Citation
Explanatory factors for local charter school policy implementation

Material Information

Title:
Explanatory factors for local charter school policy implementation the case of Colorado's school districts
Creator:
Lee, Jeongho (John) ( author )
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
neg
Physical Description:
1 electronic file (241 pages). : ;

Subjects

Subjects / Keywords:
Charter schools -- Colorado ( lcsh )
Charter schools -- Government policy -- Colorado ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )
Target Audience:
The main purpose of this dissertation is to empirically examine under what conditions Colorado's school districts more actively implement the state charter school law or why some of Colorado's school districts more actively deliver charter school services to their residents compared to other school districts. TO answer this research question, this dissertation targets Colorado's school districts and constructs the two equation models including 13 hypotheses, which are created by support of the four theoretical approaches, - policy diffusion, policy entrepreneur, policy interest group, and policy network models - and school district characteristics. The dependent variable in each equation model is operationalized by the different measurements while each equation model embraces the same 13 independent variables. To estimate both dependent variable and independent variable in these two equation models, the author makes his own unique cross-sectional dataset, which was completed by conducting a survey and collecting secondary data. Meanwhile this dataset is analyzed by the multiple ordinary least squares (OLS) regression analysis among several multivariate statistical analysis techniques. The analyzed results of the both equation models commonly reveal that the mimetic diffusion, policy entrepreneur, and policy network factors play significant roles in answering the research question. Based on the three statistically significant explanatory factors, this dissertation concludes that as existence of the three explanatory factors - neighboring school districts with previous experience for charter schools, policy entrepreneurs, and policy networks - increases, Colorado's school districts are more likely to implement the state charter school law. Regarding future studies, first, this dissertation suggests that magnet schools could be a more proper measurement to operationalize the market-based education tool (MBET) variable to improve this study's validity. Second, the dissertation addresses that a political characteristic factor will be a potential explanatory variable that can provide this dissertation's research questions with a good answer. Finally, the dissertation proposes that using a qualitative method could be a good way to obtain more accurate and in-depth results.

Notes

Thesis:
Thesis (M.S.)--University of Colorado Denver. Public affairs
Bibliography:
Includes bibliographic references.
System Details:
System requirements: Adobe Reader.
General Note:
School of Public Affairs
Statement of Responsibility:
by Jeongho (John) Lee.

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:
903048318 ( OCLC )
ocn903048318

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

EXPLANATORY FACTORS FOR LOCAL CHARTER SC HOOL POLICY IMPLIMENTATION: THE CHOOL DISTRICTS by JEONGHO ( JOHN ) L EE B.A., Kangwon National University 1997 M.A. Kangwon National University 2000 M P A University of Colorado Denver, 2002 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Public Affairs 2014

PAGE 2

This thesis for the Doctor of Philosophy degree by Jeongho (John) Lee has been approved for the Public Affairs Program by Paul E. Teske Dissertation Chair Peter deLeon Examination Chair Kelly J. Hupfeld Donald E. Klingner M. Jae Moon July 1 8 2014

PAGE 3

iii Lee, Jeongho (John) (Ph.D., Public Affairs) Explanatory Factors for Local Charter School Policy Implementation: The Case of Thesis directed by Distinguished Professor Paul E. Teske ABSTRACT The main purpose of this dissertation is to empirically examine under what chool districts more actively deliver charter school services to their residents compared to other school districts. To answer this research equation models including 13 hypotheses, which are created by support of the four theoretical approaches policy diffusion, policy entrepreneur, policy interest group, and policy network models and school district characteristics. The dependent variable in each equation model is oper ationalized by the different measurements while each equation model embraces the same 13 independent variables. To estimate both dependent variables and independent variables in these two equation models the author makes his own unique cross sectional dat aset, which was completed by conducting a survey and collecting secondary data. Meanwhile, this dataset is analyzed by the multiple ordinary least squares (OLS) regression analysis among several multivariate statistical analysis techniques. The analyzed results of the both equation models commonly reveal that the mimetic diffusion, policy entrepreneur, and policy network factors play significant roles in answering the research question. Based on the three statistically significant explanatory factors, th e dissertation concludes that as existence of the three explanatory factors

PAGE 4

iv neighboring school districts with previous experience for charter schools, policy entrepreneurs, and policy networks likely to imple ment the state charter school law. Regarding future studies, first, this dissertation suggests that magnet schools could be a more proper measurement to operationalize the market based education tool Secon d, the dissertation addresses that a political characteristic factor will be a potential explanatory variable that can provide ch question with a good answer. Finally, the dissertation proposes that using a qualitative method could be a good way to obtain more accurate and in depth results. The form and content of this abstract are approved. I recommend its publication. Approved: Paul E. Teske

PAGE 5

v DEDICATION I dedicate this dissertation to my late parents, Gwang c heol Lee and Jeong s oo Seo. Hi Dad and Mom, I am here thanks to your true and unlimited love. I understood Agape Love through you. I miss you, respect you, and love you forever. See you in Heaven. F rom Jeongho, Your youngest son

PAGE 6

vi ACKNOWLEDGMENTS I would like to give special thanks to my sisters (Og nyeo Lee, Ogh ee Lee, and Mi h ee Lee) brother (Jeong h eon Lee), sister in law (Hye r an Hwang) and brother in law (Sang g yu Lee) They have been rea lly dedicated in supporting me. Their prayers have helped me overcome a lot of barriers that made me desperate during my doctoral program. I know that they have felt my happiness and sadness with me and I thank them for their dedication. I wo uld like to thank my nephew (Hojin Choi) and nieces (Hagyeong Lee, Jiho Lee and Jiy oon Lee). All of their calls cards and letters have brought me great happiness. I want to thank them as much as I can. And, I hope that I can have more time with them soon. My doct oral study period has been really long. If I said that it was just easy, I would be a liar. It was very tough. Sometimes, I regreted my decision of starting this doctoral program. However, my mentors have helped me not to lose my motivation and pace. Dr. P aul Teske, my dissertation chair, introduced the new research topic school choice movement to me. Before I met him, my interest was environmental policy. The research topic of privat izing public education was totally new to me. However, I was able to have the charter school topic as my main research interest due to his invaluable support. I sincerely thank him for his intellectural help and guidelines. Dr. Peter deLeon showed me who a professor having reason and emotion together is. His critical comments ar e shown throughout my dissertation and his attentative heart remains in my mind. When I see him he reminds me of my Dad. I would like to thank Dr. Kelly Hupfeld and her smile. She is really busy due to her work load. However, whenever I met her, she gave

PAGE 7

vii school specialist, she also provided me with lots of academic insights helpful to the dissertation. Dr. Donald Klinger was the professor who taught my first class in the docto ral program. I cannot forget how powerful his encouragement was to me. At that time, my knowledge was not proficient enough to follow a doctoral course. Dr. Klingner led me to succeed in his class. Since our first meeting, he has continued to strongly supp ort me until now. I always thank him for the consideration he has shown me. I met Dr. M. Jae Moon in 2000 at this school. He and his wife helped me accept Jesus in 2002. Since then, I have been peaceful and powerful through many huge obstacles. I also owe him for his academic advice and help. Hopefully, I can repay him for them soon. There are more people to whom I want to convey my thanks. I must express my thanks to Dr. Linda deLeon my friend and old chair. During my entire doctoral program period, she was the kindest and the most supportive professor who gave me much happiness and joy. Magically, it took only 10 minuntes for me to throw away my gloomy feeling and loser attitidue whe n I talked with her I want to emulate her skills in being a n excellent professor. T hey will be the most powerful tools for my future At this school, t here is a faculty member I like, who starts each day as an early bird. She is Dr. Christine Martell. She came to this school as a faculty member in 2000 when I just began my Ameircan life and MPA degree. I am thankful for her help during the long period. I must not forget to express my thanks to Dr. Tanya Heikkila, the doctoral program director. She showed m However, I felt that her leadership is based on unders tanding others not authority. M y doctoral cohorts say that many conversation s with her always encourage them. I agree with them. I am thankful for her consideration to me as well. Dr. Mary Guy is a faculty

PAGE 8

viii member who knows about my academic failures and personal hardships very well. I know that she feels relieved and is happy when she see s me finishing my l ong doctoral study journey. I hope that I can show her my better and successful performance soon and we can delightfully talk about the past things in the future. I de eply thank her for helping me find what I should do in the tough situations. I also thank Antoinette San doval and Dawn Savage for their help. They have always helped me with their cordial and steady attitude s so that I could easily complete all of the formal document processes It is time to finish my long doctoral program. Now, I can say t hat I am a very happy and lucky student because I met these excellent people Based on my capabilities, my achievements during my doctoral process would not be possible without my mentors and their steady support. I live in each day, doing my best. However, my efforts are just tiny parts of my impressive performance. Completing my d octoral degree, I know that H is plan that has led me to meet such great people during my doctoral study period. I am also very glad because God will be always with me as He has always been with me. Thank y ou for your faith in me!

PAGE 9

ix TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ................................ ....... 1 Background ................................ ................................ ................................ ........... 1 American Education Risk ................................ ................................ ...................... 2 Emergence of Education Reform ................................ ................................ ... 3 Market based vs. Command controlled Education System ........................... 6 School C hoice Movement ................................ ................................ ..................... 7 Principles of Schoo l Choice Movement ................................ ......................... 8 Charter Schools ................................ ................................ ........................... 1 0 Reinventing Government Approach ................................ ................................ ... 1 2 Reinventing American Government ................................ ............................ 13 .......................... 15 Research Topic and Scope ................................ ................................ .................. 1 8 Glimpse of Theoretical Approaches ................................ ........................... 19 Key Definitions ................................ ................................ ............................ 1 9 Units o f Analysis ................................ ................................ .......................... 22 Explanation of Theoretical Approaches ................................ ....................... 22 Dissertation Chapters ................................ ................................ .......................... 2 5 I I LITERATURE REVIEW ................................ ................................ .............................. 26 Introduction ................................ ................................ ................................ ......... 26 ................................ ................................ .............. 28 Movement ................................ ................................ .. 32 Principles of School Choice Movement ................................ ...................... 32 School Choice Movement Tools ................................ ................................ .. 33

PAGE 10

x Schools ................................ ................................ ................. 3 4 Definitions of Charter Schools ................................ ................................ ..... 3 5 ................................ ......................... 37 Spread of Charter School Laws ................................ ................................ ... 41 ................................ ......................... 47 ................................ ................................ ................ 52 ................................ ........................ 53 ................................ ............ 5 9 ................................ ............... 6 2 Research Questions ................................ ................................ ............................. 6 5 I II THEORETICAL APPROACHES ................................ ................................ ................. 6 7 Introduction ................................ ................................ ................................ ......... 6 7 Factors Affecting Innovation Adoption and Implementation ............................. 7 1 Four Theoretical Models ................................ ................................ ..................... 7 1 Policy Diffusion Model ................................ ................................ ................ 7 5 Policy Entrepreneur Model ................................ ................................ .......... 80 Policy Interest Group Model ................................ ................................ ........ 8 7 Policy Network Model ................................ ................................ ................. 9 2 School District C haracteristics ................................ ................................ .......... 100 Demographic Factors ................................ ................................ ................. 1 0 1 School related Factors ................................ ................................ ............... 1 0 5 Socioeconomic Factors ................................ ................................ .............. 1 10 I V RESEARCH METHODOLOGY ................................ ................................ ................. 11 2 Overview ................................ ................................ ................................ ........... 11 2 Hypotheses ................................ ................................ ................................ ........ 11 6

PAGE 11

xi Data Collection ................................ ................................ ................................ 1 20 Survey ................................ ................................ ................................ ........ 1 20 Secondary Data ................................ ................................ .......................... 1 2 5 Data Analysis ................................ ................................ ................................ .... 1 2 7 UCINET 6 ................................ ................................ ................................ .. 1 2 7 Multivariate Analysis ................................ ................................ ................. 1 30 Two Equation Models ................................ ................................ ................ 1 30 V FINDINGS AND RESULTS ................................ ................................ ...................... 1 3 3 Descriptive Statistics ................................ ................................ ......................... 13 3 Multicollinearity ................................ ................................ ............................... 1 40 Correlation Coefficients ................................ ................................ ............. 14 2 Variance Inflation Factor (VIF) and Tolerance Level ............................... 14 3 Multiple OLS Regression Analysis ................................ ................................ .. 1 4 6 ................................ ................................ ..... 14 8 Model B s ................................ ................................ ..... 15 4 VI CONCLUSIONS AND FUTURE STUDIES ................................ ............................... 1 6 1 Conclusions ................................ ................................ ................................ ....... 1 6 1 Findings ................................ ................................ ................................ ............. 16 4 Policy Diffusion ................................ ................................ ......................... 16 4 Policy Entrepreneurs ................................ ................................ .................. 1 6 5 Policy Networks ................................ ................................ ......................... 1 6 6 Implications ................................ ................................ ................................ ....... 1 6 9 Future Studies ................................ ................................ ................................ ... 1 7 3

PAGE 12

xii REFERENCES ................................ ................................ ................................ .............. 1 7 9 APPENDI X A Approval Letter ................................ ................................ ........ 201 B Survey I to Superintendents ................................ ................................ ......... 20 2 C Survey II to Organizations ................................ ................................ ........... 2 1 5

PAGE 13

xiii LIST OF TABLES Table I I.1 Differences between Public Schools and Private Schools ................................ ........ 3 1 I I .2 ................................ ............... 39 II. 3 Number of Charter Schools in Eac h State in the 2011 2012 School Year ............... 4 0 II. 4 Cumulative Proportion of Jurisdiction Charter School Legislation by Year ............ 4 5 I 2011 2012 School Year) ................................ ................................ ................................ ................................ 5 7 I I.6 Location of Charter Schools in Colorado (2011 2012 School Year) ........................ 5 8 I I.7 Charter Schools Size in Colorado (2011 2012 School Year) ................................ ... 5 8 I I.8 Comparison of Races between Charter Schools and Conventional Public Schools in Colorado (2011 2012 School Year) ................................ ................................ ................. 6 1 I I.9 Comparison of Advanced or Proficient Students in Charter Schools a nd Conventional Public Schools (2011 2012 School Year) ................................ ................................ ........ 6 4 III.1 Theoretical Contributions of Multiple Academic Disciplines in Policy Implementation Analysis ................................ ................................ ................................ 70 III.2 Categories of Potentia l .... 7 4 I V.1 Predicted Directions of All Regressors on the Dependent Variable ...................... 1 1 9 I V.2 Comparison between Responding School Districts and Non responding School 124 I V.3 Measurement of All Independent Variables ................................ .......................... 1 2 6 V.1 Descriptive Results of All Variables ................................ ................................ ....... 1 3 9 V.2 Analyzed Results for Multicollinearity of All Independent Variables ................... 1 4 5 Multiple OLS Regression Analysis Results ..................... 15 3 V.4 Summary Multiple OLS Regression Analysis Results ..................... 1 60

PAGE 14

xiv LIST OF FIGURES Figure I I .1 2011) ................................ ................................ ................................ ................................ 46 I I .2 Cumulative Number of Charter Schools (1992 1993 Schoo l Year to 2011 2012 School Year) ................................ ................................ ................................ .................... 5 0 II. 3 Growth of Students in Charter Schools (1997 1998 School Year to 2011 2012 School Year) ................................ ................................ ................................ ................................ 51 II. 4 1994 School Year to 2011 2012 School Year) ................................ ................................ ................................ ................................ 5 6 II. 5 1994 School Year to 2011 2012 School Year) ................................ ................................ ..................... 60 I V 1 A Sample of the Matrix for the Strongest Density of Policy Networks ................ 1 2 9 I V 2 A Sample of the Matrix for the Weakest Density of Policy Networks ................. 1 2 9

PAGE 15

xv LIST OF EQUATIONS Equation I V .1 ................................ ................................ ............................... 1 3 1 I V.2 ................................ ................................ ............................... 1 3 1 V.1 ................................ ................................ ................................ ........ 1 4 3 V.2 ................................ ................................ .......... 1 4 6

PAGE 16

xvi LIST OF ABBREVIATIONS AFT BEST CCCM CDE CER CLCS COEA COMIRB CSAP CSI ELCO NCLB NEA PISA SAT SCM SSP TCAP TPS American Federation of Teachers BEST Board Colorado Department of Education Center for Education Reform Colorado League of Charter Schools Colorado Education Association Colorado Multiple Institutional Review Board Colorado Student Assessment Program Colorado Charter School Institute Education Leadership Council No Child Left Behind National Education Association Program for International Student Assessment Scholastic Assessment Test School Choice Movement Sustainable Slopes Program Transitional Colorado Assessment Program Traditional Public School

PAGE 17

1 CHAPTER I INTRODUCTION Background Scholars in the public administration and policy area have frequently sought specific causes for mechanisms of the state charter school policy adoption. However, they have rarely found satisfactory answers explaining mechanisms of the local charter school policy imple mentation as public education service delivery. To quench this intellectual thirst, this dissertation has designed and led each chapter to answer why local governments school districts vary in implementation of their own state charter school law This acad emic challenge shows that the primary intent of the dissertation is to examin e what factors cause or drive the uneven charter school policy implementation he dissertation s research topi c is represent ed by a research question: W hy and under what conditions some school districts in Colorado very actively implement the s tate charter school law, why other school districts implement it less actively, and why other school districts do not impl ement it at all ? 1 To empirically study this research question th is dissertation created and tested the hypotheses that are based on four main theoretical approaches policy diffusion policy 1 policy districts implement the state be defined as a synonym. Moreover, using the two terms interchangeably shows that this dissertation pursues the degree of public affairs, which covers both public administration and public policy academic areas. The dependent variable of this study variation of the local charter school implementation or variation of the local charter school service delivery, is expressed by a degree how actively each school district carries out the state charter school law. To measure this dependent variable, this study utilizes the percentage of charter schools among K 12 public schools in each school district and the raw number of charter schools in each school district.

PAGE 18

2 entrepreneur, policy interest group, and policy network model s and other school district characteristics. Scholars in the public administration and policy area have long been exploring this kind of research topic in multiple research areas. However, studies examining mechanisms of the local charter school policy impl ementation have been very infrequent and rare (Lee & Jeong, 2012; Lee & Kim, 2010) Colorado especially, remains an unexplored and unknown research territory in this study topic up until now This academic fact implies that th is dissertation plays a role as an academic trigg e r and as a r e search pathfinder shepherding scholars to explore the mechanisms of Colorado s local charter school policy implementation. Th is dissertation also expects to contribute the dissertation to help all main actors education le aders, practitioners, and scholars in the education policy area understand why local governments deliver more innovative education services to their r esidents. Namely, t o accompli s h this academic goal and produce an academic contribution, the dissertation must seek satisfactory causes and precise answers concretely accounting for why the uneven charter school policy implementation happens or exists American Education Risk Almost nobody doubts that the United Stat es of America is the most powerful country in the world in terms of economic and military might and capabilities (Walberg, 2007). But, its educational performance is ironically not outstanding, compared to other developed countries. S everal scholars (Chubb & Moe, 1990; Cohn, 1997 ; Lee & Jeong 201 2 ; Schneider Teske, & Marschall, 2000 ) argue that in the American traditional public school ( TPS ) system the past and current performance of conventional K 12 public

PAGE 19

3 schools has not been strong and the public have generally been dissatisfied with conventional K 12 public school education services Minorities, especially, want to receive other education services rather than the conventional K 12 public school education (Good & Braden, 2000; Lee & Kim, 2010) T h ere are several r eports and studies describing the low performance of American education A Nation at Risk, which was published in 1983, although not without its own problems, emphasizes that there are serious problems in American K 12 education and its gl obal competitiveness is low (McBeath Reyes, & Ehrlander 2008). 2 Walberg (2007) subjects. Compared to students in other countries, older American students do poorly in pointing out that the USA got low ranks in math and science (35th in math and 29th in science) among 57 industrialized countries in the 2006 Program for International S tudent Assessment (PISA) Mentioning the most recent result of PISA, Huffington P ost (2012) declares that the American student achievement is almost s imilar with the result from 6 year s ago throug hou t the 2012 PISA s evaluation. The 2012 PISA ranks America 25 th in math, 14 th in reading, and 17 th in science after it analyzes and evaluate s achievements in 40 developed countries. Emergence of Education Reform The low performance of the American K 12 education has been a big issue and worry to each new American presidential administration that wants the US A to realize the optimal position and become a leading country in the world. Particularly the 2 The National Commission on Excellence in Education published this report (Stedman, 1994)

PAGE 20

4 emergence of neoliberalism, which is representative of T h a t cherism s philosophy in England playe d a role as an ignition that stimulated American leaders to purs u e making a strong America at the end of the 1970s and the beginning of the 1980s (Finn, 2013). Amon g several political leaders, former President Ronald Reagan was the most representative leader in actually conducting this goal. He wanted his administration to be a government that strived to solve American education inefficiency and its low performan c e. Highlighting both deregulatio n of several policy processes and the rights of individuals, the Re a gan administration devoted itself to a change of the American TPS system pursuing egalitarian objectives and teaching American citizens social rule s. Following this political direction, th e Reagan administration considered and created educational policies stressing individual liberties and ingenuities in competitive circumstances (Urban & Wagoner, 2009). successors George H. W. Bush, Bill Clinton, and George W. Bush have shaped education policies focusing on both creativities and rights of individuals and better performance In reforming the American TPS system, both major parties Democratic Party and Republican Party have strangely been stressing and pursuing a similar education policy vision since the Re a accepted the main principles of neoliberalism and education policies based on Reaganomics were introduced ( Myers & Cibulka, 2008) For instance, a ne w American education aura of the school choice movement has strongly been supported by both political parties. That is, the primary directions of American education policies head to ward less money and higher goal

PAGE 21

5 accomplishments in the competitive educatio nal environments whose final destination aims at improved performance by American students. Currently, the administration of President Barack Obama has sided with the and direct ion of the American education system (Frumkin, Manno, & Edgington, 2011) The Obama administration has been striving to innovate the TPS system in order to improve poor education performance, arguing that conventional K 12 public school achievements have b een getting worse compared to Asian countries (Condon, 2010) President challenge to reform the TPS system anticipates introducing and fostering innovative policies and programs in the TPS system to a greater extent than the previous George W. Bush administration. Si nce President Obama was sworn in the 44 th American President in 2009, he has emphasized that the American education and health care systems must be changed to make the USA better and provide all Americans with high qualified public servi ces. To accomplish reform of both areas, the American government has used two different approaches market based competitions and command controlled regulations for each major reform In general, Steward, Hedge, and Lester (2008) distinguish characteristics of public policies with the terms liberal and conservative. They indicate that liberal public polic i es agree with government interventions while conservative public policies oppose g overnment interventions. 3 Thus, the current American education reform using the concept of m arket based competition resembles the conservative policy style I n contrast 3 Steward and his colleagues (2008) have adopted the words liberal and conservative policies from an in 1964.

PAGE 22

6 the present American health care system reform conducted by the comman d controlled reg ulation is close to the liberal policy style Market based vs. Command controlled Education System The market based challenges in the education reform had been supported by the late Milton Friedman a pro market economist who introduced the concept of school choice focusing on a voucher program (Abernathy, 2005 ; Anrigh, 2008; Gerson, 2013; Hess, 2010 ; Spring, 2005 ). main perspective s were direct control or involvement would produce negative results in the education system 2) the inefficient TPS system needs to take a market oriented solution to overcome its inefficiency, and 3) school competition which is the fundamental concept in the market oriented approach would facilitate excellence in the achievements of the whole T PS system (Friedman, 1962, 2002) Traditionally, the American education system had been developed as a tool to not improve individual creativity but to comply with social ordinance and responsibility (Urban & Wagoner, 2009). That is, education has been us ed as a means to make each American well behaved citizen who follows and obeys the social order To accomplish this goal, American leaders have been using a one size fits all education service form, which is harmonized with both Jeffersonism and Jacksonism that highlight citizens equality and social stability ( Nester, 2013; Urban & Wagoner, 2009) Conventional public schools that are created with this educational approach and view have been increased up until now. Peterson and Campbell (2001) declare that the twentieth century, the design of the American school system became increasingly comprehensive, uniform, centralized, and professionally directed. (p. 2).

PAGE 23

7 They found that 92% of children had enrolled in public schools by 2000 while about 7 0% of all children between 7 years old and 17 years old studied in public schools 100 years ago. Their explanation shows that American governments ha ve gradually and steadily been emphasizing and raising the command controlled TPS system by supporting conv entional public schools However public school achievements in the TPS system American citizens to follow social orders. Market oriented scholars have introduced a new market based approach to education because they believe that the noncompet itive mentality and circumstance, which are strongly embedded in the TPS system, ha ve caused low performance of the conventional public schools (Chubb & Moe, 1990; Schneider et al., 2000 ) School Choice Movement In the market based education system, students and their parents are denominated as customers (demanders) while public schools are described as companies (suppliers) that compete with other companies to survive (Abernathy, 2005; Schneider et al., 2000 ). Introducing the school choice movement based on the concept of both educational de mander and supplier, Friedman ( 1962, 2002) assert s that there are two positive views supporting the school choice movement. Fir s t, the school choice movement allows parents to have greater chances in selecting schools for their children. Second, the school choice movement stimulates conventional public schools to have better capacities because conventional public schools would drop out of the market system by losing their customers in the case that they do not show desirable and successful outcomes to educa tion customers. approach and challenge was strongly supported by

PAGE 24

8 the Ronald Reagan a dministration that pursued the goal smaller governments, better public services through the emphasis of the free market mechanism in the early 1980s. Pro sc was been originated from public choice theory ( Buckley & Schneider, 2007; Chubb & Moe, 1990; Gauri, 1998; Lee & Jeong, 2012) 4 Along with the core perspective of public choice theory, some school choice movement scholars (Bryk & Gomez, 2008; Gergen & Vanourek, 2008) indicate that command and control approach in the TPS system has shown its limitations in current educational environments. Pro school choice scholars (Chubb & Moe, 1990; Schneider et al., 2000) worry that individual liberty and ingenuity have been decreased under the TPS system which is centered on the command control governance approach Principles of School Choice Movement conven tional public schools in the TPS system which is used to providing conventional public schools with centralized and standardized guidelines controlled by powerful government entities, h ave lost the motivation to improv e their own achievement s and efficien cy because command controlled circumstances employed in the TPS system do not allow conventi on al public schools to possess their own creative curricula (Buckley & Schneider, 2007; Chubb & Moe, 1990; Schneider et al., 2000). In general, the public have been skeptical of traditional public education services created by centralized government s ( Lee & Jeong, 2012; Miron & Nelson, 2002). 4 In 1965, James Buchanan and Gordo n Tullock established the public choice communities, which are embedded in the rational choice model and self interested actors (Hill, 1999; Lemieux, 2004). Public nomics, political

PAGE 25

9 To overcome the shortcomings of the TPS system that is used in the cookie cutter approach and monopolistic characteristics school choice adherents propose that policy decision makers need to consider applying market oriented concepts to the TPS system (Buckley & Schneider, 2007; Chubb & Moe, 1990; Schneider et al., 2000). Pro school choice scholars emphasize the concept of c ompetition which is one of positive market forces (Fusarelli, 2003) They declare that American public education has been monopol istic and its achievements are not impressive And, they highlight that competition one of m arket forces is necessary for be tter education services and will be a tool to facilitate the change of the monopolistic characteristics of the TPS system that has been created by centralized government guidelines (Minron & Nelson, 2002) These school choice movement advocates point out that the competition among public schools plays a positive role in effective school achievement because public schools strive to satisfy their customers by offering new and innovative programs or facilities, which provide their customers with a better educ ational environment, in order to assure their survival (Howell & Peterson, 2006). T his interpretation is advocated by Abernathy (2005) under competition from charter schools and, perhaps, private schools will pay more attention to their customers 2). T hat is, those that cannot ( Chubb & Moe, 1990; Schneider et al., 2000; Smith & Meier, 1995). That implies that competitive circumstances in the TPS system let K 12 public schools

PAGE 26

10 provide students and parents with better qualified educational services. Finally, i ts positive influence helps improve the TPS system. 5 Based on this rationale several alternative education policies and programs appeared to facilitate competitive circumstances in the American TPS system. The logic of their appearanc e is commensurate logic: The competition in the TPS syste m makes the qualification of the public education service better and provides education customers with diverse improved education services (Friedman, 1962, 2002) A dvocates of the school choice movement propose and support marke t based concepts by introducing innovative policies and programs such as charter schools, magnet schools, vouchers, tax credits, and the open enrollment law to make the TPS system more competitive (Howell & Peterson, 2006 ) Proponents propose that the s t y le of the TPS system s governance would be more decentralized and varied and both qualification of school services and school performance would be improved if these policies or programs supporting the school choice movement bec a me prevalent in the TPS system (Chubb & Moe, 1990; Schneider et al., 2000). Charter Schools Among the aforementioned innovative institution s that school choice supporters propose, th e dissertation focuses on charter school s T he first emergence of charter 5 proponents of conventional public schools emphasize that there are several benefits th at all conventional public schools commonly possess. Their representative benefits are: 1) students and their parents are sure that teachers working in conventional public schools are certified and have passed qualifying conditions that governments require 2) students in conventional public schools have more chances to learn socialization and social interaction because conventional public schools have more diverse students with a wide variety of backgrounds compared to other educational institutions online schools, private schools, and homeschools, and 3) conventional public schools provide students with more equal educational services counseling and speech therapy or equal extra curricular activities art, band, or theater programs because they receive fina ncial support (funding) from governments compared to private schools, which do not receive funding from governments (Alleyne, 2014; Huson, 2014; Lawrence, 2014).

PAGE 27

11 schools which are established with the positive concept of the school choice movement was approximately 20 years ago (Hill & Lake, 2010; Smith Wohlstetter, Farrel, & Nayfack, 2011). Education leaders such as Ray Budd e and Albert Shanker introduced charter schools as facilitators that prom pt competition of public schools the fundamental spirit of free market economy in the TPS system (Lee & Kim, 2010; Vergari, 2002) Th e two education leaders efforts helped charter schools to be considered as innovative educational policy instruments that change the monopolistic circumstances in the TPS system and that address its weaknesses. Minnesota was the first state that adopted market based education strategies developed by scholars. It created its first charter school act in 1991 (Buckley & Schneider, 2007; Mintrom & Vergari, 1996, 1997, 1998; Renzulli & Roscigno, 2007; Schneider et al., 2000; Stoddard & Corcoran, 2007; Vergari, 1999, 2007; Wong & Langevin, 2005, 2007; Wong & Schen, 2002). The two reports published by the Center for Educatio n Reform (CER) in 2012 and 2013 show that up until the end of 2011, a total of 42 jurisdictions out of 51 jurisdictions (50 states plus Washington, D.C.) had enacted charter school laws. b ) indicates that among them, Maine was the mos t recent state adopting its own charter school law in 2011. The other report (2012) describes that Washington D.C. possesses the strongest charter school law while Mississippi ha s the weakest charter school law. 6 6 e evaluation criteria, which are composed of 1) authorizers (15 points), 2) cap (10 points), 3) operations (15 points), 4) equity (15 points), and 5) implementation (no score caps). Thus, the maximum total score starts at 55 points if a state receives the maximum score for each criterion excluding the implementation category. In 2012, Washington D.C. was scored by 46 points and Mississippi was evaluated by 1 point.

PAGE 28

12 Reinventing Government Approach Approxim ately two decades ago, the reinventing government movement hit the USA to change and reform the top down, command controlled, and centralized governance style. 7 The phrase reinventing government became more popular in the USA after it was introduced in O (1992) book and was emphasized Lenkowsky & Perry, 2000; Ruhil, Schneider, Teske, & Ji, 1999; Schachter, 1995 ; Stillman, 1996 ). The main argument of reinventi ng government movement proponents is that the top down and centralized government structure does not fit well into a contemporary world that experiences rapid technological evolution, global market competition, diverse demands of highly educated citizens, and an unhealthy fiscal situation (Osborne & Gaebler, 1992; Osborne & Plastrik, 1997; Zajac, 1997). Thus, they assert that governments need to have the decentralized, market based, and community based governance mind, and public servants have to play roles of catalytic actors that steer rather than bureaucratic actors that row in delivering public services (Frederickson, 1996; Lenkowsky & Perry, 2000; Osborne & Gaebler, 1992; Rusaw, 1997; Vito & Kunselman, 2000). When the term reinventing government was in troduced at the beginning of the 1990s, many scholars and practitioners regarded this reinventing government movement as a paradigm shift because the primary goal of this government reinvention was to create a new government structure by transforming the i nflexible and monopolistic government structure to a market based and competitive government structure shift (Kearney & Hays, 7 In the public administration and policy area, the term reinventing government or governmen t reinvention is interchangeably used with the two terms administrative reform and new public management (Kearney & Hays, 1998).

PAGE 29

13 1998). Some scholars (Denhardt & Denhardt, 2000, 2007; Russell & Waste, 1998) point out that the main principles of the reinventin g government movement are based on public choice theory, which treats a citizen as a self interested customer in the field of market or business. Reinventing government proponents emphasize that in order to satisfy this rational and self interested citizen a government must become a customer driven, mission driven, market oriented, and results oriented organization ( deLeon & Denhardt, 2000; Kamarck, 2004; Osborne, 1993; Osborne & Gaebler, 1992; Ruhil et al., 1999). movement tools can be the best instance of the reinventing gover nment movement because they almost s imultaneously started at the be g inn ing of the 1990s. With this academic insight, this section briefly looks at its contents by exploring the book written by Osborne and Gaebler in 1992 and the report conducted by former Vice Preside nt Al Gore in 1993. T hen, this section also describes the reinventing government contents shown in charter schools. Reinventing American Government reinvention in the 1990s (F rederickson, 1996; Rusaw, 1997). Their book gave both American political leaders and bureaucrats fresh ideas and strong motivation in making the classic monopolistic government more flexible and competitive (Denhardt & Denhardt, 2000, 2007 ; Stillman, 1996 ) The core theme of th eir book proposes that technology evolution, and a competitive global marketplace (Lenkowsky & Perry, 2000,

PAGE 30

14 Spring, 2010). While introducing entrep reneurial spirit th rough their book, they suggest ten reinvention principles and more than 30 strategies to help American governments overcome these new circumstances and accomplish better outcomes (Brudney, Hebert, & Wright, 1999; deLeon & Denhardt, 2000; Denhardt, 2008; Russell & Waste, 1998). 8 Some scholars (Peters & Savoie, 1994; Rivera, Streib, & Willoughby, 2000; Saint approach had a significant impact on the Clinton administration that pursued reform ing federal government and its bureaucrats. In 1993, Vice President Al Gore published a book, which w as designed as a report for President Bill Clinton, more concisely summarizing and reflecting Osborne and 008; Peters & Savoie, 1994 ; Stillman, 1996 ). President Clinton and Vice President Gore are often referred to as atypical traditional Democratic policies such as government reinvention and refor m, which are closer to the Republican political tendency (Lenkowsky & Perry, 2000; Spring, 2005). Many New Democrat policies are located in the middle of the political spectrum Thus, the New Democrats we re called centrist politicians as well. Spring (2005 ) indicates that the two Democratic organizations the Democratic Leadership Council and the Progressive Policy Institute play pivotal roles in supporting New Democrats. The reinventing g overnment trends that the Clinton administration advanced in 1993 c an be identified with t wo important elements: 1 ) a focus on customer and result orient ation and 2) the replacement of highly centralize d and hierarchical government structures with competitive and decentralized management environments. The National 8 Denhardt (2008) arranges the 10 government styles that Osborne and Gaebler propose as the following: 1) catalytic, 2) commun ity owned, 3) competitive, 4) mission driven, 5) results oriented, 6) customer driven, 7) enterprising, 8) anticipatory, 9) decentralized, and 10) market oriented government (p. 138 139).

PAGE 31

15 Perform ance Review (NPR) conducted the se two concepts of government reform to government does what it should do (Kamensky, 1996; Peters & Savoie, 1994 ). Following suggestions, Gore (1993) argue d that government re invention wa s accomplished on the basis of the mind of entrepreneurial government. He state d that the reinventing government movement can be successful by accomplishing the following four principles: 1) cutting red tape, 2) putting customers first, 3) empowering employees to get results, and 4) cutting back to basics (Stillman, 1996) The primary goal of the movement would lead governments and pub lic servant s to work better in a low cost approach (Gore, 1993; Kamensky, 1996). created on the basis of the reinventing government movement citizens as customers and regards public servants as public (bureaucratic or policy) entrepreneurs. Martin, 2001) and target to cut complicated government procedures, save budgets, contract out government functions with private companies or non profit organizations, pursue the market based competition, and emphasize final achievements. 9 The innovative suggestions and strategies introduced in the reinventing government movement affect the American TPS system to be changed and that make its conventional public 9 Taylorianism indicates the scientific management approach, which was a mainstream management style in the USA in the late nineteenth and early twentieth centuries (Kettl, 2002 ; Stillman, 1991 ).

PAGE 32

16 schools more exposed to the market based and competitive circumstances. That is to say, the school choice movement can be regarded as an education movement to reinvent the American TPS system. Two decades ago, Osborne and Gaebler (1992) pointed out that the American TPS system is a typical instance of the rule driven, command controlled, centralized, and bureaucratic model because conventional public schools in the TPS system only operate themselves on the basis of government guidelines and teach students without their unique curricula. Furthermore, in the American TPS system, education customers do not have any rights in choosing schools, and they should enroll in public schools located in their residential districts. These classic pictures shown in the TPS system describe that a change in the TPS system is rarely expected. A Nation at Risk which was written by the National Commission on Excellence in Education in 1983, warned that the TPS system had failed Americans (Osborne & Gaebler, 1992; Osborne & Plastrik, 1997 ; Spring, 2010 ). This report was strongly critical of the American TPS system and predicted that it would be difficult for the USA to keep a dominant position in economy, science, and technology areas if American education capacities and performance are not impr o ved. The advice shown in this report made Americans more interested in the school choice movement. Among 50 states, Minnesota became the most active state that follow ed A Nation at Risk providing their residents with charter school services at the beginning of the 1990s (Osborne & Plas trik, 1997). Some scholars (Carnochan, 2002; Osborne & Gaebler, 1992) indicate that the principles of charter schools include the spirit of the reinventing government movement

PAGE 33

17 moveme nt first of all, a charter school embraces the competitive spirit of the reinventing government movement Charter schools cause schools in the American TPS system to be more competitive. In this competitive circumstance, education customers as demanders c an take or leave a school, choosing a school that fits them well or avoiding a school that does not fit them well. Meanwhile, a school as a supplier can develop its clear missions that help education customers know what the school s ands for, and try to imp rove academics that parents want their children to learn. Thus, a charter school is a good example that shows the competitive spirit of the reinventing government movement Second, a charter school with autonomy is an excellent instance reflecting the dece ntralized spirit of the reinventing government movement Osborne and Gaebler (1992) and Osborne (1993) emphasize that in the centralized governance system, bureaucrats have strong power and enough information to control many ways while clients are really d ependent on them. Osborne and Gaebler (1992) indicate that it is not effective for better public service delivery and performance. A charter school stays far away from the risk of this centralized government intervention. A charter school can make its own curricula that do not follow government guidelines, and provide an organizer or teacher with a chance or power to try a new educational tool (Lee, 2014; Lee which is part of the reinventing government movement the accountability that a charter school stresses is similar with the results oriented spirit of the reinventing government movement A charter school strives to improve its student p s to

PAGE 34

18 measure the results of what a charter school performs in terms of better student performance. The final evaluations analyzing these student achievements affects authorizers to decide if a charter school operates or closes. For instance, it is mandatory for all of the charter schools in Colorado to take a student evaluation exam called do not school will be expelled After Minnesota became the first state passing charter school legislation in 1991, the No Child Left Behind (NCLB) Act was passed in 2001. This legisl ation emphasize s educational choice, performance, and testing to save and reform the American TPS system at risk (Spring, 2005). Meanwhile, the NCLB Act pursuing these methods has helped many charter schools create their own unique characteristics that emb race the spirit of the reinventing government movement The Colorado charter school case introduced in the next chapter will be a practical example of a charter school using the reinventing government movement Research Topic a nd Scope Under a strong charter school law, charter schools have impressively proliferated in Colorado (Lee & Jeong, 2012) 10 However a gap has started to appear in the charter school policy implementation among school districts. Some school districts have delivered more charter school services to their residents while other school districts have provided th eir residents with fewer charter school services. That i s, some school districts actively use a charter school policy to innovate their own TPS system while other scho ol 10 CER (2012) ranked Colorado 9 th out of 42 jurisdictions (41 states plus Washington, D.C.) having charter school law. Washington, D.C. was ranked as the first jurisdiction, which means that Washington, D.C. had the strongest charter school law in 2012.

PAGE 35

19 districts have rarely or never used a charter school policy. This situation indicates that there is a wide varia tion local charter hool districts vary in the extent to which the y actually implement the state charter school law. Glimpse of Theoretical Approaches To empirically investigate why uneven degrees of the local charter school policy implementation emerges this dissertation targets school districts in Colorado and utilizes factors representing their attributes as explanatory factors For the main intent of this study, this dissertation primarily focuses on the influence of diffused phenomena, policy entrepreneurs, interest groups, and policy networks that are nested or embedded in jurisdictions by explicating the units of analysis of Colorado school districts at the local level. Namely, t o answer th e aforementioned research question, the four main theoretical approaches the policy diffusion, policy entrepreneur, poli cy interest group, and policy network model s provide critical insights to analyze the policy phenomenon regarding why some school districts in Colorado actively implement the state charter school policy and others do not. Key Definitions Implementation of innovative institutions (policies and programs) is a general terminology that explains policy actors transitional phenomena in the public policy area (Lee & Kim, 2010) Generally, the definition of policy implementation means that policy actors in practical areas choose and carry out institutions that decision makers at the state or federal level enact. Furthermore, the term policy implementation can include the

PAGE 36

20 meaning that upper policy actors supervise lower policy actors policy proces ses as well as provide lower or supervised policy actors with financial support (Lane, 2000) Meanwhile, t he meaning of policy implementation is par a lle le d with the operational level of public management proposed by Rabb and Winstead (2003), in which polic y actors strive to achieve organizations objectives determined at the strategic level. They define public management at the strategic level and operational level as follows: Public management at the strategic level includes the stage of public policy formulation and public management at the operational level means to carry out programs or proces ses created to complete goals decided at the strategic level. Therefore, in the dissertation, the charter school policy implementation means that school districts as units of analysis operate charter schools that are authorized on the basis of the state ch arter school law oversee their operating processes, and help fund their budgets That is to say the term charter school policy implementation used in the dissertation broadly means that a school district o perate s charter schools, oversee s them, and provi de s financial aid to them If a school district does not do these actions for charter schools the sc hool district is regarded as a school district not implementing the state charter school law. Meanwhile policy implementation is expressed as public service delivery of policy supply actors as well. Brynard (2005) supports this view. He defines implementation as action where a policy actor carries out, accomplishes, fulfi l ls, or o emphasizes that this definition of implementation is equated with the definition of service delivery. In this dissertation both term s charter school policy implementation and charter school service delivery are interchange ab ly used. Along with this simi larity of both terms t his study expresses

PAGE 37

21 variation of the local charter school policy implementation as variation of the local charter school service delivery. Based on variation of the local charter school policy implementation or the local charter scho ol service delivery appearing in Colorado school districts, this disse rtation aims to explore why variation of this local charter school policy implementation (or the local charter school service delivery) emerges among Color ach chapter of this dissertation is designed to account for this variation of the local charter policy implementation seen in Colorado. Additionally, the author does not forget that charter school establishments at the local level are joint functions, whic h indicate that there are interactions of both applicants and authorizers to establish charter schools in the school districts. T he school district governments as authorizers cannot authorize and implement the charter school policy in their area if there a re no applicants (individuals or organizations) that are eager to establish and operate charter schools. This application process attenuates the pure meaning of policy implementation. To clearly define the meaning of policy implementation and to reasonably defend the joint function phenomeno n in application processes that weaken the meaning of policy implementation, this dissertation must clarify a definition (scope) of a school district and the units of analysis. In the dissertation, the meaning of the sch ool district is not school districts as governments but as their specific regional area. P o licy actors as implementers in the dissertation are not school districts as a government al unit but school districts as jurisdictions S chool districts as government s are parts of school districts as jurisdictions. This definition of

PAGE 38

22 average income and a variable of networked phenomena betwe en each school district and seven public organizations at that state level can be used together because both variables are parts consisting of attributes of school district jurisdictions. Therefore, the dissertation defines that all actions that school districts as governments conduct are parts consisting of characteristics that school district as jurisdictions portray and possess. Units of Analysis s chool districts as jurisdictions Scholars in the public administration and policy area define the units of analysis as major settings or entities that researchers want to analyze in their studies. They indicate that the units of analysis in a study can be one among individuals, groups, or geographical units (towns, cities, counties, states, or countries) in accounting for social phenomena ( Rassel, & Berner 2003 ; Remler & Van Ryzin, 201 1 ). T his dissertation embraces individuals and organizations in each school district and utilizes their attributes to explain why broad variation in the implementation of charter school policy appears among Colorado school districts This is possible because In this sense, the clarification for the units of analysis helps account for the following: 1) why many predictor variables are ral attributes and 2) why joint functions of authorizers and organizers are not serious obstacles for utilizing a terminology of policy implementation in the dissertation Explanation of Theoretical Approaches T he main theoretical approaches that the dis sertation utilizes are the four mentioned theoretical approaches. The first is the diffusion model Public administration

PAGE 39

23 and policy scholars have typically been interested in using the diffusion model in studying the policy formation stage because the policy diffusion model provides scholars with a manifest explanatory factor regional diffusion factor which helps scholars understand mechanisms of policy emulation and spread from the first innovators to other innovators. Along with the content of the pol icy diffusion model, many studies for the policy formation stage at the state level indicate that neighboring states influence other jurisdictions to formulate the same or similar policies and programs ( Berry & Berry, 1990, 1999, 2007; Rogers, 2003; Walker 1969, 1981). Stewart, Hedge, and Lester (2008) propose that both policy formation and implementation stages are continuous and sequent ial action of policy actors because after policy actors adopt a specific policy they proceed to the next stage of the p olicy process policy implement ation immediately 11 This intellectual perspective proposes that policy formulation influences the likelihood of policy implementation th is study attempt s to apply the policy diffusion model in research topic analyzing variation of the local charter school policy implementation. 12 This trial can be an academic example of actually conducting pluralism. The second is the policy entrepreneur model. S everal scholars (Mintrom, 199 7, 2000; Roberts & King, 1991, 1996 ; Teske & Williamson, 2006) prove that both policy formation and implementation stages are affected by policy entrepreneurs because policy entrepreneurs are to introduce local state and federal governments t o new 11 Th eir view produces an advanced academic contribution by applying the diffusion concept to a research topic analyzing the policy implemen t ation stage and proving its empirical role on the policy implemen t ation stage. 12 This study notices that applying the d iffusion model to the policy implementation stage research does not mean that policy formulation always leads to policy implementation. The author concurs with Brewer and ch policy formulation is not always successive to policy implementation.

PAGE 40

24 innovative polic ies and programs and to encourage these governments to choose and implement their innovative ideas. In the case of the charter school policy implementation, an increased potential to implement the charter school policy if they have very active policy entrepreneurs who prefer the school choice movement. The third is the policy interest group model. T he school choice movement might be related to the many intere sted groups that demand more effective achievements of public schools. The interest groups in public education al policies are religious groups, teacher associations, local (state or federal) education departments, parent groups, and business educational as sociations ( Fuhrman, 2001; Kirst, 2007; Malen, 2001; Mawhinney & Lugg, 2001; Mazzoni, 1995; Osguthorpe, 2002; Vergari, 2000; Wirt & Kirst, 1997 ). Each interest and concern is related to the goal of improv ing achievement s It seems that wide ranging and diverse interest groups are essential to the likelihood with which school district s implement innovative public educational policies in an effort to consolidate public school achievement s The final theoretical approach is the policy network model. O r ganizations pursuing the same goal formulate their clusters and complexes called policy networks (deLeon & Vogenbeck 2007; Mintrom, 2000; Schneider et al 2000). Their policy networks can facilitate both decision makers and policy implementers to formulate and implement innovative policies that interest groups favor. Therefore, this dissertation tests whether school districts with strong policy networks of these organizations have an increased potential to implement the state charter school poli cy.

PAGE 41

25 Disser t ation Chapters With the four primary theoretical approaches and factors representative of each school district s characteristics the dissertation strives to find more accurate answers mplementation of their state charter school policy. This dissertation consists of five parts after this section. The first part of the literature review (Chapter 2) explains the general concept of charter schools and at the local level The second part of theoretical approaches (Chapter 3) reviews main theories relevant to policy diffusion, policy entrepreneur, policy interest group, policy networks, and other explanatory variables. The third part of the research methodology (C hapter 4) develops research design including hypotheses and a statistical analysis method to test them. T h e fourth part of the analyzed results (Chapter 5) reports final statistical findings obtained through the multiple ordinary least square s (OLS) regression analysis The final part of conclusions and implications (Chapter 6) arranges the empirical findings and proposes ideas for future studies. The remaining parts cover references and appendix.

PAGE 42

26 C HAPTER II LITERATURE REVIEW Introduction As mentioned in the introduction chapter, t his dissertation is interested in explicating the primary causes of why uneven implementation happen s in the case of the charter school policy. valuable to both academic and practical fields related to public administration and policy. In fact, scholars in the public administration and policy area have not enthusiastically studied t his research topic although they have shown a lot of academic efforts in analyzing mechanisms of the charter school policy adoption at the state level. Thus, this dissertation plays a contributing role as a n academic stimulator motivating scholars to study mechanisms of the charter school policy implementation stage as much as they have researched mechanisms of the charter school policy formulation stage. On the other hand, school districts are the main local public educatio n purveyors and suppliers that discover and deliver innovative education services that their customers want to receive as well as services they are not aware of This means that th e both decision makers and practitioners in school distric ts m eaningful insights that are necessary for understanding what factors school districts must first consider in i mplementing state charter school policy prior to conduct ing new education services for their residents. To achieve these academic contribution s and goals, this dissertation is constructed on the basis of a positiv istic research approach using a quantita t ive analysis tool. Scholars ( Fis c her, 1998 2003 ; Khakee, 2003; Witte 2000) point out that disciplines in the social

PAGE 43

27 science s area obtain the outcomes of their intellectual efforts with two philosophical academic approaches postpositivism and positivism. The former aca d emic tendency pursue s an answer oriented approach while the latter academic tendency utilize s a question orient ed approach. This dissertation is inclined to use the positivism approach to seek empirical answers to the research question : W hat predictor variables contribute to explain the uneven implementation of the state charter school policy in the case of Colorad o school districts? Positivists in social sciences obtain social knowledge rules, and laws on the basis of experiments or observations like natural scien tists find natural knowledge rules, and laws (Alvesson & Skoldberg, 2009; Roth & Mehta, 2002). Their steps to reach final answers are as follows: First, positivists make hypotheses to prove social phenomena, using several theoretical appr o aches theories, models, or frameworks that can explain social phenomena. Second, they test causal inferences (hy potheses) with proper methodological analysis techniques. 13 Finally they obtain generalizable answers that can fit other observed and tested cases (samples) in the same social phenomena. Along w ith the usual analysis style hard science and empirical study of positivism, this chapter first introduces the basic logic of the school choice movement, looks at general content s of charter schools across the USA, and describes the current situation of charter schools in Colorado. T his chapter is the main chapter d rawing and explaining contents of charter schools that are utilized as the dependent variable in the two final equation model s 13 Roth and Mehta (2002) concret e ly highlight that positivists use both quantitative and qualitative analysis techniques for their empirical studies

PAGE 44

28 America s Education System All of the countries in the world commonly have two styles for education services : P ublic schools a nd private schools. Like other countries, the American education system has been develop ed on the se two main education systems. Since Protestants arrived to the Boston area, they established private schools on the basis of P rotestant denominations. These private schools became primary education suppliers that offered their residents education services until the 18 th century. After that, free tax funded public schools th century (Ho well & Peterson, 2006) Until now, both education systems have been evolving with unique characteristics in the USA This co existence means that both social equality and individual freedom of scho ol choice in American education are important (Witte, 2000 ). However, America s education system has generally emphasiz ed free tax funded public schools, which provide American s with similar education services on the basis of compulsory curricula that federal, state, and local governments make and guide ( Howell & Peterson, 2006; Lee & Jeong, 201 2 ) groups small thinkers and large doers since Thomas Jefferson as Governor of Virginia defined the two roles of public education in the late 18 th century (Vo llmer, 2010). As shown in his directions of public schools, he did not design public education to teach all children to high levels. His main ideas about public education were that all children should receive the same level of education and learn how to ob ey social rules and laws through public education services. Jeffer s shows that there are specific gaps between public schools and private schools in the USA.

PAGE 45

29 Witte (2000) in his book, The market approach to education made the clear distinction between public schools and private schools. As shown in Table II. 1, h is efforts depict that the fundamental principles of eac P ublic schools in the USA are fundamentally established to offer all childre n eq u al and universal education. On the other hand, private schools are basically created to accomplish specific 14 s ons between public schools and pri v ate schools, we can know that there are significant differences bet ween public schools and private schools. One of them is that local, state, or federal governments provide public schools with fund s 15 while private schools obtain their fund s from private or non public sources such as charitable donation s endowments, and religious organizations (Choy, 1997 ; Kennedy, 2014 ) Thus, public schools must follow the conditions and guidelines that governments impose on them. 16 However, private schools are free r from government guidelines and controls compared to public schools 17 Re garding class size, Grossberg (2014) indicates that there is a difference between public schools and private schools. A private school classroom in the urban area has fewer than 10 students compared to a 14 Religious groups have been the main establishers of American private schools. Especially, Catholic churches have been very active in establishing and operating private schools to provide children with education including their religious dogma (Witte, 2000). 15 enues count on local, state, and federal funds. However, the dependence rate of its revenues is different with every state. For instance, in the cases of Colorado and Alabama, Colorado ), state funds (44.2%), and federal state funds (70.9%), and federal funds (10%) (Howell & Miller, 1997). 16 Kennedy (2014) points out that No Child Left Behind (NCLB) strongly regulates that public schools must comply with government regulations or laws. 17 Generally, state or local governments ask private schools in their jurisdictions to follow a minimum regulation in some areas such as health, safe ty, and instruction days per year (Witte, 2000).

PAGE 46

30 public school. 18 Meanwhile, both school entities seriously take the term accountability. But, e ach school entity expresses that this accountabilit y is quite different. First of all, public schools are responsible to the public. Their accountability is measured by tools tha t governments create. In contrast, private schools are just responsible to their clients (Witte, 2000) In addition, teachers in conventional public schools must be licensed and certified by standards of the state in which they work while it is not mandato ry for teachers in private schools 2014). Based on the compari s ons between public schools and private schools, charter schools possess unique characteristics of both school entities in the terms of their budget sources and their independent operating processes Therefore, charter school s ha ve been recognized as partic ular school choice movement tool s that have a hybrid characteristic between public schools and private schools 18 A public school in the urban area generally keeps its class size at about 25 30 students while a private school has about 15 20 students per class (Grossberg, 2014).

PAGE 47

Ta ble II.1 Differences between Public Schools and Private Schools Public Schools Private Schools Attribute Democratic / Collective Independent / Private Purpose ucation, Community satisfaction Internal organization, Family satisfaction, Profit or Equivalent Statutory construction Democratically controlled, Public organization, Highly regulated Independently controlled, Pr i vate organization, Very little regulation Organizational control: personnel, budgets, work rules, physic al plant, e t c External: by political and bureaucratic authorities Internal: by board, staff, and parents Product: grade structure; curriculum ; pedagogy; classroom organization; schedule External: by political and bureaucratic authorities Internal: by board, staff, and parents Accountability To the public; Means: external assessment To families; Means: family satisfaction Source: The mark et approach to education: An an a l s first voucher program ( Witte, 2000, p.13). 31

PAGE 48

32 America s School Choice Movement Since the American education system that emphasizes public schools was created in the USA, public schools ha ve Americans grow as citizens who follow social rules and laws well. However, American ( Hess, 2008; Lee & Jeong, 201 2 ; Walberg, 2007 ). This dissatisf actory performance of publi c schools causes federal, state, and local governments to endeavor to consider and introduce school choice programs because many education leaders and professionals diagnosed and proposed Ame They believe that the monopolistic TPS system Along with this new perspective, school choice programs have been introduced as prominent tools for on reform (Belfield & Levin, 2005; Schneider et al., 2000). Principles of the School C h oice Movement The primary goals of the school choice movement are to provide the American education area with innovative and new ideas by introducing a market based con ception to the TPS system suppliers (schools) and demanders (students and their parents) with more choice rights (Belfield & Levin, 2005; Chubb & Moe, 1990; Schneider et al., 2000). Abernathy (2005) oriented firms who are more responsive to customer interests and less responsi ve to the 3).

PAGE 49

33 cookie cutter approach embedded in the TPS system. The uniform curricula that are led by governments have disrupted the develop ment of Some students are good at fine arts and sports while some students really like math or writing. However, conventional public schools could not offer proper courses that fit each r curricula must follow govern m ent guidelines. Thus, wealthy parents in the USA have considered sending their children to private schools that can offer more diverse education curricula and better circumstances to their children. are usually recognized as education institutions that students from rich families occupy. There are serious limits in the case that famil ies with medium or low income s provide their children with better education services. Finally, the existence of private schools offering better education services makes parents who earn dissatisfied with conventional public schools. School Choice Movement Tools To improve conventional public school performa nce and get more children receiving good education services like private schools offer, school choice proponents have introduced some innovative institutions vouchers, charter schools, and magnet schools into the TPS system. T h e N o Child Left Behind (NCLB) Act, which was signed by former President George W. Bush in 200 2 helps students and their parents enjoy these innovative institutions (Colvin, 2004) and encourages parents with children enrolled in persistently low performing schools to transfer them to better public or choice schools (Corwin & Schneider, 2005, p. 18). Among these innovative school choice movement

PAGE 50

34 instruments the voucher is the leading innovative education service. The publicly funded vouchers ma k e it possib le for students from poor families to receive public finance support for private schooling (Belfield & Vevin, 2005). T hat is, vouchers help parents earning l ow income s send their children to private schools by assisting with a p ortion of the private school T h e emergence of vouchers became a start ing point leading American TPS system to be exposed to more competitive and market based world. By studying thousands of institutions that were created and conducted in the USA, we can easily realize that the USA has not been a static country but an active country in respect to its future. American leaders and decision makers have conti nu ously produc ed or reproduced numerous institutions to make the USA safer and stronger. The be en still changing and shaking the TPS system by introducing several innovative education inst ruments charter schools, home schooling, magnet schools, open enrollment laws, tuition tax credits, and vouchers which are real tools supporting the school choice movement These innovative education institution roles make the monopolistic TPS system more competitive and provide education customers with more education services (Belfield & Vevin, 2005; Lee & Jeong, 2012; Schneider et al., 2000). hools C harter schools a mong these innovative education institutions are the most recent concept among a series of trials to restructure and reform the TPS system as well (Miron & Nelson, 2002). Peterson and Campbell (2001) highlight that the original idea for charter schools came from magnet schools even though they have a certain difference in

PAGE 51

35 that magnet schools have a screening system where they can select qualified students while charter schools are open to all students However, the primary strength o f these two institutions is the same compared to conventional public schools. First of all, they can make and use their own curricula, which are free r from government guidelines and th a special talent can develop and improve their ingenuity Thus, the perception that these schools offer unique and special curricula is one of the main factors that education customers consider choos ing charter schoo ls or magnet schools over convention al public schools. Because of this strength, charter schools ha ve spread very rapid ly and impressive ly across the USA since Minnesota first legislat ed charter school law in 1991 (Lee & Kim, 2010; Walberg, 2007). In the USA, charter school s have currently been recognized as one of the representative and successful education tools assisting and facilitating the school choice movement They ha ve continuously provided scholars with interesting academic issue s comparison of performance between charter schools a nd conventional public schools or mechanisms of state s charter school legislation Definitions of Charter Schools Regarding the definition of charter schools, several scholars define them. Buddin and Zimmer (2005) write that are publicly funded schools that operate outside the direct control of conventional school districts and are under the authority of a quasi is means that there is an agreement called a charter between a cha rter school and authority governments which offer funds and approve autonomy to a charter school (Jefferson, 2004). Maranto Kayes, and Maranto (2006) and Miron and Nelson (2002) call charter schools hybrid styled

PAGE 52

36 schools that mingl e posit ive characteristics (public funding and universal access) positive characteristics (autonomy, choice, and flexibility) to operate for a specifi ed period of tim e to provide students a particular educational ( Maranto, et al., 2006, p. 2 ). As d escribed in these definitions, c harter schools spirits really emphasizing both democracy and liberty because they embrace two unique characteristics democratic equality (openness to all students) and individual liberty right) Meanwhile, education customers and scholars in reality give more scores to a n advantage that charter schools can have their own unique curricula that help students develop their gifts. Based on this stren gth of charter sch ools, Mintrom (2001) defines charter schools as centrifugal innovative education institutions that are free r from government controls than traditional public schools. These definitions to charter schools indicate that charter schools are unlike conventiona l public schools, whose system employs a command and control and a top down style through state guided policy development and management. C harter schools have a more autonomous and decentralized style, in which organizers manage and operate charter schools even though such schools receive financial support from local and state governments (Gruber, 2011) Charter schools are relatively independent of state and local educational governments, compared to conventional public schools (Howell & Peterson, 2006; Levin, 2009). This hybrid style of charter schools makes it possible for

PAGE 53

37 them to reinvent the TPS system by emphasizing a market choose to send their children to charter schools rather than being assigned to a school based on attend Charter schools are public schools formed and operated by qualified actors that are free from state and local government oversight (Cohn, 1997; Tucker & Lauber, 1995). Vergari (1999) concisely explains actors which are involved in charter schools as follows: Charter schools are legally autonomous educational entities operating within the public school system under charters, or contracts. The charters are negotiated between organizers and sponsors. The organizer s may be teachers, parents, or others from the public or private sectors. The sponsors may be local school boards, state school boards, or other public authorities, such as universities (p. 390) History of America s Charter Schools The idea of charter sc hools was introduced in the USA during the late 1980s. It was advanced by Dr. Ray Budde and Mr. Albert Shanker in 1988 ( Lee & Jeong, 2012; Vergari, 2007). Their main argument is that charter schools can enhance choice in public education without harming th e roles and functions of public schools (Schneider et al., 2000; Vergari, 1999). In 1991, