologies 9 (3). Flusche, D. (2009). American Community Survey Bicycle Commuting Trends, 2000 to 2008, League of American Bicyclists : 1 7. Frank, L. D., M. A. Andresen, et al. (2004). "Obesity relationships with community design, physical activity, and time spent in cars." Am J Prev Med 27 (2): 87 96. Frank, L. D., B. E. Saelens, et al. (2007). "Stepping towards causation: do built environments or neighborhood and travel preferences explain physical activity, driving, and obesity?" Soc Sci Med 65 (9): 1898 914. Garrard J., G. Rose, et al. (2007). "Promoting transportation cycling for women: The role of bicycle infrastructure." Prev Med Geller, R. (2009). Four Types of Cyclists. Portland, Oregon, Portland Office of Transportation. Gerig, B. (2008). To Be Bike Elite, We Can't Talk The Talk, We Must Climb On And Pedal The Path. Washington Park Profile Denver, Colorado. Givoni, M. and P. Rietveld (2007). "The access journey to the railway station and its role in passengers' satisfaction with rail travel." Transport Policy 14 : 357 365.

PAGE 251

236 Glanz, K. (2009). "Measuring Food Environments: A Historical Perspective." Am J Prev Med 2009 (36 (4S)): S93 S98. Green, L. W., J. M. Ottoson, et al. (2009). "Diffusion Theory and Knowledge Dissemination, Utilization, and Integration in Public Health." Annual Review of Public Health 30 : 151 174. Gruber, K. J. and L. A. Haldeman (2009). "Using the Family to Combat Childhood and Adult Obesity." Preventing Chronic Disease 6 (3). Handy, S. L., M. G. Boarnet, et al. (2002). "How the built environment affects physical activity." Am J Prev Med 23 (2S): 64 73. Hartog, J. J. d., H. Boogaard, et al. (2010). "Do the Health Benefits of Cycling Outweigh the Risks?" Environmental Health Perspectives 118 (8): 1109 1116. Haskell, W., I. M. Lee, et al. (2007). "Phy sical Activity and Public Health. Updated Recommendation for Adults From the American College of Sports Medicine and the American Heart Association." Circulation 116 : 000 000. Henrich, J. (2001). "Cultural Transmission and the Diffusion of Innovations: Ado ption Dynamics Indicate That Biased Cultural Transmission Is the Predominate Force in Behavioral Change." American Anthropologist 31 (4): 992 1013. Jacobson, S. H., D. M. King, et al. (2011). "A note on the relationship between obesity and driving." Transpo rtation Policy 03 (008): 1 5. Jones, N. F., C. Rutt, et al. (2006). "The development of a data model for research on the environmental correlates of physical inactivity and obesity." AMIA Annu Symp Proc : 970. Kay, J. J. (2008). An Introduction to Systems Th inking. The Ecosystem Approach: Complexity, Uncertainty, and Managing for Sustainability D. Waltner Toews, J. J. Kay and N. M. E. Lister. Chichester, West Sussex, New York, Columbia University Press : 3 13. Kenworthy, J. and G. Hu (2002). "Transport and ur ban form in Chinese cities: An international comparative and policy perspective with implications for sustainable urban transport in China." Keyserling, T. C., C. D. S. Hodge, et al. (2008). "Randomized trial of a clinic based, community supported, lifesty le intervention to improve physical activity and diet: The North Carolina enhanced WISEWOMAN project." Preventive Medicine 46 : 499 510.

PAGE 252

237 Kisner, C. (2011). Integrating Bike Share Programs into a Sustainable Transportation System. Washington, D.C., National League of Cities Center for Research and Innovation. Knigge, L. and M. Cope (2009). Grounded visualization and scale: a recursive analysis of community spaces. Qualitative GIS: A Mixed Methods Approach M. Cope and S. Elwood. London, SAGE Publications Ltd : 95 114. Koetse, M. J. and P. Rietveld (2008). "The impact of climate change and weather on transport: An overview of empirical findings." Transportation Research Part D: Transport and Environment 14 (3): 205 221. Krieger, J., J. Rabkin, et al. (2009). "Hig h Point Walking for Health: Creating Built and Social Environments That Support Walking in a Public Housing Community." American Journal of Public Health 99 (S3): S593 S599. Lane, B. W. (2009). "The relationship between recent gasoline price uctuations and transit ridership in major US cities." Journal of Transport Geography 04 (002). Lee, I. M., R. Ewing, et al. (2009). "The Built Environment and Physical Activity Levels: The Harvard Alumni Health Study." American Journal of Preventive Medicine 37 (4): 293 298. Levi, J., S. Vinter, et al. (2010). F as in Fat: How Obesity Threatens America's Future. Washington, D.C., Trust for America's Health : 124. Lewis, B. A., B. H. Marcus, et al. (2002). "Psychosocial Mediators of Physical Activity Behavior Among Adul ts and Children." Am J Prev Med 23 (2S): 26 35. Link, B. G. and J. Phelan (1995). "Social Conditions as Fundamental Causes of Disease." Journal of Health and Social Behavior 1995 (Extra Issue): 80 94. Litman, T. (2004). "Transit Price Elasticities and Cross Elasticities." Journal of Public Transportation 7 (2): 37 58. Lopez Zetina, J., H. Lee, et al. (2006). "The link between obesity and the built environment. Evidence from an ecological analysis of obesity and vehicle miles of travel in California." Health & Place 12 : 656 664. Lucas, K. (2009). "Actual and Perceived Car Dependence." Transportation Research Record 2118 (2009): 8 15. MacKerron, G. and S. Mourato (2009). "Life satisfaction and air quality in London." Ecological Economics 68 (2009): 1441 1453. March etti, C. (1994). "Anthropological Invariants in Travel Behavior." Technical Forecasting and Social Change 47 (1): 75 78.

