Citation
The revolving door

Material Information

Title:
The revolving door predicting turnover (intent to stay) among fundraisers in the nonprofit sector
Creator:
Horstman, Aleah Jane
Publication Date:
Language:
English
Physical Description:
xiii, 167 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Labor turnover ( lcsh )
Fund raisers (Persons) ( lcsh )
Nonprofit organizations -- Employees ( lcsh )
Fund raisers (Persons) ( fast )
Labor turnover ( fast )
Nonprofit organizations -- Employees ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 161-167).
General Note:
School of Public Affairs
Statement of Responsibility:
by Aleah Jane Horstman.

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:
123008773 ( OCLC )
ocn123008773
Classification:
LD1193.P86 2006d H67 ( lcc )

Full Text
jT'-r
THE REVOLVING DOOR: PREDICTING TURNOVER (INTENT TO STAY)
AMONG FUNDRAISERS IN THE NONPROFIT SECTOR
by
Aleah Jane Horstman
B.S. Michigan State University, 1995
M.S.S.A. Case Western Reserve University, 1997
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment of
Doctor of Philosophy
Graduate School of Public Affairs
2006


This thesis for the Doctor of Philosophy
Degree by
Aleah Jane Horstman
has been approved
by
Date


Horstman, Aleah, J. (Ph.D., Graduate School of Public Affairs)
The Revolving Door: Predicting Turnover (Intent to Stay) among Fundraisers in the
Nonprofit Sector
Dissertation directed by Professor Linda deLeon
ABSTRACT
Despite its importance to nonprofit organizations effectiveness, turnover among
fundraisers is an understudied topic. Fundraisers are ever more vital to the sector due
to the increased pressure on nonprofits to recruit and retain qualified employees as
well as acquire financial resources. Previous studies have indicated an average length
of service among fundraisers of three to four years, too short for effective
development activity. Using the Causal Model of Turnover developed by James
Price, this research surveyed fundraisers about various determinants that affect their
intent to stay with their current employer. Multiple regression analysis using 26
variables, found the following variables have a strong relationship with intent to stay:
job satisfaction, commitment to mission, distributive justice, promotional chances,
job involvement, support of supervisor, and search behavior. The goal of this research
is to add to the scarce literature on this topic and provide nonprofit managers with
helpful information on retaining fundraisers.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
Signed


DEDICATION
To my husband, Keith, who had more faith in me to complete this dissertation than I
had in myself. Even though getting my Ph.D. was my goal, he took it on as one of his
own. His love, enthusiasm, and encouragement gave me the strength I needed to see
this through.


ACKNOWLEDGMENTS
There are many people who supported and assisted me. First, and foremost, is Linda
deLeon who gave me the flexibility I needed to work on the dissertation at my own
pace, and who was always there when I needed her feedback or advice. Her
experiences in academia and management have been key to helping ensure the quality
of this research.
Other committee members, Stephen Block, Mary Feller, Jody Fitzpatrick, and
Jennifer Wade, gave me their feedback and opinions. As a seasoned fundraiser and
manager, Mary Feller offered a unique practitioners point of view. Jody Fitzpatrick
reviewed my data and helped me think through the analysis. Stephen Block and
Jennifer Wade utilized their expertise in nonprofit management to guide me. These
varying perspectives were vital to the creation of this study. James L. Price, who was
not officially on the committee, shared his intimate knowledge of The Causal Model
of Turnover and sent me valuable articles that I might not have otherwise accessed.
There are many others who assisted me. My husband, Keith, was always there to
motivate me and to occupy our daughter, giving me the study time I needed. My
parents, Ron and Marilyn, have been encouraging throughout my years of education.
My father has been a mentor and role-model and was invaluable as my first reader,


spending countless hours giving me thoughtful feedback on how to make my writing
clearer and better organized. Judy Gordon edited the final version with great care and
precision. My family was never short on enthusiasm and support for my endeavors.
In addition, I am grateful to my employer, Planned Parenthood of the Rocky
Mountains, for allowing me to be flexible with my schedule and for providing
financial support to help me further my education.
Lastly, the Association of Fundraising Professionals has been a wonderful sponsor of
this research. They were very helpful in connecting me to other people in the field
who also were interested in the turnover among fundraisers. I look forward to
working with them to disseminate this research.


TABLE OF CONTENTS
Figures..................................................................xi
Tables................................................................xii
CHAPTER
1. INTRODUCTION..............................................................1
The Purpose of the Study................................................1
Literature Review.......................................................3
The Nonprofit Sector as an Employer..................................3
Problem Statement: Turnover Within the Nonprofit Sector..............5
The Fundraising Profession.......................................... 8
Problem Statement: Turnover Among Fundraisers.......................12
The Research Questions..............................................15
The Rate of Turnover Among Fundraisers..............................16
Determinants Found in the Nonprofit/Fundraising Profession Literature.17
The Significance of the Study..........................................19
The Practical Significance of the Study.............................19
The Theoretical Significance of the Study...........................21
vii


2. THEORETICAL FRAMEWORK
.23
Turnover Models....................................................23
The Causal Model of Turnover.......................................32
Hypotheses.........................................................42
H!.............................................................43
H2.............................................................43
H3.............................................................44
H4.............................................................44
H5............................................................ 44
3. METHODOLOGY........................................................47
Data Collection....................................................47
Sample.............................................................50
Measurement Tool...................................................58
Data Analysis......................................................61
Reliability and Validity...........................................62
Conclusion.........................................................63
4. RESULTS........................................................... 64
The Average Length of Service......................................65
The Main Variables Predicting Fundraisers Intent to Stay..........67
viii


Testing the Assumptions for Multiple Regression..................67
Handling of Missing Continuous Data (Non-Demographic)............70
Explanation of Various Data Analyses in Chapter 4................70
Multiple Regression Analysis of the 26 Independent Variables.....74
Analysis of the Hypotheses...........................................81
Hi...............................................................81
H2...............................................................87
H3...............................................................92
H4...............................................................93
H5...............................................................94
Job Satisfaction, Organizational Commitment, and Search Behavior
as Dependent Variables...........................................95
Conclusion......................................................... 102
5. CONCLUSION..........................................................104
Contributions to the field..........................................104
Limitations.........................................................106
Implications for Academics and Recommendations for Future Research..109
The Average Length of Service...................................109
The Variables................................................. Ill
IX


Implications for Practitioners......................................117
Conclusion.........................................................124
APPENDIX
A: AFPs Professional Obligations......................................126
B: Letter to Sample Population.........................................127
C: Follow-Up Postcard to Sample Population.............................128
D: Survey Instrument...................................................129
E: Data Journal........................................................148
F: Research Matrix.....................................................151
G: Responses to Survey Question No. 68.................................154
H: Human Subjects Research Committee Approval..........................160
BIBLIOGRAPHY..........................................................161
x


LIST OF FIGURES
Figure
2.1: Comparison of Mobleys Model and the Steers-Mowday Model..........20
2.2: Comparison of the Steers-Mowday Model and Prices Model...........22
2.3: The Causal Model of Turnover......................................36
2.4: The Adjusted Causal Model of Turnover Using Intent to Stay
as the Dependent Variable..........................................46
3.1: Average Length of Service Formula.................................61
xi


LIST OF TABLES
Table
3.1: Sample by Gender........................................................52
3.2: Sample by Ethnicity.....................................................52
3.3: Sample by Age...........................................................54
3.4: Sample by Salary...................................................... 55
3.5: Sample by Years of Experience...........................................55
3.6: Sample by Education Completed...........................................56
3.7: Sample by NPO Sub-Sector................................................57
4.1: Average Length of Service Descriptive Statistics.......................65
4.2: Variable Definitions....................................................68
4.3: Descriptive Statistics Means and Standard Deviation for
All Variables...........................................................69
4.4: Correlation for Predictor Variables on Intent to Stay..................72
4.5: Results from Backward Elimination Regression Analysis..................74
4.6: Results from Multiple Regression Analysis for Hypothesis 1.............82
4.7: Results from Multiple Regression Analysis for Hypothesis 2.............88
4.8: Multiple Regression Results for Kinship Responsibility..................90
4.9: Results from Multiple Regression Analysis for Hypothesis 3............. 92
xii


4.10: Results from the Pearson Correlation for Hypothesis 4.................93
4.11: Results from Multiple Regression Analysis for Hypothesis 5............95
4.12: Multiple Regression Results Using the Intervening Variables
as Dependent Variables................................................98
xm


CHAPTER 1
INTRODUCTION
The Purpose of the Study
Nonprofit organizations have been vital to American society for over a
century. They play an important role in providing education, social services, housing,
and health care. As the number of nonprofit organizations grows, the challenges they
face also grow. One challenge is the increased competition for funding and
employees. Recruitment, motivation, and retention of qualified staff and volunteers
have become harder due to increased competition. The resulting work environment is
most demanding (Barbeito & Bowman, 1998: 2). Another challenge is the aging of
the nonprofit workforce. Nonprofits will find recruiting more difficult as baby
boomers move towards retirement (Joslyn, 2002a). It is important for nonprofit
scholars and practitioners to give more attention to these challenges in order to
sustain and improve the nonprofit community. The range of essential nonprofit
contributions to our communitiesfrom healing the sick to showcasing artistic
expressiondepends on the vitality of our nonprofit workforce (Peters,
Femandopulle, Masaoka, Chan, & Wolffed, 2002: 22). Unfortunately, many in this
workforce are leaving their employers at a detrimental rate
1


(Joslyn, 2002b; Peters et al., 2002; Taketa, 2002). For executive directors, the
average tenure is three to five years (Peters et al., 2002).
Since the 1980s, nonprofit organizations have come under increasing pressure
to be efficient and effective. As a result, nonprofit managers need to utilize stronger
leadership techniques to manage staff and to raise funds (Barbeito & Bowman, 1998).
Nonprofits are experiencing an elevating reliance on fundraisers for financial stability
(Wagner, 2002). At the same time, the turnover rate for fundraisers (also called
development officers) is becoming a matter of great concern (Peters et al., 2002). Past
studies illustrate that fundraisers have an average length of service of every 3 to 4
years, jeopardizing the capacity of nonprofits to remain financially healthy (Schwinn,
2002; Schwinn & Sommerfield, 2002; Sommerfeld, 2002a; Anonymous, 2003). For
planned giving officersa subset of fundraisersthe average turnover is even
higher, at every 2 years (Ashton, 2005). According to Mai Warwick, a fund-raising
consultant in California,
[turnover among fundraisers] represents a major cost, and is one of the
most significant factors holding back many leading nonprofit
organizations from fulfilling their fundraising potential. The issue is
really worth academic study to give charities the facts they need to
tackle the issue (Schwinn & Sommerfield, 2002: 41).
The purpose of this dissertation is to enhance the nonprofit sectors
understanding of the determinants impacting fundraisers voluntary turnover, and
thereby help nonprofits retain their fundraising staff. Nonprofit managers can make
better human resource decisions if they are equipped with the knowledge of why
2


turnover is occurring. The research questions are: (1) What is the average length of
service among fundraisers? and (2) What are the main variables predicting
fundraisers intent to stay in their current job?
Literature Review
The Nonprofit Sector as an Employer
The growth of the nonprofit sector affects not only the people it serves, but also the
people it employs. Nonprofit organizations are commonly referred to as charities,
third sector organizations, tax-exempt, philanthropic, and voluntary associations. By
definition, nonprofits are regulated under the federal tax code utilizing the charitable
status under section 501(c)(3) of the tax code. This group, also called 501(c)(3)
organizations, makeup the majority of nonprofits and serve the broad public through
social services, education, religion, scientific advancement and relief of poverty.
Another type of nonprofit organization falls under the 501(c)(4) section of the tax
code. These organizations differ from the 501(c)(3) organizations in that they are not
tax exempt. The majority of their activities include lobbying and political work
(Boris, 1999). For the purpose of this dissertation, when referring to a nonprofit
organization it will be defined as a 501(c)(3), not a 501(c)(4). The discussions
around the research sample in Chapters 3, 4 and 5 also will be referring to 501(c)(3)
employees only.
3


Estimates show that the nonprofit sector grew slowly from the 19th century to
about 1960, and since then has grown rapidly (Hammack, 2001). Numbering 12,500
in 1940 and 50,000 in 1950; by 1967, there were 309,000 secular nonprofits; by 1977,
there were 790,000; and by 1989, just under a millionan eighty-fold increase in 40
years (P. D. Hall, 1994: 19). Other accounts of nonprofit sector growth show that in
1989 there were 1,262,000 nonprofits and in 1996 there were 1,455,000, an increase
of 15.3 percent over 7 years. Today, that number is over 1.8 million organizations (V.
Orlando from Independent Sector, e-mail communication, April 21, 2005). The
growth in the number of nonprofits has been substantially faster than the population
growth (Boris, 1999).
Translating these numbers into salaries and wages these organizations pay out
is helpful in understanding the extent of the nonprofit sectors influence on the
nations economy. In 1930, nonprofit expenditures accounted for about 1.6 percent of
all wages and salaries; by 1960, they accounted for more than 3 percent. Their share
has continued to grow.
Nonprofit expenditures as a share of U.S. wages and salaries increased
by 36 percent in the 1970s, then doubled between 1980 and the mid-
1990s. In the year 2000, nonprofit organizations account for as much
as 10 percent of U.S. employment and perhaps 9 percent of the
domestic economy as a whole (Hammack, 2001: 165).
The Independent Sector (2004) reports that 64 percent of workers in the
nonprofit sector are employed by health organizations and educational institutions.
However, the most rapid growth is occurring in social service agencies. A forecast
4


done by the Independent Sector states that between 2000 and 2010, the health
services industry is estimated to add 2.8 million jobs. Social and human services will
add another 1.2 million jobs (2). The nonprofit sector will have to make sure that
nonprofit managers and their human resource professionals are equipped with the
skills and strategies to recruit and retain the quality employees needed for these
millions of new jobs.
The growth in the nonprofit sector has drawn the attention of scholars
(Hammack, 2001, Barbeito & Bowman, 1998) and practitioners to the fact that
competition among nonprofit organizations is real. As a result, more sophisticated
ways of bringing in and retaining quality staff, volunteers, and funding has become a
topic of concern and discussion.
Problem Statement: Turnover Within the Nonprofit Sector
Turnover has long been a focus of study by organizational management
scholars. Employee turnoverpeople leaving organizationsis a major
organizational phenomenon. In recognition of this fact, turnover is included in many
definitions of organizational effectiveness (Mobley, 1982 b). Scholars and
practitioners have focused on turnover and retention because these are issues that
impact businesses bottom lines. Businesses and organizations can realize a financial
loss from turnover due to an increase in recruiting, hiring, and replacement costs
5


(Mobley, 1982 b; Fitz-Enz, 1997; Chambers, 2001; Peters et al., 2002). Turnover
drains organizations resources through advertising, through time spent interviewing
and training new staff, and through opportunity costs during the period the new
employee is becoming familiar with the job and the organization (Peters et al., 2002).
Specifically, turnover costs can range from
... six months of an hourly workers salary to 18 months of salary for
a professional employee. If a 1,000-employee hospital loses 84 nurses
per year (an average for a hospital with 400 nurses), it will spend
nearly $5 million per year in replacement costs alone, or more than
$18,000 every business day (Drizin, 2002: 4).
Turnover of top staff in nonprofit organizations is recognized by scholars and
practitioners as a problem. A study done by the Neighborhood Reinvestment
Corporation found that among their 184 affiliates, the average length of service for
the top executives was three years. This lack of tenure was causing turmoil among
the organizations and preventing them from raising funds efficiently (Marchetti,
1999). According to Paul Light, it is important to reduce the sectors well-known high
turnover rate in order to deal with the shrinking workforce issue (Joslyn, 2002b). The
Council of Michigan Foundations (2004) states that nonprofit executives used to
average 10 years tenure, but that number has now dropped to 5 years. Recent
national studies predict this declining trend to continue. A 2002 study found that one-
third of nonprofit executive directors on the Hawaiian Islands expected to leave their
jobs by 2004, and only 27 percent planned to remain in their jobs past 2007 (Taketa,
6


2002). A similar scenario is seen for nonprofits nationwide. This trend toward short-
tenure executives can do lingering harm to any agencys culture and performance
(Peters & Wolfred, 2001: 6).
A 2002 study (Peters et al.) found that 64 percent of nonprofit managers said
the impact of turnover and position vacancy was high or very high. Only nine percent
said it had no impact at all. The Peters study found that the responses to this question
did not vary significantly by agency mission or budget size. Within these same
nonprofits a disproportionate number of administrative, non-program staff (including
fundraisers) leave their jobs. The majority of nonprofits operate with administrative
(budget) rates of 10 to 20 percent, yet administrative staff represent 47 percent of
employees that leave their jobs (2). One reason for this is that administrative staff
have more transferable skills than program staff. As a result of nonprofits struggling
to retain administrative staff, some agencies have stretched their previous limits by
offering senior fundraisers higher salaries than those of the executive directors hiring
them (Peters et al., 2002).
Not all turnover is bad. There are also positive organizational consequences,
namely the displacement of poor performers and the influx of new knowledge and
ideas (Mobley, 1982 b; Griffeth & Horn, 2001; Sommerfeld, 2002b). However,
research indicates that high exit rates tend to negatively affect organizational
effectiveness due to hiring costs, opportunity loss during vacancy, and resources
spent on training new employees (Alexander, Nuchols, & Bloom, 1994).
7


Additionally, the Peters (2002) study showed that nonprofit managers felt that the
ideal voluntary turnover of their staff would be zero percent.
The Fundraising Profession
Fundraisers are people who acquire revenues from private sources for
nonprofit organizations (Wagner, 2002). Private sources generally come from
individuals, corporations and foundations. More recently, nonprofits are charging
their fundraising staff to not only bring in dollars from private resources, but from
government and commercial activities as well. This is called Total Resource
Development (Pollack, 2003). Nationwide, nonprofits rely on fees and dues for
approximately 38 percent of their income; government for 31 percent; other sources
for 11 percent; and private gifts (i.e., foundations, corporations, and individuals) for
the remaining 20 percent (Hodgkinson, Nelson, & Sivak, 2002). These percentages
may shift as government continues to cut back in funding nonprofits, which will
increase nonprofits dependency on fundraisers (Herbst, 2005).
Fundraisers are not only important to the health of the countrys philanthropic
activities, but they also play a vital role in the health and stability of the nonprofit
sector. According to Wagner (2002), fundraising [does not] take place in a vacuum.
It is an essential function in the entire scope of the nonprofit sector (13). In 2003,
Americans gave $241 billion to nonprofits (The Center on Philanthropy at IU, 2003).
8


Many of these donations came in response to an ask made by charities through their
fundraising staff.
Robert Hartsook (1996) stated that fundraising is an important part of the
success of any nonprofit institution. Serious attention to the means and type of
supervision of the fundraising staff can be critical to success (41). For those
nonprofits that rely on private contributions for a large portion of their budget, this
statement weighs even more heavily. An example of this reliance is National Public
Radio, which obtains 85 percent of its budget from private contributions (National
Public Radio, 2004).
Using data from a 2002 study done by The Urban Institute (Hager, Rooney, &
Pollack, 2002), a conservative estimate can be made of nearly 200,000 fundraisers
(defined as staff members whose main responsibility is to raise money for their
organization) in the United States. That number is significantly larger than an
estimate of 25,000 in 1995 (Hohler, 1995). Historically, fundraisers were responsible
for acquiring private donations for nonprofit organizations (Duronio & Tempel,
1997).
In 1952, Amaud Marts determined that the tremendous growth in charitable
giving since 1900 was attributed not only to a burgeoning growth in wealthy
Americans, but to a new way of encouraging people to givethe modem fundraising
campaign. Organizations such as the March of Dimes and United Way revolutionized
fundraising. They started what some call the democratizing of philanthropy or the
9


mass campaign for small donations. As these fundraising strategies became more
organized, so did the individuals who ran them (Hodgkinson, 2002; Hodgkinson et
al 2002).
As the new profession of fundraising emerged in the mid-1930s, the American
Association of Fund Raising Counsel (AAFRC) was created. This led to the
development of professional standards for fundraisers. Today, the largest association
for fundraisers is the Association of Fundraising Professionals (AFP), previously the
National Society of Fund Raising Executives (established in 1960). (See Appendix A
for AFPs Statement of Ethical Principles.) AFP reported 1,899 members in 1979;
20,000 members in 2002 (Hodgkinson, 2002; Hodgkinson et al., 2002); and 26,000 in
2004 (AFP, 2004) throughout the United States, Canada, Mexico, and China. Since
all fundraisers are not members of AFP, the actual number of fundraisers is larger
than 26,000, and may be closer to the estimated 200,000 fundraisers aforementioned.
The largest influx of fundraisers was seen in the 1980s as a response to the cut
in federal funding of nonprofit organizations. Around the same time, nonprofits
started hiring internal fundraisers instead of consultants. In 1980, the creation of the
Independent Sector, a membership organization now consisting of 800 nonprofits,
was created. This led to donors and fundraising professionals working together in a
more formalized fashion due to increased networking and educational opportunities
(Wagner, 2002).
10


In the 1990s, the development of the profession continued as more individuals
chose fundraising as a career. The escalating growth in the number of nonprofits
required more fundraisers. In addition, there was increased activity in research and
publications, and more colleges and universities offered fundraising classes and
certifications (Wagner, 2002).
Fundraising is a growing and important profession, yet there is limited
empirical research on fundraisers. The Association of Fundraising Professionals,
Indiana Universitys Center on Philanthropy, search firms, and nonprofit associations,
such as the Council for the Advancement and Support of Education (CASE), have
done some initial studies on the demographics of fundraisers. One of the most
comprehensive empirical studies done on a national level is Margaret Duronio and
Eugene Temples book, Fundraisers: Their Careers, Stories, Concerns, and
Accomplishments (1997). The authors surveyed 1,651 respondents regarding their
age, gender, education, salary, professional history, current positions, and attitudes
towards fundraising. (Details of this study will be discussed further in Chapter 2).
Today, the demographics of fundraisers can be better understood by looking at
AFPs 2005 Compensation and Benefits Study, which is a survey of AFPs
membership. This study confirmed that 68 percent of the membership is female and
32 percent is male. Less than 10 percent self-identified as non-white. The types of
nonprofits most represented are educational organizations, health services, and
human/social services. These fundraisers tend to be older, with 60 percent between
11


the ages of 45 and 64. The study found that among fundraisers in the United States,
the average salary is $80,685; for males it is $95,879 and for females it is $73,295
(Brown, Healey, Maehara, & Williams, 2005). This is good demographic
information, yet there is much more to learn about this population; such as why they
joined the fundraising profession, what career or academic path they came from, and
what they perceive as the opportunities and challenges of working as fundraisers.
Problem Statement: Turnover Among Fundraisers
The impact of staff departure is not equal for all positions in nonprofit
organizations. Obviously, a receptionists departure has a different effect than the
departure of a fundraiser. Since a fundraiser generates resources and a receptionist
typically does not, there is a greater opportunity cost associated with the exit of a
fundraiser (Peters et al., 2002). Because fundraisers are often charged with building
relationships with donors on behalf of the organization, turnover negatively affects
those relationships and thereby, donors contributions. Donors who feel that they are
passed from fundraiser to fundraiser may not be as inclined to give (Schwinn, 2002).
A specialist in staffing of fundraisers stated that a major impact of fundraisers
turnover is confusion among donors. They like to know that the people who ask
them to support the organization also stay and support it (C. Murray, personal
communication, October 5,2004).
12


It is difficult to estimate exactly how much of an organizations fundraising
capacity is lost each time a fundraiser leaves, but the loss can be estimated by figuring
the amount of time the position is vacant and the amount of money that the fundraiser
raised in a year. William J. Smith, a business consultant calculates that
... if a fundraiser who brought in $2 million a year leaves and it takes
four months to find a replacement, a third of the $2 millionabout
$667,000in future income is lost. If it then takes six months for the
replacement person to get fully set up in the new job, the charity can
expect to lose as much as $1.16 million, Mr. Smith says, because the
new hire may bring in 30 to 50 percent less in gifts while (s)he is
learning the ropes (Schwinn & Sommerfield, 2002: 40).
Some scholars believe the turnover rate problem is overstated. In exploring
turnover in the field, Duronio and Temple (1997) found that the average length of
service for female fundraisers was 3.31 years and 4.42 for male respondents. Looking
at those with 10 or more years experience in the profession, the average length of
service increased to 4.6 years for women and 5.72 for men. Duronio and Temple
concluded that
... the turnover rate problem is overstated ... our research data
suggest that turnover in [fundraising] is related to the rapid growth of
the field, the resulting opportunities for advancement for [fundraisers],
and the strong competition for experienced [fundraisers]. The
implication is that as growth of the field slows, the field itself will
become more stable (57).
That time of a slowing field may have arrived. According to recent trends,
data suggest that the nonprofit labor market has weakened significantly. While the
nonprofit sector maintained a good pace in the 2001 recession and its aftermath, 2004
13


data show a halt in employment growth in the sector nationwide. Between 1990 and
2004, the average employment growth in the nonprofit sector was 2.4 percent. In the
year ending July 2004, that rate had fallen to .5 percent (Irons & Bass, 2004). The
nonprofit sector really has ground to a halt over the last year and a half... there is
something else going on here beyond the normal ebb and flow of the sector (Jensen,
2004: 23). Even though this data does not focus on fundraisers per se, the historically
rapid growth of the nonprofit workforce, which directly affects the growth of the
fundraising profession, is changing. Therefore, the turnover issue needs to be
revisited.
Other experts disagree that the turnover phenomenon is overstated. The
Chronicle of Philanthropy (Schwinn & Sommerfield, 2002) featured an article called
The Revolving Door Dilemma, which indicates that for all the talk of high
turnover among [fundraisers], the issue has received little recent attention from
researchers, charity associations, or foundations (40). The authors reported that
charity executives, board members, and other fundraisers are frustrated by the
problems caused by high turnover among fundraisers. This is a problem that, at least
anecdotally, appears to be going strong or getting worse, despite a weak national
economy that has prompted people in other professions to curtail their job hunting
(Schwinn & Sommerfield, 2002: 39).
In New York City, the issue gained attention from The New York Times
(Herbst, 2005). Nonprofits in New York say that the problem of high turnover among
14


fundraisers is not only worrisome, it is critical. One executive suggested that demand
for fundraisers has outpaced the supply and that there is a terrible dearth of good
fundraisers (1).
The problem is not only a problem for organizations, it also has become an
issue for individuals who go into interviews with the revolving door reputation of
the profession. According to Colette Murray, 2004 Chair of AFP and an executive for
a search firm specializing in development positions, we are always asked to find
candidates who will make a long-term commitment (C. Murray, personal
communication, October 5, 2004). In a conversation with a fundraiser in higher
education, he stated that in his recent job search interviews, several of his potential
employers questioned him about his intentions regarding staying in the new position
for a significant length of time, as they were tired of staff leaving after a year (P.
Gaines, personal communication, August 13, 2004).
The Research Questions
Based on a literature review of the fundraising profession, research questions
will look at how the situation is now and why turnover is happening. The research
questions are: (1) What is the average length of service among fundraisers? and (2)
What are the main variables predicting their intent to stay in their current job?
15


The Turnover Rate Among Fundraisers
A 2002 study done by CASE (Council for Advancement and Support of
Education) found that fundraisers have the lowest average length of service of the
five divisions in higher education institutional advancement (development,
advancement management, advancement services, alumni relations, and
communications/marketing). In fact, even though fundraisers were the second highest
paid group of the five, they spent at least one year less in their positions and
institutions than did employees in the other four disciplines. The Council for
Advancement and Support of Education found that, on average, fundraisers stayed in
their current positions 3 years and at their current institutions 6 years (Lajoie &
Pollack, 2002). Although this study was limited to higher education, it is an indication
that fundraisers may have a shorter length of service than their colleagues in
institutional advancement.
In 2002, AFP surveyed 1,215 people who, on average, said they had been in
their current jobs for 3 years or less (Schwinn & Sommerfield, 2002). DRG
(Development Resource Group), a New York search firm, found that in 2003
fundraisers were staying in their current positions longer than previous DRG surveys
indicated. Nineteen percent stayed 5 to 6 years, 52 percent stayed 3 to 4 years and 16
percent stayed 1 to 2 years (DRG, 2003).
A 2004 study conducted by AFP found that 24 percent of fundraisers had
changed jobs the previous year. This turnover rate is down from 2002, when a similar
16


study found 36 percent of the profession switching jobs. The same study found that
61 percent of fundraisers plan to stay in their jobs indefinitely, up from 42 percent in
2003 (Hall, 2004).
Another study concluded that among all the sectors (nonprofit, for profit, and
government), the average employee stays in his or her job 3.4 years (Drizin, 2002).
For nonprofit executive directors it is about 3 to 5 years (Peters & Wolfred, 2001).
Therefore, the 2003 data showing fundraisers leaving every 3 to 4 years (DRG, 2003)
is similar to other professions average length of service. Yet several Chronicle of
Philanthropy articles discuss the turnover rates among fundraisers as high
(Baldwin, 2002; Schwinn, 2002; Schwinn & Sommerfield, 2002: 39; Sommerfeld,
2002b) and higher, compared to other professions (H. Hall, 2004:18). Even if the
data do not consistently support that fundraisers turnover rate is high compared to
other professions, the fundraising profession is questioning the reasons for what is
seen as a revolving door problem.
Determinants Found in the Nonprofit/Fundraising Profession Literature
Regardless of how fundraisers compare to other professionals in turnover rate,
nonprofits experience hardship as a result. Therefore, in conjunction with the
argument that there is a problem of turnover among fundraisers, there is also the
question of why it is happening. Regardless of the labor market (as hypothesized by
Duronio and Temple), people leave jobs for different reasons. In order to decrease the
17


nonprofit sector turnover rate among fundraisers, practitioners need to understand the
related determinants. The determinants suggested by interviews in Duronio and
Tempers 1997 study included: (1) a desire for personal achievement over dedication
to an organization, and (2) a lack of support from their board and president (Duronio
& Tempel, 1997).
The Chronicle of Philanthropy (Schwinn & Sommerfield, 2002) cited AFPs
survey of 1,215 people who said the reasons for their turnover included: (1) stress
from the demands of the job, especially from the board and president; (2) ability to
get a higher salary from another organization; (3) desire to obtain a job with a more
prestigious title; and (4) lack of agreement with their supervisor. Common
determinants, such as support from supervisor and board and a desire for personal
achievement (i.e., salary and title) are evident in these two studies.
Yet other studies have found different issues and variables relating to the
turnover of fundraisers. The president of DRG hypothesized that the economy is a
factor. Scarce jobs and sluggish giving patterns have made many [fundraisers]
reluctant to jump ship ... charities may have grown more patient with the
[fundraising] process and are putting less pressure on slow-producing employees
(Anonymous, 2003: 33). The DRG survey found that 46 percent of the 109 nonprofit
organizations responding stated that fundraisers who left their current positions did so
for quality of life reasons such as spending more time with family or deciding to
obtain additional education (DRG, 2003). These conclusions were made from a low
18


number of respondents, all of whom were speculating on behalf of fundraisers who
had left the organization. Therefore, the reliability and validity of this research is
questionable.
In summary, various factors found in the literature for fundraisers voluntary
turnover are: (1) fundraisers relationships with their organizations president and
board (Duronio & Tempel, 1997; Schwinn & Sommerfield, 2002); (2) availability of
higher paying jobs in the field (Duronio & Tempel, 1997; Schwinn & Sommerfield,
2002); (3) quality of life needs (Anonymous, 2003; DRG, 2003); (4) the strength of
the economy (Anonymous, 2003; DRG, 2003); (5) self-interest motivators such as job
title and salary (Duronio & Tempel, 1997; Schwinn & Sommerfield, 2002;
Anonymous, 2003; DRG, 2003); and (6) commitment to the organizations mission
(Duronio & Tempel, 1997; Schwinn & Sommerfield, 2002). The theoretical model
discussed in Chapter 4 will further examine these and other determinants.
The Significance of the Study
The Practical Significance of the Study
Practically, the big picture goals of this research are to help nonprofits retain
fundraisers so that they can (1) reap the benefits of having tenured fundraisers and (2)
avoid the costs associated with turnover. As discussed in the problem statement,
19


tenured fundraisers are beneficial because they have gotten beyond that first-year
learning curve, during which they do minimal fundraising (Schwinn & Sommerfield,
2002), and they tend to have better-developed relations with their donors, which leads
to more consistent contributions (Schwinn, 2002).
Research estimates that the cost of turnover ranges between 93 percent to 200
percent of a leavers salary, depending on skill and responsibility levels (Griffeth &
Horn, 2001). For the average fundraiser making $80,685 and leaving their job, it
would cost their employer between $75,037 and $161,370. Indeed, if nonprofits could
minimize avoidable turnover, they would also circumvent these costs, ultimately
improving their efficiency.
Nonprofit managers need to better understand the reasons for fundraisers
turnover so that they can implement retention strategies that combat the related
determinants. Some human resource professionals rely on surveys to predict who will
quit a company and the reason for their leaving. Such predictions can assist managers
in forecasting future turnover problems. In addition, scholars have long known that
employee surveys diagnose the causes of voluntary turnover more accurately than the
perceptions of supervisors or exit interviews by human resource managers on why
people quit (Griffeth & Horn, 2001). For the majority of nonprofits that do not take
the time to survey their own employees on these issues, this study will add insight
into the main reasons why fundraisers leave, and can act as a tool for managers to
improve retention rates.
20


As the fundraising profession continues to grow (Wagner, 2002), the demand
for profession-based research will increase. Specific topics of interest to the
population, such as turnover, should be studied to support the development of the
field.
The Theoretical Significance of the Study
Theoretically, there are two major objectives for this research: (1) to add to
nonprofit management literature and (2) to extend the application of employee
turnover theory to a new populationfundraisers.
Most theoretical models improve with use. This dissertation will use an
existing theoretical model, the Causal Model of Turnover, which historically has
measured for-profit and military employee populations, and adapt it for fundraisers.
Scholars dedicated to turnover research now have additional knowledge to support, or
refute, theories important to the field of turnover research.
Management technologies within the nonprofit sector have historically
borrowed from the for-profit sector and the public sector. Because nonprofits have an
audience and goals that differ from the other sectors, their management tools must be
modified to meet their unique needs (Block, 1990). The literature on nonprofit
management and human resources is growing. This research project adds to this
important movement. This dissertation, and its conclusions, will encourage future
research on human resource issues within the nonprofit sector.
21


The following chapters will provide details on the theoretical model chosen
for this research. They will also cover the research methodologies used and the
results found regarding fundraisers and their intent to stay in a job.
22


CHAPTER 2
THEORETICAL FRAMEWORK
Turnover Models
In the field of organizational development and behavior, theories around
retention and voluntary turnover have been developing for years. Turnover is an
important field of study because it adversely influences organizational effectiveness
or the ability of organizations to meet their goals (Price, 2001). Turnover is the
movement of members across the boundary of an organization (Price, 2001: 600).
Voluntary turnover is turnover that is self-initiated instead of organization-initiated
(Mobley, Griffeth, Hand, & Meglino, 1979: 496). It is also referred to as quitting or
avoidable turnover (Price, 1997: 532). The term turnover in this dissertation
means voluntary turnover.
The measurement of turnover among organizational scholars is typically done
by measuring turnover within an organizational setting. There are several formulas
used to measure turnover, most of which look at employee records (Price, 1997). It
is also common to measure the average length of service within an industry or
profession (Filipowski 1992, Pennington 2003). For the purposes of this dissertation,
the term turnover will be used when looking at the literature; however, the
dissertations actual research will be measuring the average length of service (not
23


turnover). This will be described further in Chapters 3 and 4. Formal inquiry on
employee turnover first appeared in the mid-1950s with research indicating a link
between turnover and employee dissatisfaction (Mowday, Koberg, & McArthur,
1984). Since the 1950s, over a dozen turnover models have been proposed and 1,000
studies have been conducted (Horn, Caranikas-Walker, & Prussia, 1992). In 1973,
Lyman W. Porter and Richard M. Steers reviewed 60 studies from the turnover
literature. They found that more comprehensive studies were needed to look at the
many possible variables related to turnover (Porter & Steers, 1973). The turnover
theory landscape received most of its shape between 1977 and 1981 when scholars
became more sophisticated in developing comprehensive models (Steel, 2002).
Thomas Lee and Richard Mowday (1987) suggested that the three main
turnover models of that time were from (1) William H. Mobley, (2) Richard M. Steers
and Richard T. Mowday, and (3) James L. Price. Although decades have passed, the
models developed by these scholars continue to shape current research.
The scholar most commonly referenced in the turnover literature is William
H. Mobley. His work has been one of the most influential turnover models because it
serves as a foundation for other scholars to confirm and/or question. (Horn &
GrifFeth, 1991; Horn et al., 1992; Sager, Griffeth, & Horn, 1998). Mobleys 1978
conceptual framework has attracted more research attention than any other turnover
theory (Horn et al., 1992: 890).
24


Mobley began developing his framework by criticizing the existing turnover
research, which was limited to establishing economic and job dissatisfaction as
factors contributing to turnover. Economists and personnel researchers focused on
vacancy rates and economic activity to understand turnover; behavioral scientists
tried to understand the correlation between job dissatisfaction and turnover. Mobley
believed that these approaches were conceptually simplistic and empirically
deficient bases for understanding the employee turnover process (Mobley et al.,
1979: 493).
Mobley believed that there was more to the phenomenon of turnover and
advocated for a better-developed conceptual model. Mobley (1977) originally
hypothesized that the causal relationship leading to turnover was made up of
intermediate linkages of variables. He looked at job attitudes and withdrawal
cognitions (thoughts about quitting), which were associated with the decision to leave
a job. If a person is unhappy at a job, his or her dissatisfaction will arouse thoughts of
quitting. Those thoughts will then lead to a consideration of seeking another job.
Therefore, a decision to leave, he theorized, was indirectly related to actual turnover.
The second model Mobley (1979) developed came from his realization that
turnover was much more comprehensive and complex than his original framework
suggested. He expanded his model to include a broader range of variables that could
25


influence a persons decision to leave. Instead of concentrating on the intermediate
linkages within the decision process, he focused on the complex relationships
between factors (job- and non-job-related) that can influence the decision process.
In order to study the complex relationships between determinants, he
organized his new model into individual demographic and personal variables (i.e.,
age, tenure, sex, family responsibilities, and education) and work environment
variables (i.e., pay, promotion, supervision, and peer group relations). In addition, he
included job satisfaction, organizational commitment, involvement, job attachment,
met expectations, job content factors, the external environment, and intentions to
search and quit into his model.
Both of his models have been frequently challenged (Mowday et al., 1984;
Lee & Mowday, 1987; Horn et al., 1992; Sager et al., 1998), and from those
challenges, a variety of cognitive constructs have developed using additional factors
that Mobley did not consider. For example, Horn et al (1992) compared the construct
of thoughts of quitting, search decisions, and quit intentions among the military,
civilian, nursing, and non-nursing professions. He found that military turnover
experienced a different process than non-military turnover, mainly due to the
occupational market. The effect of thoughts of quitting among the military might not
lead to quit intentions as quickly as other occupations because of the militarys innate
relocation process. Military personnel often commit to multi-year employment due to
strict enlistment policies.
26


In contrast, because of chronic nursing shortages, nurses show a greater
correlation between quit intentions and actual quitting than do military personnel.
Nurses have an easier ability to implement their intent to leave because of their
confidence in finding other work in the market (Horn et al., 1992). The fundraising
profession may be similar to the nursing profession because it also shows signs of a
high market demand (Duronio & Tempel, 1997; Herbst, 2005). Chapters 4 and 5 will
look at this assumption further.
The second of the three models is the Steers-Mowday model developed by
Richard M. Steers and Richard T. Mowday (Steers & Mowday, 1981). This model
includes a sequence of variables that leads to turnover. Steers and Mowday argued
that job expectations that have been met (or not met) lead to the intention to stay (or
leave). The intention to leave, plus the influence of non-work variables and
alternative job opportunities, then leads to actual leaving. The Steers-Mowday model
is unique among most of the turnover models offered because it considers more non-
job related variables, such as the spouses job and time left for family. The main
difference between the Steers-Modway model and Mobleys model is the order of
events leading to turnover (see Figure 2.1). Mobley theorized that search processes
precede intention to leave, whereas Steers and Mowday theorized that search
processes follow intention to leave (Lee & Mowday, 1987: 724).
27


MOBLEY Search Processes Intent to Leave
STEERS-MOWDAY Intent to Leave > Search Processes
Figure 2.1: Comparison of Mobleys Model and Steers-Modway Model
The third significant model of turnover was developed by James L. Price. He
began developing this model in 1972 with colleague Charles W. Mueller and graduate
students at the University of Iowa. Next to Mobley, Price has received the most
systematic research attention (Lee & Mowday, 1987). Prices approach to developing
a model came from a wish to look beyond the models that were predominantly
developed by economists and to consider non-economic factors. Price believed that
the economic models focused too heavily on monetary variables and did not
incorporate enough non-monetary variables.
Like many other scholars in this field, Price focused on creating a
comprehensive model. Price theorized that the interaction between job satisfaction
and job opportunities occurred just prior to an employee leaving an organization. He
focused on five main determinants that were directly related to job satisfaction: pay
(income), integration (workers cohesion), instrumental communication (the
transmission of information necessary to do a job), formal communication (officially
transmitted information), and centralization (the degree to which power is
differentially distributed).
28


Prices model differed from that of the Steers-Mowdays model in that Steers
and Mowday proposed that the immediate antecedent of an employees leaving is
the interaction of the intention to leave and alternative job opportunities. Price,
instead, hypothesized that the immediate antecedent of an employees leaving is the
interaction of job satisfaction and alternative opportunities (Lee & Mowday, 1987:
722) (see Figure 2.2). Price also put more emphasis on structural variables
(autonomy, distributive justice, job stress, pay, promotional changes, routinization,
and social support). Even though Mobley and Mowday included structural variables
in their models, they have more of a social psychological focus, looking at
satisfaction, utility, and values. Price admits that all the models show an overlap of
multiple variables, but there is a distinct difference in the emphasis on the types of
variables (economic, social psychological, structural) used in each model (Price,
2001).
STEERS-MOWDAY Intent to Leave <-* Alternative Opportunities Leaving
PRICE Job Satisfaction Alternative Opportunities Leaving
= interaction
Figure 2.2: Comparison of Steers-Modway Model and Prices Model
Price spent his career developing this original theory into the Causal Model of
Turnover, one of the most respected in the field (Price, 2004). The model was
developed over several decades and went from incorporating six main variables to
29


incorporating over 20 variables. With continued advancement and research, however,
turnover models continue to receive criticism. Much of the criticism comes from the
lack of quality empirical research and validation studies (Lee & Mowday, 1987; Horn
et al., 1992; Steel, 2002).
Robert P. Steel (2002), in his analysis of the current field of turnover research,
outlined its three limitations. The first is the gaping disparity between the conceptual
importance of job search mechanisms and their empirical significance (347). Most
turnover models give job search mechanisms a central role. Yet the average
correlation between perceived employment alternatives and turnover is actually
insignificant at r = .13. Even though researchers have tried to explain this disparity,
they have not been completely successful, and Steel (2002) argues that this disparity
has caused a blemish on turnover researchs otherwise enviable record of scholarly
progress (347).
The main concern is that giving job search mechanisms a central role in
research fails to explain why some people leave their jobs without any kind of job
search. Steel suggests an evolutionary search model, which theorizes that the search
behavior of individuals is more complex than previous models depict. He theorizes
that individuals receive various levels of information or feedback about the job
market, which influences their search behavior. While he acknowledges that static
cohort methodology has a role in understanding turnover, he suggests a longitudinal
methodology to appropriately test his evolutionary model (Steel, 2002).
30


The second gap Steel identified is the fields lack of research sensitivity
towards the amount of time the turnover process takes. The popular static and cross-
sectional studies do not appropriately capture the decision-making process of
turnover.
Although turnover theorists consistently choose to portray turnover as
the product of a time-lapse process, relatively little progress has been
made in the empirical literature toward accommodating the temporal
aspects of turnover. Longitudinal, episodic forms of research would
appear to be particularly well suited to the study of decisional
processes with strong evolutionary overtones, but turnover researchers
have relied almost exclusively on static methods of investigation
(347).
Steel concludes that the disciplines lack of progress could be repaired with the
addition of quality studies to validate the field.
His third criticism is that the majority of turnover research is on
occupationally homogeneous samples, which show less variance than occupationally
diverse samples. He recommends that the field pay more attention to those small
variances in order to expand the understanding of the turnover process. Clearly,
turnover studies must capitalize on all key sources of intrastudy variance if they are to
maximize predictive yields (358).
After reviewing the leading models in turnover theory, as well as the criticism
of past research, the framework developed by James L. Price (2001) was selected for
this study. Steels recommendation of paying more attention to small variances in an
occupationally homogeneous sample will be discussed in the chapter on the studys
31


results. However, this research will not attempt to apply Steels recommendations of
incorporating an evolutionary model or using longitudinal methodology due to
limited time and resources. Even though this project will not be able to overcome all
the weaknesses identified in this literature review, the strengths of Prices model,
especially its comprehensiveness, will be emphasized.
The Causal Model of Turnover
The Causal Model of Turnover (CMT) serves as a framework to understand
the determinants of voluntary turnover. The CMT was chosen for this study because it
has been well tested (Price, 1990; Price & Kim, 1993; Price, 1995; Kim, Price, &
Mueller, 1996; Iverson, 2000; Kim, 2000; Price, 2001, 2004), and it is an
explanatory model which includes variables from the three main perspectives:
economic, psychological, and sociological. This is different from the other existing
models, which are not explanatory (Price, 1990).
The CMT has been previously used to study the military and the medical
workforce. By adapting it to a new populationthe fundraising profession
additional knowledge will be available for future researchers. For example, this study
may discover that certain CMT variables are or are not applicable for this profession.
Future researchers may choose to add or delete variables from the model if their
population is similar to fundraisers.
32


Historically, turnover research predominantly focused on men and blue-collar
workers. More recently, research has expanded to study women and various service
industries (personal communication, Dr. James Price, October 27,2004). This is an
important advancement in the field because there is still much to learn about the
distinct differences found in certain populations. For example, women and men often
differ in factors such as pay, promotion, and turnover rates. Testing this and other
turnover models in various populations and industries is important for research to be
applicable to a wide-range of people. The fundraising profession, for example, has
been looking at the salary discrepancy between women and men for years (Boice,
2006), but has done very little empirical testing to better understand this phenomenon.
Turnover research originally asked study respondents their reasons for leaving
a job. James Price changed this type of research by asking respondents to describe
their work and then measuring leavers versus stayers. The CMT is based on this
method of research (personal communication, Dr. James Price, October 27, 2004). A
challenge to the former style of research is that after an employee leaves a job, it may
be difficult to find him or her to ask them why they left. Instead, using this new type
of research, a researcher can access all employees at one point in time and then, after
a period of time, analyze data from the surveys against those who left and those who
stayed.
There are 23 identified variables (or determinants) in this model (see Figure
2.3). They are categorized into exogenous (originating from without) variables;
33


intervening (coming between) variables; and endogenous (originating from within)
variables. The variables in the model are discussed in the order that Price (2001)
discusses them mostly by how they are organized in the model. They are not listed in
order of significance. The exogenous variables are listed vertically on the left and
connect to the intervening variables which are illustrated horizontally (see Figure
2.3).
There are three categories of exogenous variables: (1) environmental, (2)
individual, and (3) structural. There are two environmental variables: (1) opportunity
and (2) kinship responsibility. Opportunity is the availability of alternative
employment, a determinant often emphasized by economists. In this model, increased
opportunity leads to greater turnover. Kinship responsibility is the level of obligation
one has to local relatives such as spouse and children. It is hypothesized that an
increase in kinship responsibility reduces turnover. Price recently added two other
environmental variables to the CMT relating to kinship responsibility: the extent to
which an employing organization is responsive to kinship concerns and the extent to
which both the husband and wife have careers (Price, 2001).
The second category of exogenous variables are individual variables: (1)
general training, (2) job involvement, (3) positive affectivity, (4) and negative
affectivity. General training is the transferability of skills and education between
employers. The CMT shows that the higher the transferability of skills, the higher the
number of turnover. Job involvement is the exertion of effort on the job or the
34


degree to which an employee is willing to work (Price, 1997: 411). It is argued that
job involvement positively affects job satisfaction, which negatively impacts turnover
(Price, 2001). Positive and negative affectivity indicate the individuals feelings about
the job. Positive affectivity selectively perceives the favorable aspects; negative
affectivity selectively perceives the non-favorable aspects. Positive affectivity is
believed to indirectly decrease turnover through an increase in job satisfaction;
negative affectivity indirectly increases turnover through a decrease in job satisfaction
(Price, 2001).
The third category of exogenous variables are structural variables: (1)
autonomy, (2) procedural justice, (3) distributive justice, (4) stress, (5) pay, (6)
promotional chances, (7) routinization, and (8) social support. Autonomy is the
degree to which an employee exercises power in relation to his or her job. The
assumption is that autonomy decreases turnover through its positive relationship to
job satisfaction. Procedural justice or equality is defined as the degree to which rights
are applied equally to all employees.1
1 Price (2001) discusses procedural justice and recommends that it be used in turnover research the
evidence for procedural justices positive impact on job satisfaction is of such a magnitude as to
require its inclusion in future research (611-612). He also provides questions for a survey to measure
procedural justice. However, he does not include it in his model because of the lack of literature on this
determinant. Due to his recommendation that it be used for research, it is included for this study.
35


u>
Os
Figure 2.3: Causal Model of Turnover


Distributive justice or equity is the extent to which rewards and punishments
are associated with job performance. The CMT theorizes that both procedural justice
and distributive justice decrease turnover through positive impact on job satisfaction
(Price, 2001). Job stress is a determinant that includes four types of stress: resource
inadequacy, role ambiguity, role conflict, and workload. The CMT posits that job
stress increases turnover through its negative impact on job satisfaction. Pay includes
the salary and benefits employees receive for their services to the employer. This
model theorizes, as do most, that an increase in pay decreases turnover (Price, 2001).
Promotional chances refer to potential for vertical mobility within an
organization. The CMT argues that promotional chances indirectly decreases turnover
by positively affecting job satisfaction and organizational commitment. Routinization
is the level of repetitiveness in a job. Routinization increases turnover by its negative
impact on job satisfaction.
Social support is the availability of informal aid with job-related problems.
Price (2001) distinguishes three types of support: supervisory (support from
supervisor), peer (support from coworkers), and kinship (support from family and
friends). Social support (specifically supervisory) decreases turnover through its
positive impact on job satisfaction and organizational commitment.
The intervening variables include: (1) job satisfaction, (2) organizational
commitment, (3) search behavior, and (4) intent to stay. Job satisfaction is the extent
to which employees enjoy their job. Organizational commitment is the level of
37


loyalty one feels towards their employing organization. Both of these determinants
are believed to have a negative correlation with turnover.
Search behavior is the extent to which employees look for other job
opportunities. The CMT shows that search behavior increases turnover. Intent to stay
is the degree to which employees plan to continue at their current employment. The
CMT hypothesizes that intent to stay decreases turnover due to employees intentions
to stay in a job and not leave (Price, 2001).
James Price (2001) identified some concerns related to the CMT. He
acknowledges that demographic variables (gender, age, etc.) are not included in this
model. Demographic variables do not precisely indicate the conditions that produce
variations in turnover. That is because demographic variables are believed to affect
the identified determinants, but are not determinants themselves. For example, marital
status does not directly affect turnover, but it does affect kinship responsibility. Price
notes that demographic variables can be used as control variables in the data analysis
process. Another concern, or warning, identified by Price is that not all of the
exogenous variables affect turnover through job satisfaction, organizational
commitment, search behavior, and intent to stay. Determinants of opportunity,
kinship responsibility, and general training may, or may not, directly impact turnover.
Prices point is that there may be multiple paths from the exogenous variables to
turnover.
38


To implement the CMT into this research proposal, a few alterations have
been made. First, instead of turnover being the dependent variable, intent to stay is
the dependent variable. Due to the nature of this study, the subjects are individuals
from different organizations; this research is not based on one organization. In
addition, due to time and resource limitations associated with this dissertation project,
a longitudinal study cannot be conducted. Most research measures actual turnover a
year after the beginning of a study. Therefore, actual turnover cannot be measured.
However, intent to stay has been identified as the closest variable to turnover in the
causal chain of explanation (Price & Kim, 1993; Price, 2001; Thatcher, Stepina, &
Boyle, 2003). Because of this, scholars who wish to study turnover commonly focus
on intent to stay (Price & Kim, 1993: 125). Using intent to stay instead of turnover
as the dependent variable limits the researchers ability to identify those in the sample
who quit their jobs, even if they have a high sense of intent to stay, due to other
factors; such as, a spouse getting relocated out of state, the need to care for a family
member, or a sudden organizational change which creates a working environment that
is no longer desirable. Therefore, even though intent to stay is often used by scholars
when turnover cannot be measured, it is not as accurate as measuring actual turnover.
Second, the variable pay will be split into two variables, actual pay (money
and its equivalents which employees receive for their services to the employer) and
perception of fairness of pay (the degree to which employees perceive their pay as
39


being fair). Price only includes actual monetary pay; however, literature supports the
distinguishing perception of fairness of pay as an important factor in turnover (Kim,
2005).
The third alteration adds a few variables that relate specifically to fundraisers
turnover (intent to stay). Through an analysis of the limited research on fundraisers,
seven variables have been identified that are worth further investigation, including:
(1) fundraisers relationship with their organizations president and board (Duronio &
Tempel, 1997; Schwinn & Sommerfield, 2002); (2) availability of higher paying jobs
in the field (Duronio & Tempel, 1997; Schwinn & Sommerfield, 2002); (3) quality of
life needs (Anonymous, 2003; DRG, 2003); (4) the strength of the economy
(Anonymous, 2003; DRG, 2003); (5) self-interest motivators such as job title and
salary (Duronio & Tempel, 1997; Schwinn & Sommerfield, 2002; Anonymous, 2003;
DRG, 2003); (6) commitment to the organizations mission (Duronio & Tempel,
1997; Schwinn & Sommerfield, 2002; Brown & Yoshioka, 2003; Moore, 2004); and
(7) unrealistic fundraising goals.
The first factor, relationship with the president and board, can be partially
measured within the CMT determinant of social support. Support from the
organizations president is represented through supervisory support. Support from the
board of directors, however, is an important variable that is unique to the field of
fundraising. Fundraisers often rely on board members to assist them with asking
donors for money. If board members do not understand or acknowledge the important
40


work of fundraisers, fundraisers may feel unsupported and their tasks may become
more difficult because they are raising money without the assistance of the board.
Based on the possible influence of the support of the board of directors on
fundraisers intent to stay, board of directors support will be added as a fourth type
of social support.
The second factor, availability of other higher paying jobs, can be measured
within the CMTs determinant of opportunity. The third factor, quality of life needs,
includes decisions to leave a job to stay home with children or other family members.
This is represented in the model through kinship responsibility. The fourth factor, the
economy, will not be added as a determinant since this study was conducted at one
point in time. The fifth factor, self-interest motivators, is already represented within
the CMT through the promotion and pay determinants.
The sixth factor, commitment to mission, is similar to the variable
organizational commitment, but is not exactly the same. Organizational commitment
is more broadincluding a belief in an organizations goals and values, a willingness
to exert effort on behalf of an organization, and a strong desire to maintain
membership in an organization (Price, 1997). Commitment to mission focuses on the
mission only. For example, a fundraiser might have a strong commitment to an
organization that provides summer camp programs for kids, but (s)he is not really
passionate about the mission. That same fundraiser might be more passionate about
working on behalf of abused kids. (S)he could work for an agency that protects
41


abused children, and therefore, have a high commitment to mission, but a low
organizational commitment if (s)he does not support the goals of the agency in
fulfilling its mission. Because commitment to mission is not represented in the CMT,
it will be added as new variable with a positive correlation with intent to stay.
The seventh factor, unrealistic fundraising goals, is not represented in the
model so it will be added as a fifth type of job stress. This additional variable is in
response to informal conversations with fundraisers about job stress resulting from
their inability to accomplish the unrealistic fundraising goals set by their organization
(see Figure 2.4).
Hypotheses
The hypotheses in this study are based on the CMT and the aforementioned
adjustments. Due to the large number of variables in this study, the variables were
positioned into five different hypotheses and arranged by their positive or negative
impact on intent to stay, by their type of determinant (exogenous or intervening), and
whether or not a variable was part of the original CMT or newly added for this study.
The researchers goals in developing these hypotheses are to (1) find a manageable
way to interpret the variables and data within the context of this complex model and
(2) discover and understand the role of the significant determinants within the model.
42


Hi: The group of exogenous variables containing opportunity,
general training, negative affectivity, job stress (including
unrealistic fundraising goals), and routinization will have a
negative effect on intent to stay.
This hypothesis includes the exogenous variables that have a negative
correlation with intent to stay. All are in the original model, except for
unrealistic fundraising goals. The factor of unrealistic fundraising goals was
included because it is considered along with the other job stresses: resource
inadequacy, role ambiguity, role conflict, and workload.
H2: The group of exogenous variables containing kinship
responsibility, job involvement, positive affectivity, autonomy,
distributive justice, procedural justice, pay (actual salary), pay
(perception), promotional chances, and social support (coworkers,
supervisor, and family/friends) will have a positive effect on intent
to stay.
This hypothesis includes the exogenous variables that have a positive
correlation with intent to stay. All are in the original model.
43


H3: The group of intervening, endogenous variables containing
job satisfaction and organizational commitment will have a
positive effect on intent to stay.
This hypothesis includes the two intervening variables, job satisfaction
and organizational commitment, which have a positive correlation with intent
to stay. They are both in the original model.
H4: The intervening, endogenous variable of search behavior will
have a negative effect on intent to stay.
This fourth hypothesis includes the one intervening variable, search
behavior, which has a negative correlation with intent to stay. It is in the
original model.
H5: Commitment to mission and support of the board of directors will
have a positive effect on intent to stay.
This last hypothesis includes the remaining two of the three new variables,
commitment to mission and support of the board of directors. Both will have a
positive correlation with intent to stay. Support of the board could have instead been
included with Hypothesis 2 along with the other support determinants, but it was
44


included here because the fact that it is a new variable is weighted with higher inquiry
interest from the researcher than the fact that it is a support variable. The CMT and
the five hypotheses will be further analyzed in Chapters 4 and 5.
45


fc.
OS
Figure 2.4: The Adjusted Causal Model of Turnover Using Intent to Stay as the Dependent Variable


CHAPTER 3
METHODOLOGY
This chapter describes the demographics of this studys sample and outlines
the methodological approach to the study. First, a description of the data collection is
offered. Second, the samples demographics are compared to the demographics of a
sample of fundraisers involved in a previous study by the Association of Fundraising
Professionals (AFP) (Brown, Healey, Maehara, & Williams, 2005) to add insight to
how representative this dissertations sample is to the fundraising population. Third,
the changes made to the measurement tool are discussed. Fourth, a description of the
quantitative data analysis methods is offered. Fifth, reliability and validity issues
around this study are analyzed. Last, concluding remarks summarize the findings of
this chapter.
Data Collection
The methodology of this research was greatly influenced by James Price and
the CMT (Causal Model of Turnover). Methodological decisions mirrored those of
previous CMT studies in order to maintain as much of the validity and reliability of
the model and its measurement tool as possible. The research process followed a
quantitative approach.
47


Primary data were collected from fundraisers2 3 using a survey. Following past
research using the CMT, individuals were studied, not organizations. Fundraisers
were randomly identified from two sources. One source was the Association of
Fundraising Professionals. The second source was an institution that serves
fundraisers, which asked to remain anonymous. Therefore, it will be referenced as
the anonymous institution from here forward.
The Association of Fundraising Professionals was asked to be a source of
subjects because it is the leading organization that serves fundraisers through a
membership association. In addition, AFP sponsored this research project with a
grant, which illustrates their commitment to supporting research on the fundraising
population.
The researcher approached one other institution in addition to the Association
of Fundraising Professionals in order to obtain a sample that was not solely AFP
members. Previous research done by AFP has been criticized because it is limited to
their membership and cannot be generalized to the entire fundraising population.
There are only a small handful of institutions that have a significant database of
fundraisers. These participating institutions (AFP and the anonymous institution)
were the first two to be approached and were very responsive to assisting with this
study.
2 Fundraisers are defined as spending at least 50 percent of their job raising funds for a nonprofit
organization.
3 According to its website, For more than 40 years, AFP has been the standard-bearer for
professionalism in fundraising (AFP, 2006).
48


The anonymous institution was chosen because it has a large database
(approximately 26,000 fundraisers or people interested in learning about fundraising).
It provides education to fundraisers, from intermediate to advanced. The anonymous
institution does not track whether participants are AFP members or not.
Both AFP and the anonymous institution maintain databases of fundraisers
that represent all arenas of the nonprofit sector (education, health, social services, the
arts, etc.) and various sized organizations. At the researchers request, each of these
two sources supplied approximately 600 subjects through their own random computer
sampling. Approximately 70 subjects were discarded from the lists because they were
out of the country, and this study intended to focus on the United States. There was
only one duplicate name found between the two lists.
The two participating institutions provided names, titles, organizations, and
addresses of the subjects, but not e-mail addresses. A letter (see Appendix B) was
sent to a total of 1,129 subjects in June 2005 introducing them to the study and asking
them to access and participate in a Web-based survey at www.survevmonkev.com.
Prior to the mailing, the survey was piloted with approximately 30 participants.
Feedback on how to improve the survey was received and used to make changes
before the final mailing. These changes will be discussed later in the chapter.
A reminder postcard (see Appendix C) was sent to the 1,129 subjects two
weeks after the initial letter was sent. The letter and postcard encouraged people to e-
mail the researcher to request the results of the study: 100 people responded to this
49


opportunity. The data of 228 participants were collected from
www.survevmonkev.com six weeks after the final mailing, for a 20 percent response
rate.
Five of the 228 surveys were excluded from the study because respondents
selected b on question no. 3 (see Appendix D), indicating that they did not work for
a 501(c)3 nonprofit organization. This studys focus is fundraisers in the nonprofit
sector. Forty-five additional surveys were excluded based on their responses to survey
question no. 4: Answering a or b to no. 4 indicated that they spent less than 50
percent of their time on fundraising. This studys focus is on those employees who
consider themselves fundraisers and spend the majority of their job doing fundraising.
An additional nine surveys were discarded because respondents did not answer a
majority of the questions. The final sample size was 169.
Sample
The sample of 169 is being compared to that of the 2005 AFP Compensation
and Benefits Study of its membership (Brown et al., 2005). The Association of
Fundraising Professionals study (United States respondents only) is being used
because AFP is the leading organization doing research on fundraisers and has access
to the fundraising population. In fact, this is the only recent report on fundraisers
available to use as a comparison. The main difference between AFPs sample and this
studys sample is that all of the 664 fundraisers (a 22 percent response rate) from the
50


AFP sample are AFP members; this dissertation study incorporated both AFP
members and non-members. The study participants were not asked whether they were
AFP members, but it is assumed that the sample includes both AFP members and
non-members since 50 percent of the initial pool came from AFPs database. It is
unknown what percentage of fundraisers from the anonymous institution are AFP
members as this question was not included in the survey.
Some demographic variables were included in the survey measurement,
however, most are not used in the CMT. Demographic variables were included not
for their theoretical significance but for their ability to create a comprehensive dataset
and give additional information about these subjects. For example, if a researcher
wanted to do additional research on the fundraising profession using this dataset,
demographic variables could be helpful in detailing the sample population and
looking at different research questions.
The demographic variables to be discussed are gender, ethnicity, age, salary
(also an independent variable), years of experience, and education. The representation
of nonprofit sub-sectors also will be reviewed. The other demographic variables not
mentioned here are marital status and number of children, which will be discussed in
Chapter 4 in regards to the independent variable kinship responsibility.
The gender distribution in this study is 75.4 percent (N=126) female and 24.6
percent (N=41) male (see Table 3.1). In the 2005 AFP survey, 66.7 percent of
respondents (from the United States) were female, 32.1 percent male. Of the 169
51


respondents in the current study, 167 answered the question on ethnicity (see Table
3.2). Less than 10 percent of this studys sample are non-white. Hispanic/Latino was
the second most frequent ethnicity with six respondents. The ethnic demographics are
very similar to the ethnic demographics of AFPs study.
Table 3.1: Sample by Gender
Valid Percent
Gender Dissertation AFP
Valid 1 (Female) 75.4 66.7
2 (Male) 24.6 32.1
Total 100.0 98.8
N 167 656
Table 3.2: Sample by Ethnicity
Valid Percent
Ethnicity Dissertation* AFP
Caucasian, not Hispanic 92.2 94.1
Asian or Pacific Islander 2.4 0.8
Hispanic/Latino 3.6 1.5
Multi-Ethnic 0.6 0
Other 1.2 3.3
Total 100.0 98.8
N 167 656
*Not Valid Due to Low Representation: African American and Alaskan Native
52


The median age of this studys sample is 39 and the mean age is 40.71 (see
Table 3.3). This sample is younger than that of AFPs study. Over 65 percent of AFP
respondents were between the ages of 45 and 64; only 35 percent of this studys
sample are in that bracket. One reason for this difference could be related to AFPs
hypothesis that AFP members are weighted toward older practitioners (Brown et al.,
2005), and this sample includes people who are not AFP members.
Membership in AFP entails paying a fee, and therefore, it may be more likely
for fundraisers in established, stable, larger nonprofits to be members (assuming their
employer is paying the fee for them). Small grassroots organizations may not have the
funds to pay for an AFP membership. Established nonprofits can attract higher
experienced fundraisers, who are likely to be older than non-experienced fundraisers.
The anonymous institution may be serving younger fundraisers than those that AFP
serves. This anonymous institution does not have a membership fee, only fees for the
cost of the education programs, which may attract people who are newer to the
profession, and therefore, likely to be younger on average.
53


Table 3.3: Sample by Age
Valid Percent
Age Dissertation AFP
Under 25 4.8 0
25-34 32.1 4.2
35-44 27.4 23.2
45-54 20.4 35.2
55-64 15.0 31.8
65 or more 1.2 3.9
Total 100.8 98.3
N 169 653
Salary was measured in increments of $20,000starting at $30,000 or less to
$130,000 or more (see Table 3.4). This study found a mean salary of $56,712.5 This
is significantly less than the mean salary in AFPs sample, which was $80,685.
Unfortunately, AFP data on salary were not available in a similar form for
comparison purposes. The difference is probably due to this studys having a younger
sample. It is also likely to be correlated to years of experience in the field. Over half
(64.1 percent) of this studys sample reported having nine years or less of experience
in fundraising (see Table 3.5). This is a high percentage compared to AFPs sample
5 Taking the mid-point of each salary level and multiplying it by the frequency of respondents, adding
them together, and dividing by N (total number respondents).
54


with only 13.7 percent of fundraisers having nine years or less experience. This issue
of years of experience will be discussed further in Chapter 4 where average length of
service is examined.
Table 3.4: Sample by Salary
Valid Percent
Salary Dissertation*
<$30K 5.4
$30-49K 43.7
$50-69K 28.7
$70-89K 13.2
$90-109K 5.4
$110-129K 1.8
>$130K 1.8
Total 100.0
N 167
* AFP raw data for this demographic were not available
Table 3.5: Sample by Years of Experience
Valid Percent
Years of Experience Dissertation AFP
0-4 42.5 2.7
5-9 21.6 11
10-14 14.4 22.4
15-19 9.0 24.4
20-24 10.8 19.9
25-29 0.6 11.6
30 or more 1.2 7.2
Total 100.0 99.2
N 167 659
55


In a comparison of education level, the two samples are quite similar. In this
studys sample, 43.8 percent (N=74) completed college and 37.3 percent (N=63)
completed graduate school (see Table 3.6). The Association of Fundraising
Professionals sample is also well-educated, with 35 percent having a college degree
and 45 percent having a masters degree. This similarity reinforces the observation
that the main differences between the two samples are age and tenure, not education.
Table 3.6: Sample by Education Completed
Valid Percent
Education Level Dissertation* AFP
Completed Grade School 0.6 0.5
Some College 3.0 3.9
Completed College 43.8 35
Some Graduate School 15.4 15
Completed Graduate Degree 37.3 45
Total 100.0 99
N 169 655
*Not Valid: Completed High School
The types of nonprofit sub-sectors with the highest representation in this
studys sample and AFPs study are education, social/human services, and health
organizations (see Table 3.7). In this studys sample, 72.70 percent (N=123) of the
56


respondents fall into these three sub-sectors; a little more than the 61 percent in
AFPs sample. Comparing these samples to the nonprofit sector in general shows that
these fundraisers under-represent the health and religious sub-sectors, and over-
represent the arts/culture sub-sector. This may be due to the fact that many health
related organizations have a large portion of their budget from patient fees and public
sources and do not need a large fundraising staff. Likewise, religious organizations
receive donations during weekly services, and do not need specialized fundraising
staff. The arts/culture sub-sector may a good example of nonprofits that need
substantial fundraising staff due to limited commercial activities and public funding.
Table 3.7: Sample by b PO Sub-Sector
Valid Percent
NPO Sub-Sector Dissertation* AFP Nonprofit Sector**
Arts/Culture 8.3 8.3 1.9
Advocacy/Law 1.8 1.4 .2
Religious 4.1 4.2 11.8
Trade, Professional or Association 2.4 0.9
Health/Medical 19.5 19.0 41.9
Human/Social Services 25.4 18.4 18
Education 27.8 25.3 21.9
Environmental 4.7 2.4
Scientific Research 0.6 0.6
Other 5.3 17.53 4.2
Total 99.9 98.03 99.9
N 169 657
*Not Valid: Federation Appeals
** (Independent Sector, 2001)
57


In conclusion, the most significant finding in the comparison between this
studys sample and AFPs sample are the differences in age, compensation, and
tenure. The surprise is not that there is a correlation between these variables, but that
these two samples would have this significant of a difference, particularly in age. In
their report, AFP (2005) includes anecdotal evidence that its membership represents
an older, more tenured group of fundraisers. It is possible that the anonymous
institution attracts younger, less experienced fundraisers. The survey tool used in this
study did not ask respondents if they are AFP members, so it cannot be statistically
tested here. Future research should consider this question to better understand if, and
why, AFP attracts older fundraisers.
These findings also reinforce the fact that fundraising is still developing as a
profession and that describing the true population of fundraisers is challenging. Data
on this population are limited to a few institutions. Until a clearinghouse for
fundraisers, or some way to track these professionals is implemented, it is difficult to
generalize this research to the entire population.
Measurement Tool
James Prices (2001) measurement tool used in the CMT was replicated as
much as possible in order to preserve the reliability of the tool in measuring the
impact of the determinants. A few adjustments were made to Prices tool to
58


appropriately adapt it to this study. The first adjustment was in the number of
questions per variable. Instead of using four to six items (questions) per variable, this
survey used primarily two, and sometimes three, resulting in a total of 75 questions.
The process of selection included finding the two or three items per variable from
Prices tool that were the most different from one another, and often selecting a
positive response and a negative response. Without this adjustment, the survey tool
would have been too lengthy, which might have negatively impacted the response
rate (Dillman, 2000).
The second adjustment was to include questions for the new variables:
commitment to mission, support of the board, and unrealistic fundraising goals. In
order to maintain the validity of the tool as much as possible, wording used in the
survey tool for these items was taken from items used in similar variables. For
example, question number 40 measures the variable support from coworkers: I
enjoy working with most of my coworkers. That same language was used for question
number 38 to measure the new variable support from the board of trustees: I enjoy
working with most of the people on the board of trustees.
The third adjustment was the addition of an introduction. The introduction
included occupational-specific items to determine whether or not the respondents
worked for a nonprofit, which sub-sector they worked in, if they were self-employed
or employed by an organization, and if the respondents fundraised at least 50 percent
of their time. This section was added because the studys focus is on people who
59


work in the nonprofit sector and consider themselves fundraisers (defined as
fundraising at least 50 percent of the time). The sub-sector question determined what
type of nonprofit (education, social services, etc.) they worked for to help create a
comprehensive dataset. It is common for practitioners and scholars to perceive, or
analyze, nonprofits based on their type. Even though this research did not analyze the
differences among the sub-sectors, future research may want to look at that more
closely.
The fourth and fifth adjustments were the addition of two questions: (1) to
measure the average length of service (to be discussed in detail in Chapter 4) and (2)
question No. 68 asking respondents why they left their last job (see Appendix D).
The sixth adjustment was the addition of basic demographic information
questions at the end of the survey (as discussed at the beginning of this chapter). The
final adjustments were minor changes to the language used for a few questions and
the clarification of confusing questions based on feedback from those in the pilot
study. For example, using the wording current employer instead of present
employer. Another example of a question that was unclear to the pilot study
participants was: Compared to other people, lam more likely to think of my job as a
career. Based on the feedback, the researcher changed the wording to: I view my job
as part of my career*. (* Career is defined as a line of work that an employee plans
to pursue for many years.)
60


Data Analysis
A formula was created to answer the research question, What is the average
length of service among fundraising professionals? The formula used to determine
the average length of service is the difference between current age and age when
entered the fundraising profession divided by the number of fundraising jobs held
(see Figure 3.1). The average length of service among the 169 fundraisers in this
study was calculated using Microsoft Excel.
Age now Age when entered the profession
Number of fundraising jobs
Figure 3.1: Average Length of Service Formula
Average length of
service
Multiple regression analysis was used to answer the research question, What
are the main variables predicting fundraisers intent to stay? Multiple regression
analysis can estimate the effects of each of the determinants in the CMT on intent to
stay. Primary data from the surveys were downloaded from www.survevmonkev.com
into a Microsoft Excel format. From there the data were inputted into SPSS for
analysis. A data journal was kept throughout this process to track the reversing and
recoding of date (see Appendix E).
61


Most questions used a Likert scale response option (strongly agree, agree,
neither agree nor disagree, disagree, or strongly disagree). Responses ranged from
five to one (5=strongly agree, l=strongly disagree). Responses were reversed for
negative items (see Appendix F for the Research Matrix). Per Prices (2001)
recommendation, the responses were averaged instead of summed to get a score for
each variable for the purpose of comparison.
Reliability and Validity
Reliability and validity were considered in this research. According to Price,
the measures generally possessed acceptable validity and reliability (2001: 617).
Factor analysis was used to provide data about discriminant and convergent
validity. Most of the research, especially since about 1980 used Confirmatory
Factor Analysis provided by LISREL. Reliability is assessed with alpha. The
measures generally possessed acceptable validity and reliability. The
psychometric properties of the measures are available in the sources cited (Price,
2001:617).
Even with the adjustments made to the instrument, the researcher feels that the
instrument closely reflects the survey suggested by the CMT (2001).
As in many research projects, there are threats to validity. Findings that come
from a study that have external validity may be generalizable to other cases or
populations (O'Sullivan & Rassell, 1999). In this research design, the use of random
sampling has helped increase external validity. However, because the sources of the
subjects were institutions that serve fundraisers in exchange for a fee, it is likely that
62


part of the fundraising population who cannot afford, or choose not to participate in,
these types of institutions will not be represented in this study. Until there is a way to
track the majority of fundraisers in this country, it is difficult to generalize this
research to the entire population. Hopefully, future research will continue to seek out
ways to identify and understand the actual population of fundraisers.
Conclusion
In comparison to the AFP 2005 Compensation and Benefits Study, this
dissertations sample is younger, less experienced, and makes less money. The main
reason for the differences is most likely related to the fact that this sample was not
solely AFP members. Therefore, future research done on the fundraising population
should include both AFP members and non-members to strengthen the ability to
generalize findings to the entire fundraising population. It is also recommended that
future studies find a way to include fundraisers who come from nonprofit
organizations that cannot afford educational programs and other fee-based services
for their development staff. Those people may not be well represented in this sample,
nor that of AFPs studies. This is a concern for this researcher because if people in
smaller, more grassroots organizations were not well represented in this research, the
results may only represent those who work in larger, more developed nonprofits and
may differ in their intent to stay. This and other possible limitations of this research
will be further discussed in Chapter 5.
63


CHAPTER 4
RESULTS
This chapter discusses the results of the data analysis. The first research
question (What is the average length of service among fundraisers?) is analyzed
using a simple mathematical formula. The second research question {What are the
main variables predicting fundraisers intent to stay?) is analyzed using multiple
methods. First, all of the independent variables are run through multiple regression
analysis. In order to determine the most significant variables, a backward elimination
regression is used. The variables that survive the elimination process are discussed in
further detail. Second, the five hypotheses are tested to determine the number of
variance in intent to stay that each group explains. The variables that are
statistically significant are discussed. Third, the intervening variables (job
satisfaction, organizational commitment, and search behavior) are used as dependent
variables and regressions are run to test the path of the exogenous (environmental,
individual, and structural) variables as illustrated by the adjusted Causal Model of
Turnover (CMT). Lastly, a conclusion reviews the findings. 7
7 Backward elimination (or backward deletion) is a form of statistical (or stepwise) regression that
takes all independent variables into the statistical equation and then are deleted one at a time if they are
not significant in the regression, leaving only those variables that are significant (Tabachnick & Fidell,
2001: p 117).
64


What is the Average Length of Service?
The average length of service for this studys sample is 3.66 years (see Table
4.1). This formula does not capture those fundraisers who may have been out of the
workforce to care for family, or to take time off for education. A limitation of this
formula is that it does not take into account those who have not had a continuous
employment experience. People take off from working for periods of time for
various reasons; unfortunately, this formula does not capture that reality. Therefore,
the average length of service of 3.66 years for this sample may be inflated.
Table 4.1: Average Length of Service Descriptive Statistics
Average Length of Service
Average All 3.66
Females 3.50
Males 4.17
N 169
In 1997, Duronio and Temple found that the average length of service for
female fundraisers was 3.31 years and 4.42 for males. This dissertation project found
the average length of service for females was 3.50 years and 4.17 for males. This
indicates that the average length of service is still lower for women than men, but it
has not changed much within the profession since 1997. Looking at previous
research, it may be possible that the average length of service is increasing. In 2002,
AFP surveyed 1,215 people who, on average, said they had been in their current jobs
65


for three years or less (Schwinn & Sommerfield, 2002). In 2003, a New York search
firm, DRG, surveyed fundraising professionals about job retention: 19 percent stayed
5-6 years, 52 percent stayed 3 to 4 years and 16 percent stayed 1 to 2 years (DRG,
2003). These studies are consonant, although not statistically comparable with the
findings of this research.
A 2004 study conducted by AFP found that 61 percent of fundraisers plan to
stay in their jobs indefinitely, up from 42 percent in 2003 (Hall, 2004). The possible
recent shift in intent to stay found in the research of AFP, DRG, and this dissertation
may be related to the halt in growth in the nonprofit sector workforce. Since 1990, the
average employment growth in the nonprofit sector has been 2.4 percent. In the year
ending July 2004, the rate had fallen to 0.5 percent (Irons & Bass, 2004). Other
possible reasons for the change in the average length of service include a maturing
profession, the improved pipeline that is feeding trained fundraisers into nonprofit
organizations, and the trend of nonprofits merging. These hypothesized reasons will
be further discussed in Chapter 5.
The 1997 study by Duronio and Temple looked at fundraisers with 10 or more
years of experience in the profession and found the average length of service
increased to 4.6 years for women and 5.72 years for men (from 3.31 years and 4.42
years with less than 10 years, respectively). This study found that those with 10 or
more years experience in the profession had a significantly longer length of service of
6.48 years compared to those with less than 10 years of experience (an average of
66


only 2.11 years). Unfortunately, this dissertation does not confirm that there has been
a recent increase in length of service among fundraisers. A longitudinal study would
need to be done to confirm this theory.
What are the Main Variables Predicting Fundraisers Intent to Stay?
Before analyzing the variables within their respective hypotheses, a backward
elimination multiple regression analysis was run on all of the 26 independent
variables. See Table 4.2 for definitions of the variables. See Table 4.3 for the
descriptive statistics of all 26 independent variables and the dependent variable.
Testing the Assumptions for Multiple Regression
The following assumptions were tested prior to running a multivariate
regression: (1) linearity, (2) homoscedasticity, (3) normality, and (4) independence.
Intent to stay was analyzed through histograms, which indicated a normal curve.
Though it was slightly skewed negatively (-.690), it did not require transformations.
The correlations table (see Table 4.4) confirms that there is no threat of
multicollinearity (the highest correlation between variables is job satisfaction and
promotional chances at .709).8
8 Correlations are assumed to be low (.1-.3), moderate (.4-.6), and moderately high to high (.6-1)
(Goodwin, 2003).
67


Table 4.2: Variable Definitions9
Variable Definition
Intent to Stay Intent Extent to which employees plan to continue member- ship with their employer
Job Satisfaction JobSatisfaction Extent to which employees like their work
Autonomy Degree to which an employee exercises power relative to his/her job
Commitment to Mission Mission Extent to which employees have loyalty to the mission of their employer*
Organizational Commitment OrgCommit Extent to which employees have loyalty to their employer
Routinization Extent to which jobs are repetitive
Distributive Justice DistJustice Extent to which rewards and punishments are related to job performance
Procedural Justice ProcedJustice Degree to which rights are applied universally tc all employees
Promotional Chances PromoChance Degree of potential occupational mobility withir an organization
Job Involvement Joblnvolve Willingness to exert effort on the job
Job Stress (Inadequate Resources) StressIR Lack of means to perform a job
Job Stress (Conflict) StressConflict Inconsistent job obligation
Job Stress (Ambiguity) StressAmb Unclear job expectations
Job Stress (Workload) StressWork Number of effort required by a job
Job Stress (Unrealistic Fundraising Goals) StressGoal Unrealistic fundraising goals*
Job Support (Supervisor) SupportSup Assistance with job-related problems from supervisor
Job Support (Family and Friends) SupportFamFr Assistance with job-related problems from fami and friends
Job Support (Coworkers) SupportCoWo Assistance with job-related problems from coworkers
Job Support (Board of Trustees) SupportBOT Assistance with job-related problems from the board of trustees*
9 Definitions with an are the authors. All others are quoted directly from Price (2001).
68


Table 4.2: Variable Definitions Continued
Variable Definition
General Training GenTraining Extent to which the knowledge and skills required for a job are transferable between employers
Opportunity Availability of alternative jobs in the environment (locally and non-locally)
Positive Affectivity PositiveAfF Dispositional tendencies to experience pleasant emotional states
Negative Affectivity NegativeAff Dispositional tendencies to experience unpleasant emotional states
Search Behavior Search Degree to which employees are looking for otherjobs
Perception of Fairness of Pay PayPercep Degree to which employees perceive their pay as being fair*
Kinship Responsibilities KinshipResp Existence of obligations toward relatives living in the community
Salary Money and its equivalents which employees receive for their services to the employer
Table 4.3: Descriptive Statistics: Mean and Standard Deviation for All Variables
Variables Mean Standard Deviation
Intent 3.5483 1.02376
JobSatisfaction 3.9911 .97051
Autonomy 3.8876 .79756
Mission 4.0266 .80041
OrgCommit 3.9941 .75787
Routinization 2.2278 .79799
DistrJustice 3.3757 .74758
ProcedJustice 3.3373 1.07259
PromoChance 3.4083 1.01133
Joblnvolve 3.1243 .81244
69


Table 4.3: Descriptive Statistics: Mean and Standard Deviation for All
Variables Continued
Variables Mean Standard Deviation
StressIR 2.7160 .87921
StressConflict 2.6686 .85197
StressAmb 2.6627 1.01704
StressWork 3.9704 .81230
StressGoal 2.6746 .89509
SupportSup 3.6864 1.04049
SupportFamFr 4.2101 .73514
SupportCoWo 3.9615 .80178
SupportBOT 3.3314 .74978
GenTraining 4.4941 .59509
Opportunity 3.0000 .99253
PositiveAff 3.8432 .70961
NegativeAff 3.1953 .90979
Search 2.5325 .89333
PayPercep 2.8195 1.04439
KinshipResp 3.3228 .80576
Salary 2.82 1.194
N= 169
Handling of Missing Data
There were only two questions on the survey that produced a significant
percentage of missing values. Question No. 57 asked respondents about the career of
their significant other: they were instructed to skip the question if they did not have a
significant other. Thirty-three respondents skipped this question.
Question No. 75 asked respondents if they have children living at home.
Participants were instructed to answer the question only if they have children. Fifty-
four of the 82 participants who did not have children skipped this question. The other
28 may have misunderstood the instructions and answered the question anyway, most
70




likely as not having children living at home. However, there may have been
participants who do not have children of their own, but have children living in their
home. A weakness of this question is that it was not clear for people in this latter
situation.
Taking these two questions (No. 57 and No. 75) out of the missing data
analysis, less than 1 percent of the values had missing data. However, because theses
data are non-continuous, the answers were left blank.
Explanation of Various-Data Analyses in Chapter 4
Due to the researchers interest in looking at the 26 variables in a variety of
ways, an explanation of this chapters order of analyses is helpful. This chapter
introduces the bigger analyses first (correlation table and backward elimination
multiple regression analysis of all 26 variables) and then moves to the smaller
analyses of each of the five hypotheses. It concludes with a more complex analysis
which follows the various paths that the 26 independent variables take to intent to
stay (using job satisfaction, organization commitment and search behavior as
dependent variables). Explanations of findings will be given throughout the chapter.
71


Multiple Regression Analysis of the 26 Independent Variables
Due to the large number of independent variables in this analysis, a backward
multiple regression was rim to eliminate those independent variables that are not
significant. An alpha of .10 will be used for all of the statistical analyses. The small
sample size of this research results in low power; therefore, the alpha is set at .10 to
increase the power. Seven variables survived (stayed in) the backward elimination
analysis: job satisfaction, commitment to mission, distributive justice, promotional
chances, job involvement, support of supervisor, and search behavior (see Table 4.5).
Table 4.5: Results from Backward Elimination Regression Analysis (Displaying
Standardized Betas)
Beta
JobSatisfaction 29i ***
Mission in **
DistrJustice .120 **
PromoChance .247 ***
Joblnvolve .075 *
SupportSup .097 **
Search -.257 ***
F value 77.822 ***
N 169
Adjusted^ .764
a. Dependent Variable: Intent
b. 05 enter, .10 removed
c *** p< .0001; ** p<.05; *p<.10
Most of these significant variables are consistent with the Causal Model of
Turnover (CMT) literature. Discussing in order of beta weights, the highest variable
74


2 4.4: Correlation for Predictor Variables on Intent to Stay
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
Intent - - - - - - - -
0.782 0.492 0.442 0.569 0.413 0.564 0.388 0.722 0.432 0,385 0.398 0.512 0 164 0.308 0.494 0,059 0.372 0 377 0.261 0.185 0.325 0.178 0.664 0.287 0.139 0.117
Jobsatisfaction - - - - - - - -
0.641 0.362 0 543 0.397 0.526 0.394 0.709 0 334 0434 0.393 0.582 0.165 0.395 0.462 0.063 0 346 0.431 0.281 0.258 0.409 0.224 0.556 0.236 0.188 0.179
Autonomy 0.198 0.330 0.427 0.438 0.370 0.513 0.217 0.320 0.425 0.514 0.086 0.287 0.315 0.109 0.235 0.278 0.252 0.175 0.316 0.098 0.275 0.250 0.196 0.256
Mission - - - - - - - -
0.427 0.182 0.192 0.205 0293 0.469 0.112 0 115 0 161 0098 0.210 0.158 0.092 0 178 0.387 0 167 0.096 0.275 0.121 0.312 0039 0.152 0.007
OrgCommit 0.254 0.414 0.457 0.466 0.361 0 229 0.373 0 476 0.202 0.244 0 500 0.007 0.312 0 325 0.136 0.069 0.259 0.003 0.409 0.102 0.131 0.047
Routinization - - - - - - - - - - - - - -
0.463 0.270 0.306 0.155 0.156 0.269 0.374 0.223 0.221 0.317 0,122 0.131 0.246 0.267 0.150 0.380 0.154 0.235 0.146 0.243 0.329
DistrJustice - - - - - - - -
0.464 0.499 0.237 0,382 0.324 0.446 0 167 0.213 0.318 0.041 0.343 0.282 0.199 0.177 0,197 0.125 0.380 0.273 0 066 0.189
ProcedJustice - - - - - - - -
0 365 0.216 0.263 0.367 0.398 0.004 0.186 0.287 0.067 0.382 0.256 0.147 0.094 0.124 0.158 0.245 0.176 0.049 0.107
PromoChance - - - - - - -
0.279 0.419 0.384 0.556 0.126 0.272 0 440 0.110 0 190 0.351 0.253 0.194 0 318 0.200 0 438 0.293 0.094 0.122
Joblnvolve - - - - - - -
0.195 0.116 0.123 0.262 0.026 0.208 0.089 0.241 0.177 0.037 0.120 0.100 0.059 0.350 0.057 0.062 0.024
StressIR - - - - - - -
0.461 0 452 0.156 0.556 0.324 0.008 0.282 0.314 0.129 0.162 0.134 0.249 0.277 0.303 0.096 0.143
StressConflict - - - - - - - - -
0.670 0.155 0.412 0.480 0.063 0 309 0.334 0.106 0.062 0.176 0.228 0.230 0.290 0.014 0.133
StressAmb - - - - - - - - - -
0.136 0 443 0 548 0.124 0.280 0 472 0 232 0 116 0.313 0.226 0 311 0 266 0.132 0 264
StressWork 0.132 0 031 0.003 0.030 0 076 0.158 0.046 0,238 0.275 0.200 0 217 0.124 0.033
StressGoal - - - - - - - -
0.310 0.067 0.205 0.383 0.131 0.071 0.206 0.278 0.166 0.169 0.140 0.110
SupportSup 0.181 0.248 0.325 0.147 0.061 0.176 0.223 0.320 0.159 0.060 0,025
72


; 4.4: Correlation for Predictor Variables on Intent to Stay Continued
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
SupportFamFr 0.130 0.191 0.163 0.090 0.102 0.030 0.069 0.004 0.054 0.053
SupportCoWo 0.216 0,147 0.077 0.166 0.102 0.398 0.109 0.013 0.027
Support BOT 0,161 0.008 0.287 0.207 0.202 0.191 0.193 0.117
GenTraining 0.003 0.373 0.132 0.162 0.111 0.050 0.103
Opportunity 0.062 0.043 0.156 0.204 0.132 0.205
PositiveAff 0.280 0.160 0.075 0.277 0.194
NegativeAff 0.053 0.107 0.040 0.127
Search 0.223 0.056 0.025
PayPercep 0.083 0.367
KinshipResp 0.254
Salary
73


is job satisfaction, which is frequently cited in turnover literature as having a
relatively important effect on turnover (Kim, Price, & Mueller, 1996; Price, 1997).
There were two items used in the survey to measure job satisfaction: (1) lam very
satisfied with my job, and (2) I do not find enjoyment in my job.
Responses to these items, and those in the upcoming discussion, were based
on a 5-point Likert scale from strongly agree to strongly disagree. There is no clear
reason why fundraisers would be different from people in other occupations or
populations regarding the effect of job satisfaction on their intent to stay in a job.
The second highest variable is search behavior which, as predicted, had a
negative impact on intent to stay. There were two items used in the survey to measure
search behavior: (1) I rarely seek out information about job opportunities with other
employers, and (2) I almost always follow up on job leads with other employers that I
hear about.
The turnover literature includes a variety of theories regarding the effect of
search behavior and where it is in the causal chain, as discussed in Chapter 2. The
focus on search behavior is viewed by Robert P. Steel as a limitation (Steel, 2002).
He found that when looking at all turnover research, the correlation with search
behavior and turnover is r=.13. The correlation found here (r =.664) is significantly
higher. Note that this study regressed search behavior on intent to stay, not on
turnover, which could explain for the disparity between the two correlations
(.13 and .664).
75


Another explanation for the large negative beta and high correlation of search
behavior in this study is the fact that it is very likely that this sample of fundraisers
has easy access to the Internet, a popular venue for job announcements. Future
turnover research should consider technological capabilities when looking at the
importance of search behavior. It is much easier and more convenient to search for
jobs than it was 10 years ago.
The variable with the third highest beta is promotional chances. The two items
used in the survey instrument to measure promotional chances were: (1) I am in a
dead-end job, and (2) There is a good chance to get ahead with my employer.
Previous turnover research has validated the importance of promotional
chances as a predictor of turnover, especially research by sociologists (Kim et al.,
1996). Indeed, for most fundraisers a promotion is the only way to make more money
and have a more prestigious title, if they want to stay with their organization. Duronio
and Tempel (1997) predicted this to be important to fundraisers a decade ago.
The next largest variable is distributive justice (or equity) within an
organization. The two items used in the survey instrument to measure distributive
justice were: (1) Promotions by my employer are almost totally based on seniority,
and (2) Competent employees are well rewarded by my employer.
This finding is interesting because distributive justice is not overly
emphasized in the CMT literature nor the nonprofit literature. Distributive justice is
the concept of being rewarded for good performance. Furthermore, it relates to how
76


others in the organization are being rewarded for their performance and the equity of
rewards for performance among all employees. This indicates a similar motivation to
that of promotional chances. Being recognized and rewarded by the organization
seems to be important to this sample of fundraisers.
The reward system for many fundraisers is not the same as for personnel in
the for-profit sector. Many fundraisers, especially those who are members of AFP,
follow a strict ethics policy that indicates they cannot be paid through a commission-
type arrangement. Whether someone raises $100,000 or $1 million, that individual
should be salaried or paid by the hour. A fundraiser should not get a cut or percentage
of funds raised. Chapter 5 offers suggestions for nonprofit managers on how to ensure
that fundraisers have a high sense of distributive justice.
The significance of distributive justice and promotional chances indicates that
rewards and career advancements are important to this sample of fundraisers. The
determinants (distributive justice and promotional chances) are mildly correlated
(.499); therefore, there is not an issue of multicollinearity.
The next variable in rank is commitment to mission. This new variable was
added to the CMT because of its predicted importance for employees in the nonprofit
sector. Very little empirical research has been done on commitment to mission, so it
is noteworthy to see that it came out ahead of so many other variables. Commitment
to mission has one of the highest means (4.03) of the independent variables (see
Table 4.3). It confirms the hypothesis of practitioners who argue that fundraisers do
77


what they do because they are dedicated to the mission of the organization for which
they raise money. The two items used in the survey instrument to measure
commitment to mission were: (1) lam passionate about the mission of the
organization I work for, and (2) If I had to choose one mission to be affiliated with, it
would be that of my current employer.
The next variable is support from supervisor. The significance of supervisory
support is not very surprising as both the CMT literature and nonprofit literature
discuss the importance of the relationship between employee and supervisor. Looking
at Table 4.3, support from supervisor has one of the highest standard deviations
(1.04), indicating that supervisory support meant a lot to some respondents, and little
to others. Support from supervisor was measured via two items in the survey: (1) My
immediate supervisor cannot be relied on when things get tough on my job, and (2)
My immediate supervisor shows a lot of concern for me on my job.
According to the CMT, supervisory support affects intent to stay through a
correlation with both organizational commitment and job satisfaction. This indirect
correlation with supervisory support and intent to stay will be discussed later in this
chapter. In many organizations the fundraising staff is small in comparison to the
programmatic staff, which makes support from a supervisor even more important for
fundraisers to ensure that they do not feel isolated or unimportant. In addition,
because fundraising has been referred to as an art not a science (Grayson, 2002),
fundraisers may appreciate extra support from their supervisors in decision making
78


and strategizing. Employees in scientific professions such as accounting, may feel
that they have more control over the outcomes of their efforts. Employees in artistic
professions, such as fundraising, may feel that they have less control over the results
of their efforts. A fundraiser may spend lots of time cultivating a likely prospect,
using all the right tricks of the trade, and yet, never see a donation from that prospect.
The final variable is job involvement, which is the degree to which an
employee is willing to work. Synonyms for job involvement found in the turnover
literature include motivation, central life interest, and the Protestant work ethic
(Price, 2001: 605). Job involvement was measured via two items in the survey: (1)
Most of my interests are centered around my job, and (2) I have strong ties with my
present job which would be difficult to break.
The significant variables in this research thus far indicate a theme among this
sample population of fundraisers that focuses on their advancement and career path
within the nonprofit sector. Job involvement, promotional chances, distributive
justice, and commitment to mission support the notion that these employees want the
opportunity to move up in their organization, be busy and involved, get rewarded
fairly for their efforts, and advance the mission of the organization.
It is surprising that organizational commitment, which is an intervening
variable along with search behavior and job satisfaction, is not significant in this
study. It had a moderately high mean at 3.99 (see Table 4.3). The survey asked the
79


following two questions to measure organizational commitment: (1) I do not really
care about the fate of my current employer (the organization), and (2) My current
employer inspires the very best in me in the way of fob performance.
There are a variety of instruments in the literature that can be used to measure
organizational commitment (Price, 1997). The measurement tool in this study used
two items to measure this variable, which may not have been enough, and therefore,
problematic for such a comprehensive determinant. Organizational commitment
includes a variety of concepts such as a strong belief in and an acceptance of the
organizations goals and values, a willingness to exert considerable effort on behalf of
the organization, and a strong desire to maintain membership in the organization
(Price, 1997: 337). The measurement tool may or may not be the reason for this
variables lack of significance in this study; nonetheless, it is recommended that a
more comprehensive instrument be used in future research.
Another explanation for the lack of significance of organizational
commitment in this study is because it reflects the nonprofit sector and commitment
to mission, which is a new competing variable in the analysis. Since commitment to
mission is significant in the study, it may be so important to fundraisers that
organizational commitment is less relevant. The correlation table (see Table 4.4)
indicates that the determinants measured two different things. Yet, it is possible that
commitment to mission used up the variance that would have otherwise (in the for-
80


profit sector) gone to organizational commitment. Further research should inquire
how the determinant commitment to mission affects organizational commitment in
the nonprofit sector.
Analysis of the Hypotheses
Regression analysis was run for each of the five hypotheses, each of which
included only the variables in that hypothesis. Even though the correlation table (see
Table 4.4) and beta weights (see Table 4.5) give insight into answering the five
hypotheses, individual regressions were run in order to see how much variance each
group has on intent to stay and to determine if any variables that were insignificant in
the previous analysis are significant here.
Hi: The group of exogenous variables containing opportunity,
general training, negative affectivity, job stress from insufficient
resources, job stress from conflict, job stress from ambiguity, job
stress from workload, job stress from unrealistic fundraising
goals, and routinization will have a negative effect on intent to
stay.
81


Table 4.6: Results from Multiple Regression Analysis for Hypothesis 1 (Displaying
Standardized Betas)
Beta
StressIR l OO *
StressAmb -.230 **
Routinization -.190 **
F value 11.235 ***
N 169
Adjusted R1 .354
a. Dependent Variable: Intent
b. *** p< .0001; ** p<.05; *p<.10
Table 4.6 shows that Hypothesis 1 is partially supported. This group of
variables explains 35 percent of the variance in intent to stay at a significance level of
<.0001. However, only stress from inadequate resources, stress from ambiguity, and
routinization are statistically significant and negatively correlated with intent to stay.
None of these were significant in the backward elimination analysis of the 26
independent variables.
The survey instrument measured stress from inadequate resources via two
items: (1) I have adequate equipment/technology to do my job, and (2) I have enough
support services to do my job. It is possible that stress from inadequate resources was
significant due to the fact that this was tested in the nonprofit sector, which may be
less likely to have the financial resources for the latest and greatest technology, or
82


needed support such as administrative assistance. It would have been interesting to
measure the budget size of the organizations to see if that affected the way
respondents viewed their access to adequate resources.
As in any profession, nonprofit or for-profit, if employees do not have the
resources they need to do their job, the resulting stress is negatively correlated with
their intent to stay. For fundraisers, technology is very important for tracking donors,
writing grants, and having access to e-mail and the Internet. In addition, if fundraisers
are spending all of their time doing administrative work, they will not have time to
meet with donors. The significance of this variable may be indicating that fundraisers
do not want to fail, and if they feel that their organization is not giving them the
resources they need to be successful fundraisers, they will go somewhere else.
The next significant variable in Hypothesis 1 is stress from ambiguity. In the
Handbook of Organizational Management, job ambiguity is described as having three
dimensions: performance criteria, work method, and scheduling. Performance criteria
refers to the standards used in evaluating employees performance. Work method is
the actual procedures employees use to accomplish their work. Scheduling is defined
as the sequencing of activities for employees to do their job (Price, 1997).
The survey used the following two items to measure this variable: (1) I know
exactly what is expected of me in my job, and (2) I have to work under vague
directives. These statements do not address the three types of stress from ambiguity,
83


as they are more general. However, they do address the main issues of ambiguity
unclear expectations and vague directives.
Stress from ambiguity also seems to be a typical determinant of intent to stay
for all professions. But for fundraisers, being in an art profession, it may be even
more important. Focusing on performance criteria and work methods, it is important
for fundraisers to know how much money they need to raise, what the money is for,
who they should ask for the funding, and how they should go about raising those
funds. The fundraising profession has not produced extensive guidelines on how to
measure fundraisers performance, which may be a challenge for some organizations.
Quite often, organizations evaluate fundraisers on how much money they raise.
Others evaluate fundraisers on their activity (i.e., grants written, prospects researched,
donors contacted, etc.). Some look at both the funds raised and the activity level of
the fundraiser. Regardless of how a fundraiser is evaluated, it is important for the
expectations to be clarified beforehand. It is understandable that if fundraisers are
unclear about what is expected of them, which will lead to stress, and therefore, a
desire to leave the organization. This is similar to stress from inadequate resources, in
that a fundraiser might perceive this type of negative situation as being a set up for
failure.
The other significant variable in Hypothesis 1 is routinization. Routinization is
the degree to which a job is repetitive (Price, 1997: 521). It is typically more
relevant in clerical and laboring jobs than in professional jobs (Price, 1997). Since
84


Full Text
Peters, J., Femandopulle, A., Masaoka, J., Chan, C., & Wolfred, T. (2002). Help
wanted: Turnover and vacancy in nonprofits (Study). San Francisco:
CompassPoint Nonprofit Services.
Peters, J., & Wolfred, T. (2001). Daring to lead: Nonprofit executive directors and
their work experience. San Francisco: CompassPoint Nonprofit Services.
Pollack, R. H. (2003). Beyond philanthropy. CASE Currents, 20-25.
Price, J. L. (1990). Economic, psychological, and sociological determinants of
voluntary turnover. Journal of Behavioral Economics, 19(3), 321-336.
Price, J. L. (1995). A role for demographic variables in the study of absenteeism and
turnover. The International Journal of Career Management, 7(5), 26-32.
Price, J. L. (1997). Handbook of organizational measurement. International Journal
of Manpower, 75(4/5/6), 303-558.
Price, J. L. (2001). Reflections on the determinants of voluntary turnover.
International Journal of Manpower, 22(7), 600-624.
Price, J. L. (2004). The development of a causal model of voluntary turnover. In R.
Griffeth & P. Horn (Eds.), Innovative theory and empirical research on
employee turnover (pp. 3-32). Greenwich, CT: Information Age Publishing.
Price, J. L., & Kim, S.-W. (1993). The relationship between demographic variables
and intent to stay in the military: Medical personnel in a U.S. Air Force
hospital. Armed Forces & Society, 29(1), 125-144.
Sager, J., Griffeth, R. W., & Horn, P. (1998). A comparison of structural models
representing turnover cognitions. Journal of Vocational Behavior, 53, 254-
273.
Schwinn, E., & Sommerfield, M. (2002, April 18). Revolving door dilemma The
Chronicle of Philanthropy, pp. 39-42.
Schwinn, E. (2002, July 25). Keeping donors in the fold. The Chronicle of
Philanthropy, 14, pNA.
Sommerfeld, M. (2002a, February 21). Keeping up after the campaign. The Chronicle
of Philanthropy, 14, pNA.
166


Sommerfeld, M. (2002b, December 12). Self-employed consultants: What works and
what doesn't The Chronicle of Philanthropy, 15, pNA.
Steel, R. P. (2002). Turnover theory at the empirical interface: Problems of fit and
function. Academy of Management Review, 27(3), 346-360.
Steers, R. M., & Mowday, R. T. (1981). Employee turnover and post-decision
accommodation processes. In L. L. Cummings & B. M. Staw (Eds.), Research
in organizational behavior (Vol. 3, pp. 235-281). Greenwich, CT: JAI Press.
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.).
Boston: Allyn and Bacon.
Taketa, K. (2002). Nonprofit leadership needs support to reduce high rate of burnout
and turnover. Retrieved September 17, 2004, from
http://the.honoluluadvertiser.com/article/2002/Oct/24/bz/bzl 6a.html/?print=o
n.
Thatcher, J. B., Stepina, L. P., & Boyle, R. J. (2003). Turnover of information
technology Workers: Examining empirically the influence of attitudes, job
characteristics, and external markets. Journal of Management Information
Systems, 79(3), 231-261.
The Center on Philanthropy at IU. (2003). Giving USA 2003 (Annual Report):
AAFRC Trust for Philanthropy.
Wagner, L. (2002). Careers in fundraising. New York: John Wiley & Sons, Inc.
167