PAGE 253

238 Martens, K. (2004). "The bicycle as a feedering mode: experiences from three European countries." Transportation Research Part D 9 : 281 294. Martens, K. (2007). "Promoting bike and ride: The Dutch experience." Transportation Research Part A 41 : 326 338. Moreno, I. (2010). LaHood rides bike in Denver to tout share program. The Denver Post Denver, Colorado, The Denver Post. Mostafa, M. M. (2011). "Does globalisation affect consumers' pro environmental intentions? A multilevel analysis across 25 countries." International Journal of Sustainable Development & World Ecology 2011 Nazelle, A. d., M. J. Nieuwenhuijsen, et al. (2011 ). "Improving health through policies that promote active travel: A review of evidence to support integrated health impact assessment." Environment International 37 (2011): 766 777. Neuman, W. L. (2003). Social Research Methods: Qualitative and Quantitative Approaches Boston, MA, A and B. New York City Department of Health and Mental Hygiene (2009). Physical Activity in New York City: New Yorkers Can Easily Integrate Exercise into Their Daily Routine. NYC Vital Signs New York, NY, New York City Department of Health and Mental Hygiene. 8: 3: 1 4. Newman, P. and J. Kenworthy (2006). "Urban Design to Reduce Automobile Dependence." Opolis 2 (1): 35 52. Office of Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, et al. (2001). T he Surgeon General's Call To Action To Prevent and Decrease Overweight and Obesity 2001. U.S. Department of Health and Human Services. Rockville, MD : 1 39. Ogden, C. L. and M. D. Carroll (2010). Prevalence of Overweight, Obesity, and Extreme Obesity Among Adults: United States, Trends 1960 1962 Through 2007 2008. C. f. D. C. a. Prevention. Atlanta, GA, Centers for Disease Control and Prevention. Ogilvie, D., F. Bull, et al. (2011). "An applied ecological framework for evaluating infrastructure to promote wa lking and cycling: the iConnect Study." American Journal of Public Health 101 (3): 473 481. Oliver, L. N., N. Schuurman, et al. (2007). "Comparing circular and network buffers to examine the influence of land use on walking for leisure and errands." Interna tional Journal of Health Geographics 6 (41).

PAGE 254

239 Osher, C. N. (2010). Bike agenda spins cities toward U.N. control, Maes warns. The Denver Post Denver, Colorado. Pabayo, R., L. Gauvin, et al. (2010). "Sustained Active Transportation is associated with a favora ble body mass index trajectory across the early school years: Findings from the Quebec Longitudinal Study of Child Development birth cohort Preventive Medicine 2010 (50): 559 564. Pekmezi, D. W., C. J. Neighbors, et al. (2009). "A Culturally Adapted Physi cal Activity Intervention for Latinas." Am J Prev Med 37 (6): 495 500. Pendola, R. and S. Gen (2007). "BMI, auto use, and the urban environment in San Francisco." Health Place 13 (2): 551 6. Petosa, R. L., R. Suminski, et al. (2003). "Predicting Vigorous Phy sical Activity Using Social Cognitive Theory." Am J Health Behav 27 (4): 301 310. Pikora, T., B. Giles Corti, et al. (2003). "Developing a framework for assessment of the environmental determinants of walking and cycling." Soc Sci Med 56 (8): 1693 703. Princ e, S. A., M. S. Tremblay, et al. (2011). "Neighbourhood differences in objectively measured physical activity, sedentary time and body mass index." Open Journal of Preventive Medicine 1 (3): 182 189. Pucher, J. and R. Buehler (2006). "Why Canadians cycle more than Americans: A comparative analysis of bicycling trends and policies." Transport Policy 13 : 265 279. Pucher, J. and R. Buehler (2009). "Integrating Bicycling and Public Transport in North Ameri ca." Journal of Public Transportation 12 (3): 79 104. Pucher, J. and L. Dijkstra (2003). "Promoting safe walking and cycling to improve public health: lessons from The Netherlands and Germany." Am J Public Health 93 (9): 1509 16. Pucher, J., Z. R. Peng, et a l. (2007). "Urban transport trends and policies in China and India: Impacts of rapid economic growth." Transport Reviews 27 Rabl, A. and A. d. Nazelle (2012). "Benets of shift from car to active transport." Transport Policy 2012 (19): 121 131. Ragin, C. C. (1999). "Using Qualitative Comparative Analysis to Study Causal Complexity." Health Services Research 34 (5): 1225 1239.

PAGE 255

240 Regional Transportation District. (2011). "RTD Bus and Light Rail Station GIS layers." Retrieved April 13, 2011, 2011, from http://gis.rtd denver.com/MapServer/datadownload.aspx Reynolds, C. C. O., M. Winters, et al. (2010). Active Transportation in Urban Areas: Exploring Health Benefits and Risks Vancouver, BC, Nati onal Collaborating Center for Environmental Health. Richardson, A. J. (2000). Seasonal and weather impacts on urban cycling trips. TUTI Report 1 2000 Victoria, Australia, The Urban Transport Institute. Rietveld, P. (2000). "Non motorised modes in transpor t systems: a multimodal chain perspective for The Netherlands." Transportation Research Part D 5 : 31 36. Rihoux, B. (2006). "Qualitative Comparative Analysis (QCA) and Related Systematic Comparative Methods." International Sociology 21 (5): 679 706. Robertson Wilson, J. E., S. T. Leatherdale, et al. (2008). "Social Ecological Correlates of Active Commuting to School Among High School Students." Journal of Adolescent Health 42 : 486 495. Robinson, W. S. (1951). "The Logical Structure of Analytic Induc tion." American Sociological Review 16 (6): 812 818. Rogers, E. M. (2003). Diffusion of Innovations New York, The Free Press. Rohwer, G. (2011). "Qualitative Comparative Analysis: A Discussion of Interpretations." European Sociological Review 27 (6): 728 74 0. Rojas Rueda, D., A. d. Nazelle, et al. (2011). "The health risks and benefits of cycling in urban environments compared with car use: health impact assessment study." British Medical Journal 2011 (343). Roux, L., M. Pratt, et al. (2008). "Cost Effectiven ess of Community Based Physical Activity Interventions." Am J Prev Med 35 (6): 578 588. Rovniak, L. S., E. S. Anderson, et al. (2002). "Social Cognitive Determinants of Physical Activity in Young Adults: A Prospective Structural Equation Analysis." Annals o f Behavioral Medicine 24 (2): 149 156. Sallis, J. F., R. B. Cervero, et al. (2006). "An ecological approach to creating active living communities." Annu. Rev. Public Health 27 : 297 322. Sallis, J. F., L. D. Frank, et al. (2004). "Active transportation and p hysical activity: opportunities for collaboration on transportation and public health research." Transportation Research Part A 38 (2004): 249 268.

PAGE 256

241 Sallis, J. F. and N. Owen (2002). Ecological Models of Health Behavior. Health Behavior and Health Education K. Glanz, B. K. Rimer and F. M. Lewis. San Francisco, CA, John Wiley and Sons, Inc. Sallis, J. F., B. E. Saelens, et al. (2009). "Neighborhood built environment and income: Examining multiple health outcomes." Social Science and Medicine 68 (7): 1285 1293. Shaheen, S. A. and S. Guzman (2011). "Worldwide Bikesharing." Access 39 (Fall 2011): 22 27. Shaheen, S. A., S. Guzman, et al. (2010). "Bikesharing in Europe, the Americas, and Asia: Past, Present, and Future Transportation Research Record 2143 (2010): 159 167. Shaheen, S. A., H. Zhang, et al. (2011). "Hangzhou Public Bicycle: Understanding Early Adoption and Behavioral Response to Bikesharing in Hangzhou, China." Transportation Research Record 2247 : 34 41. Shannon, T., B. Giles Corti, et al. (2006). "Ac tive commuting in a university setting: Assessing commuting habits and potential for modal change." Transport Policy 13 (2006): 240 253. Shaw, S. L. and X. Xin (2003). "Integrated land use and transportation interaction: a temporal GIS exploratory data anal ysis approach." Journal of Transport Geography 11 (2): 103 115. Stadler, G., G. Oettingen, et al. (2009). "Physical Activity in Women: Effects of a Self Regulation Intervention." Am J Prev Med 36 (1): 29 34. Sternfeld, B., C. Block, et al. (2009). "Improving Diet and Physical Activity with ALIVE: A Worksite Randomized Trial." Am J Prev Med 36 (6): 475 483. Stralen, M. M. v., H. Vries, et al. (2009). "The working mechanisms of an environmentally tailored physical activity intervention for older adults: a random ized controlled trial." International Journal of Behavioral Nutrition and Physical Activity 6 (83). Thompson, N. J., D. Sleet, et al. (2002). "Increasing the use of bicycle helmets: lessons from behavioral science." Patient Education and Counseling 46 (2002) : 191 197. Tilahun, N. Y., D. M. Levinson, et al. (2007). "Trails, lanes, or tra c: Valuing bicycle facilities with an adaptive stated preference survey." Transportation Research Part A 41 : 287 301. U.S. Census Bureau. (2010). "S0801. Commuting Characteris tics by Sex, Denver County, Colorado." 2009 American Community Survey 1 Year Estimates

PAGE 257

242 Retrieved Jan. 28, 2011, 2011, from http://factfinder.census.gov/servlet/STTable?_bm=y& state=st& context=st& qr_name=ACS_2009_1YR_G00_S0801& ds_name=ACS_2009_1YR_G00_& tree_id=307& redoLog=false& geo_id=05000U S08031& format=& _lang=en U.S. Census Bureau (2010). The American Community Survey Questionnaire. U. S. D. o. Commerce. Washington, D.C., U.S Department of Commerce, Economics and Statistics Administration. U.S. Census Bureau (2011). B01001: Sex by Age, 2 010 American Community Survey 1 Year Estimates, Denver County, Colorado. U. S. C. Bureau. Washington, D.C., U.S Census Bureau. U.S. Census Bureau (2011). B19001: Household Income in the Past 12 Months (In 2010 Inflation Adjusted Dollars), Denver County, Co lorado. U. S. C. Bureau. Washington, D.C., U.S. Census Bureau. U.S. Census Bureau (2011). DP 1 Denver County, Colorado: Profile of General Population and Housing Characteristics: 2010. U. S. C. Bureau. Washington, D.C., U.S. Census Bureau. U.S. Census Bu reau (2011). S0801 Commuting Characteristics by Sex 2010 American Community Survey 1 Year Estimates (State of Colorado and Denver County). U. S. C. Bureau. Washington, D.C., U.S. Census Bureau. U.S. Census Bureau (2011). S0801: Commuting Characteristics by Sex, United States. U. S. C. Bureau. Washington, D.C., U.S. Census Bureau. U.S. Census Bureau (2011). S1501: Educational Attainment, Denver County, Colorado. U. S. C. Bureau. Washington, D.C., U.S. Census Bureau. U.S. Census Bureau. (2012). "American Fact Finder." Retrieved January 7, 2012, 2012, from http://factfinder2.census.gov/faces/nav/jsf/pages/index.xhtml UCLA: Academic Technology Services Statistical Consulting Group. (20 11). "Ordered logistic regression." Retrieved Aug. 1, 2011, from http://www.ats.ucla.edu/stat/stata/whatstat/whatstat.htm#orderedlogistic Van de Ven, A. H., D. E Polley, et al. (1999). The Innovation Journey New York, Oxford University Press. Wang, Y. and M. A. Beydoun (2007). "The obesity epidemic in the United States Gender, age, socioeconomic, racial/ethnic, and geographic characteristics: A systematic review and meta regression analysis." Epidemiol Rev 29 : 6 28.

PAGE 258

243 Wang, Y., M. A. Beydoun, et al. (2008). "Will All Americans Become Overweight or Obese? Estimating the Progression and Cost of the US Obesity Epidemic." Obesity 16 (10): 2323 2330. Warburton, D. E. R., C. W. Nicol, et al. (2006). "Health benefits of physical activity: the evidence." Canadian Medical Association Journal 174 (6): 801 809. Weiner, B. J., M. A. Lewis, et al. (2009). "Using organization theory to understand the determinants of effective imple mentation of worksite health promotion programs." Health Education Research 24 (2): 292 305. Wen, C. P., J. P. M. Wai, et al. (2011). "Minimum amount of physical activity for reduced mortality and extended life expectancy: a prospective cohort study." Lance t 2011 (378): 1244 1253. Westword. (2011). "Best Bicycle Challenge 2011: Tour de B cycle." Westword Best of Denver Awards 2011, from http://www.westword.com/bestof/2011/award/best bicycle challenge 1770614/ Wilcox, S., M. Dowda, et al. (2009). "Maintenance of Change in the Active for Life Initiative." Am J Prev Med 37 (6): 501 504. Woodcock, J., O. H. Franco, et al. (2010). "Non vigorou s physical activity and all cause mortality: systematic review and meta analysis of cohort studies." International Journal of Epidemiology 40 (1): 121 138. Wright, F. L. (1945). When Democracy Builds Chicago, University of Chicago Press. Xie, J., S. Sreeni vasan, et al. (2011). "Social consensus through the inuence of committed minorities." Physical Review E 84 (2011): 011130 1 to 011130 8. Yoo, J., Y. Yoon, et al. (2010). Importance of positive reputation for Smartphone adoption. 2010 International Conferen ce on Information and Communication Technology Convergence (ICTC) Jeju, South Korea : 314 318.

PAGE 259

244 AP PENDIX

PAGE 260

245 *=" Denver B cycle Select' Station Observation Sheet

PAGE 261

246 B. Denver B cycle User Survey Instrument Welcome You are being asked to be in this research study because you are a Denver B cycle user. If you join the study, you will be asked to complete a survey regarding your use of Denver B cycle. The survey takes approximately 7 minutes to complete. This study is designed to learn more about who uses the Denver B cycle system and how the bikes are used, to determine health and environmental benefits. Possible discomforts or risks include those associated with using a computer. There may be risks the researchers hav e not thought of. Every effort will be made to protect your privacy and confidentiality by maintaining the anonymity of responses. Responses are collected through secure web interface (https) and stored on a secured server. No sensitive information is col lected. You have a choice about being in this study. You do not have to be in this study if you do not want to be. If you have questions, you can call Andrew Duvall at 303 887 6964. You can call and ask questions at any time. You may have questions about your rights as someone in this study. If you have questions, you can call HSRC (Human Subject Research Committee). Their number is (303) 315 2732. By completing this survey, you are agreeing to participate in this research study. Demographic informat ion 1) What is your gender? ( ) Female ( ) Male 2) What is your age? ( ) 18 24 ( ) 25 44 ( ) 45 64 ( ) 65 and over 3) What is your ZIP Code where you live? Please enter your ZIP code like this: 01234 or like this: 01234 5678 _____________________________ _______________

PAGE 262

247 Demographic information (continued) 4) What is your ethnicity? ( ) Hispanic or Latino ( ) Not Hispanic or Latino 5) What is your race? ( ) Native American/Alaska Native ( ) Asian ( ) Black/African American ( ) Native Hawaiian or Other Pacific Islander ( ) White/Caucasian ( ) Other/Multi Racial: _________________ ( ) Decline to Respond Demographic information (continued) 6) What is the highest education you have completed? ( ) 12th grade or less ( ) Graduated high school or equivalent ( ) Some college, no degree ( ) Associate's degree ( ) Bachelor's degree ( ) Master's degree ( ) Doctorate or professional degree 7) Is your annual household income from all sources ( ) Less than $25,000 ( ) $25,000 to $34,999 ( ) $35,000 to $49,999 ( ) $50,000 to $74,999 ( ) $75,000 to $99,999 ( ) $100,000 to $124,999 ( ) $125,000 to $149,999 ( ) $150,000 or more Health status and health related quality of life 8) Would you say that in general your health is: ( ) Excellent ( ) Very good ( ) Good

PAGE 263

248 ( ) Fair ( ) Poor ( ) Don't know/Not sure 9) About how much do you weigh without shoes? ____________________________________________

PAGE 264

249 10) About how tall are you without shoes? ( ) 54 inches (4'6") ( ) 55 inches (4'7") ( ) 56 inches (4'8") ( ) 57 inches (4' 9") ( ) 58 inches (4'10") ( ) 59 inches (4'11") ( ) 60 inches (5'0") ( ) 61 inches (5'1") ( ) 62 inches (5'2") ( ) 63 inches (5'3") ( ) 64 inches (5'4") ( ) 65 inches (5'5") ( ) 66 inches (5'6") ( ) 67 inches (5'7") ( ) 68 inches (5'8") ( ) 69 inches (5'9") ( ) 70 inches (5'10") ( ) 71 inches (5'11") ( ) 72 inches (6'0") ( ) 73 inches (6'1") ( ) 74 inches (6'2") ( ) 75 inches (6'3") ( ) 76 inches (6'4") ( ) 77 inches (6'5") ( ) 78 inches (6'6") ( ) 79 inches (6'7") ( ) 80 inches (6'8") ( ) 81 inches (6 '9") ( ) 82 inches (6'10") ( ) 83 inches (6'11") ( ) 84 inches (7'0") ( ) 85 inches (7'1") ( ) 86 inches (7'2") Health status and health related quality of life (continued) 11) Thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good? ( ) 0

PAGE 265

250 ( ) 1 ( ) 2 ( ) 3 ( ) 4 ( ) 5 ( ) 6 ( ) 7 ( ) 8 ( ) 9 ( ) 10 ( ) 11 ( ) 12 ( ) 13 ( ) 14 ( ) 15 ( ) 16 ( ) 17 ( ) 18 ( ) 19 ( ) 20 ( ) 21 ( ) 22 ( ) 23 ( ) 24 ( ) 25 ( ) 26 ( ) 27 ( ) 28 ( ) 29 ( ) 30 ( ) Don't know / Not sure 12) Thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good? ( ) 0 ( ) 1 ( ) 2 ( ) 3 ( ) 4 ( ) 5

PAGE 266

251 ( ) 6 ( ) 7 ( ) 8 ( ) 9 ( ) 10 ( ) 11 ( ) 12 ( ) 13 ( ) 14 ( ) 15 ( ) 16 ( ) 17 ( ) 18 ( ) 19 ( ) 20 ( ) 21 ( ) 22 ( ) 23 ( ) 24 ( ) 25 ( ) 26 ( ) 27 ( ) 28 ( ) 29 ( ) 30 ( ) Don't know / Not sure 13) During the past 30 days, for about how many days did poor physical or mental health keep you from doing your usual activities, such as self care, work, or recreation? ( ) 0 ( ) 1 ( ) 2 ( ) 3 ( ) 4 ( ) 5 ( ) 6 ( ) 7 ( ) 8 ( ) 9 ( ) 10

PAGE 267

252 ( ) 11 ( ) 12 ( ) 13 ( ) 14 ( ) 15 ( ) 16 ( ) 17 ( ) 18 ( ) 19 ( ) 20 ( ) 21 ( ) 22 ( ) 23 ( ) 24 ( ) 25 ( ) 26 ( ) 27 ( ) 28 ( ) 29 ( ) 30 ( ) Don't know / Not sure Bicycle ownership and skill level 14) Do you own a bicycle? ( ) Yes ( ) No 15) How would you classify yourself as a bicyclist? ( ) Experienced ( ) Moderate ( ) Beginner Bicycle transportation use 16) How many times a week do you ride a bicycle for: 1 or fewer times 2 to 3 times 4 to 6 times 7 or more times Commuting (Work/School) ( ) ( ) ( ) ( ) Errands/Shopping ( ) ( ) ( ) ( ) Exercise ( ) ( ) ( ) ( )

PAGE 268

253 Recreation ( ) ( ) ( ) ( ) 17) How many times a week do you check out a Denver B cycle bike? ( ) 1 or fewer times ( ) 2 to 3 times ( ) 4 to 6 times ( ) 7 or more times 18) Which types of Denver B cycle subscriptions have you purchased? Check all that apply. [ ] Annual subscription [ ] Monthly subscription [ ] Weekly subscription [ ] Daily subscription [ ] I have not purchased a subscription 19) Did you try out the Denver B cycle system with a daily or other short term subscription before you purchased an annual subscription? ( ) Yes ( ) No ( ) I have not purchased an annual subscription Active transportation 20) Do you ride a bike or walk when you might have otherwise taken a car? Always Most of the time Sometim es Rarely Never On a bike OTHER THAN a Denver B cycle bike... ( ) ( ) ( ) ( ) ( ) On a Denver B cycle bike... ( ) ( ) ( ) ( ) ( ) While walking... ( ) ( ) ( ) ( ) ( ) 21) During the past 30 days have you walked or biked ten or more blocks to get to routine destinations, such as work, school or running errands? ( ) Yes ( ) No Bicycling to transit 22) Do you bicycle to or from other public transit modes (bus or light rail) to reach your destination? No Yes, to a bus Yes, to a light rail

PAGE 269

254 line line On a bike OTHER THAN a Denver B cycle bike... ( ) ( ) ( ) On a Denver B cycle bike... ( ) ( ) ( ) 23) Is a bicycle connection to public transit (bus or light rail) essential to your ability to commute by bicycle? ( ) Yes ( ) No Bicycle safety 24) Do you wear a helmet when you ride a bike? Always Most of the time Sometimes Rarely Never On a bike OTHER THAN a Denver B cycle bike... ( ) ( ) ( ) ( ) ( ) On a Denver B cycle bike... ( ) ( ) ( ) ( ) ( ) Other transportation 25) Do you own or have access to a car? ( ) Yes ( ) No 26) What means did you use most often to get to work or school last week? ( ) Carpooled ( ) Bus/train public transit (excluding taxicab) ( ) Bus/train public transit with a bicycle (OTHER THAN a Denver B cycle) portion ( ) Bus/train public transit with a Denver B cycle portion ( ) Bicycled (OTHER THAN a Denver B cycle) ( ) Bicycled on a Denver B cycle ( ) Walked ( ) Worked at home ( ) Other: _________________ ( ) Drove alone Optio n for further research in the use of Denver B cycle

PAGE 270

255 27) Are you interested in participating in additional voluntary research on the use of Denver B cycle? Additional research involves informal interviews and travel diaries. If you are selected for partici pation, you will receive a gift in consideration of your time. ( ) Yes ( ) No Contact info for further research If you are interested in participating in additional voluntary research on the use of Denver B cycle, please provide your contact information below. Otherwise, leave the contact information spaces blank. If you are selected for further research, you will be asked to: Keep a travel diary of the trips you make over a 24 hour period, once a month from July to November 2010 Participate in one i nformal interview about your use of Denver B cycle and your bicycling habits. The interview will last about 60 minutes and will be conducted in a public place of your choosing during September or October 2010. Following completion, you will receive a gift card in appreciation of your time. Name ____________________________________________ Email address ____________________________________________ Phone number Please type in your area code and phone number like this example: 303 555 5555 ____________________________________________ Thank You! Thank you for taking the Denver B cycle user survey. Your response is very important to us.

PAGE 271

256 C. Semi structured Interview Question Guide for Denver B cycle Select' Participants Introduction: Thanks for meeting me here today. I have a few questions to ask you about how you use Denver B cycle and about being physically active. Your responses will help me to better understand how you and users like you feel about Denver B cycle and riding bicycle s for transportation. If you are unsure about any question, just ask me to clarify. Please think of this as an informal conversation. The main goal is for me to understand your ideas. Introduction ;9 !4&7#3002/'$*3,7#,&*4%13)%33-7#-*"$',0&#])3.#%3.&#$3#,&')&" $#G&,H&)#I ` 050(&#"$'$*3,7#1)*&]#%*"$3)5#3]#"&(]9 # Denver B cycle initial experience <9 E%5#-*-#532# ]*)"$#1&4*, #2"*,4#G&,H&)#I ` 050(&v# # =9 G*-#532#]*,-#',5$%*,4#'132$#2"*,4#G&,H&)#I ` 050(&#-*]]*02($#$3#2,-&)"$',-v # U]#"37#&O/('*,9 # Activities integrated with Denver B cycle >9 G&"0)*1&#%36#532#2"&#G&,H&)#I ` 050(&9# # ?9 E%'$#-&"$*,'$*3,"#-3#532#43#$3#6%&,#532#')&#)*-*,4#'#I ` 050(*+&v# # '9 G 3#532#03..2$&#$3#63)+v#c3#$3#)&"$'2)',$"v#G3#&))',-"v#E%'$# &("&#-3#532#-3v # @9 G3&"#2"*,4#G&,H&)#I ` 050(&#]*$#*,$3#532)#(*]&"$5(&v#E%5#3)#6%5#,3$v # Bicycling behavior A9 T36#%'HȔ)#1*050(&#)*-*,4#%'1*$"#0%',4&-# 6%*()#"*,0& #2"*,4#I ` 050(&v# # '9 G3#532# 1*050(&# .3)&7#(&""#3)#'132$#$%&#"'.&#'"#532#-*-#1&]3)& # 2"*,4#I ` 050(& v# # d9 U]#532#)*-$%&)#1*+&"7# %'"#$%'5#532#)*-&#$%&.#0%',4&-#']$&)#2"*,4#I ` 050(& v #T36v # ^9 T'H&# ',5 #])*&,-"7#]'.*(5#3)#0363)+&)"#"$')$&-#$3# 2"&#I ` 050()#$3# )*-&#'# 1*+&0'2"Ȕ#-3v # Physical activity/a ctive transportation ;89 G3#532#)&42(')(5#"&$#'"*-&#$*.&#$3#&O&)0*"&v# # ;;9 G3#532#03,"*-&)#)*-*,4#'#G&,H&)#I ` 050(&#$3#1&#&O&)0*"&v#E%5#3)#6%5#,3$v #

PAGE 272

257 ;<9 C* ,0"*,4 #G&,H&)#I ` 050(& 7#%36#-3#532#]&&(#'132$#2"*,4#'#1*+&#$3#4&$# ')32,-#$36,v # ;=9 G3#532#$%*,+ # G&,H&)#I ` 050(&#%'"#0%',4&-#$%&#"30*'(#"$'$2"#3)#'00&/$'1*(*$5# 3]#1*050(*,4#]3)#$)',"/3)$'$*3, v #T36v # ;>9 T36#(*+&(5#')Ȕ#$3#03,$*,2&#$3#2"&#I ` 050()#)*-&#'#1*050( &#]3)# $)',"/3)$'$*3,#']$&)#$%&#"5"$&.#0(3"&"#3)#*,$3#,&O$#5&')v # Perceived quality of life/benefits ;?9 T 36#-3#532#$%*,+# G&,H&)#I ` 050(& #6*((#']]&0$#$%&# J2'(*$5#3]#(*]&#*,#$%&# 0*$5v# # ;@9 E%'$7#*]#',57#%&'($%# &]]&0$" # -3#532#'$$)*12$&#$3 # 532)#2"] #G&,H&)#I ` 050(&v # ;A9 E% '$7#*]#',57#&03,3.*0# &]]&0$" # -3#532#'$$)*12$&#$3 # 532)#2"] #G&,H&)#I ` 050(&v # Effects on other modes ;d9 G&"0)*1Ȕ)#0')#2"&9#G*-#532)#0')#2"�%',4%*("*,4#G&,H&)#I ` 050(&v#XO/('*,9 # ;^9 T'"#G&,H&)#I ` 050(&# 0%',4&-#$%&#(*+&(*%33-#$%'$# 532#6*((#2"&#/21(*0# $)'," /3)$'$*3,v #T36v # Perceived comfort and safety <89 E%&,#$%*,+*,4#'132$#1&*,4#*,#$)']]*0#'"#'#1*050(*"$7#632(-#532#03,"*-&)# 532)"&(]#$3#1&_ # '9 C$)3,4#',-#]&')(&""_#R3$#*,$*.*-'$&-#15#$)']]*0g#)*-&#)'*,#3)#"%*,&g# '132$#; ` <:#3]#$%&#/3/2('$*3,9 # 19 X,$%2"&-#',-#03,]*-&,$_# Q*-*+&"#]3)#$)*/"g#&,o35#1*+&#(',&"#',-# 1*+&6'5"g#03,]*-&,$#)*-*,4#*,#$)']]*0g#'132$#;8 ` ;?:#3]#$%&# /3/2('$*3,9 # 09 U,$&)&"$&-#12$#03,0&),&-_#E*((#)*-,#$)'*("#',-#/')+6'5"g#,3$# 03,]*-&,$#*,#$)']]*0g#-3#,3$#]&&(#"']&#*,#$)']]*0#&H&,#6%&,#1*+&#(',&"# &O*"$g# $%&#t.*--(&#%'(]7l#'132$#?8:#3]#$%&#/3/2('$*3,9 # -9 R3#6'5#,3#%36_#E*((#,3$#)*-&#'#1*050(&#]3)#$)',"/3)$'$*3,#1&0'2"&# $%&5#')&#,3$#*,$&)&"$&-#3)#-3#,3$#+,36#%36#$3#)*-&g#'132$#'#$%*)-# Z==:[#3]#$%&#/3/2('$*3,9 # <;9 E%'$#')Ȕ)#1*44&"$#"']&$5#03,0&),"#6%*(&#)*-*,4# '#I ` 050(*+&v # Impressions of Denver B cycle <<9 E%'$#-3#532#(*+&# .3"$# '132$#2"*,4#G&,H&)#I ` 050(&v # <=9 E%'$#-3#532#-*"(*+&# .3"$# '132$#2"*,4#G&,H&)#I ` 050(&v # <>9 V]#$%&#/&3/(Ȕ#+,36#6%3#-3,l$#2"&#G&,H&)#I ` 050(&7#6%'$#-3#532#$%*,+# ')&#$%&*)#.'*,#)&'"3,"v #

PAGE 273

258
PAGE 274

259 D. Code Definitions Table D. 1 Code Definitions CODE DEFINITION Activity tracking Reference to tracking physical activity or use of Denver B cycle Website Person visited personal account page on website to review logged ride information Smartphone Person used smartphone to track activity or locate system resources Associated health b ehavior Participation in other health behaviors in tandem with Denver B cycle use Walking Walking for recreation or transportation Riding personal bike Riding a personally owned bike for recreation or transportation Dietary effects Changes in diet, eating behavior or access to alternative food in conjunction with Denver B cycle use Awareness Reference to increased awareness of Denver B cycle, usually station locations or details of use Social channels Increased awareness of the system through peers or family Health channels Increased awareness of the system via health connection Media channels Increased awareness of the system through news media or advertising Barrier Issues that act as impediments to participation or use of the system Distance to B cycle station Usually refers to too great of a distance to a station from home or work Cost for use The price of an annual membership is too expensive, usually noted by a short term or occasional user Perceived safety Reference to perceived safety of riding a bicycle on the street in traffic, or of bicycling in general

PAGE 275

260 CODE DEFINITION Clothing Reference to bicycle specific clothing or wrinkled/disheveled clothing after riding Health barrier People spoke of limited participation because of health problems Behavior change Acknowledged change in behavior or reference to altered habits regarding active transportation Self efficacy Reference to ability to maintain a changed behavior, primarily continued use of Denver B cycle Changes to attitudes Alteration of attitudes regarding physical activity Maintenance of behavior Description of how a person maintains continued use of Denver B cycle Biking habits changed Change to total amount of biking after participation in Denver B cycle Benefits Any advantage meas ured or perceived as a result of participation Health benefit General reference to health benefit, such as "feeling good" Economic benefit General reference to economic benefit without specific details Parking costs Reference to money saved as a result of reduced car parking costs, often as a daily parking rate Fuel costs Reference to fuel cost savings Food costs Participants talked about saving money because of access to less expensive lunch options Value of membership Perceived return in value from the price of annual or short term membership Bicycle transportation Reference to transportation use of a bicycle, whether Denver B cycle or privately owned bike Perceptions of bike transportation How a person views bicycling within the range of transportation options Changes to bike acceptability Opinion about any change or alteration of social attitudes regarding the appropriateness of bike transportation

PAGE 276

261 CODE DEFINITION Changes to quality of life in city Opinion about whether Denver B cycle has or is able to change quality of life in Denver Decline in use Reference to an individual's decline in use of the system as a result of seasonal changes Weather Reference to the impact of weather on the decision whether or not to use Denver B cycle System closure Reference to closure of Denver B cycle during the winter months Daylight Savings Time Change in the amount of daylight toward the end of the day, and concerns about being seen by cars Winter Effects of cold and snow on the willingness or desire to use De nver B cycle Dedication to the system Some people spoke of Denver B cycle as a public trust, and felt that as users they had a responsibility to assist in monitoring for maintenance Feedback to the system Reference to action or desire to act to provide feedback to the system operators Destinations People spoke of specific destinations that they visited via Denver B cycle Diffusion behavior Behavior or actions that contributed to the diffusion of participation or knowledge of Denver B cycle Innovator Actions or behaviors suggesting traits of innovators, as described in diffusion theory Early adopter Actions or behaviors suggesting traits of early adopters, as described in diffusion theory Power of influence Reference to an individuals ability to infl uence the behavior of others Recruitment Actions by participants that served to recruit others to participate in the intervention Others ride because you do Instances in which participants described specific cases where they knew others used Denver B cycle because of them

PAGE 277

262 CODE DEFINITION Act as proponent Actions in which a participant functioned as an advocate for the system, such as assisting or providing information to new users Goals Reference to goals set by participants, often to achieve a higher rider rankin g on the website Health effects Reference to identified health effects attributed to use of Denver B cycle Weight loss Measured or perceived weight loss for which the participant attributed in part or in whole to use of Denver B cycle Invigoration Feelings of increased energy levels or strength Self empowerment Feelings of increased self worth or accomplishment related to participation Stress relief People spoke of using Denver B cycle as a stress reliever during the workday, or during a commute Influences on use Factors that people noted as having an impact on whether they used the system Time availability Comparison of time needed to make a trip via Denver B cycle versus a car Traffic Perceptions of traffic speed or congestion on routes to rea ch destinations Distance Usually in reference to destinations that were too far to be practical to reach by bike Learn about B cycle The manner in which a person first learned about or were exposed to Denver B cycle Reason to first use Denver B cycle Description of why a person first used the system Difficulties learning the system Problems experienced or observed by participants during use of Denver B cycle Peer introduction Reference to introduction to the system by family, friends or coworkers Dan Maes Introduction to the system as a result of the Dan Maes incident

PAGE 278

263 CODE DEFINITION John Hickenlooper Introduction to the system because of media exposure or observation of Mayor Hickenlooper Lifestyle Reference or assessment of how Denver B cycle fit within the li festyle of a participant Effects on family Spillover effects of active transportation to family members Modification of habits Reference to lifestyle changes, such as restaurant choices, in order to better adapt to Denver B cycle station locations Mode choice General reference to transportation mode choices Car use Self assessed frequency of car use, or reference to car ownership Replace car trips People spoke of consciously attempting to reduce their car use by using Denver B cycle B cycle with transit People talked about using the system in conjunction with trains or buses General transit Reference to general transit use Motivation General reference to sources of motivation to initiate or continue to use Denver B cycle Competition Motivation derived through competition with peers or other users, often in reference to rider ranking on the website Benefits Motivation linked to benefits, either experienced or anticipated Enjoyment Participants talked about enjoying riding the bikes, citing fun a reason for continued motivation Physical activity General reference to engagement in physical activity Set aside time for exercise Some participants spoke of consciously planned activities, such as time at a gym or running Consider B cycle as exercis e Whether or not a participant considered using Denver B cycle to be physical activity Increased active transportation Reference to increase in active transportation behavior

PAGE 279

264 CODE DEFINITION Safety concerns Comments regarding concerns about the safety of using Denver B cycle or of riding bikes in general Adherence to laws Knowledgeable of or adherence to laws governing the use of bicycles in public places Cars and safety Reference to fear or distrust of cars while riding in mixed traffic Bicycling skill Self assessed bicycling skill level using the Portland Bureau or Transportation definitions Helmet use Comments or opinions about helmet use while riding a bicycle Social support General reference to socially supportive activities with regard to Denver B cycle Peer activity Reference to activities engaged with friends, family or workmates while using Denver B cycle Tour de B cycle Reference to the Tour de B cycle challenge ride System impressions General impressions or opinions regarding the function of Denver B cycle as a system Impressions of the bikes Comments on the function, suitability, or performance of Denver B cycle bikes Finding station locations Comments on the ease of finding stations near origination or destination points System problems Software or hardware problems that inhibited use Dependability Reference to the dependability of Denver B cycle to meet the transportation needs of individual users Like most What a person liked most about using Denver B cycle Like least What a person least liked about using Denver B cycle Improvements Suggestions to improve the function of the system or in how to recruit new users Types of use General reference to how individuals used Denver B cycle

PAGE 280

265 CODE DEFINITION Commuting Reference to using Denver B cycle as all or part of a regular commute Utilitarian Reference to using Denver B cycle for errands or other utilitarian purposes Entertainment Reference to using Denver B cycle to access entertainment or sports venues Recreation Reference to using De nver B cycle for recreation purposes Work related Reference to using Denver B cycle for work related purposes, such as attending meetings or making deliveries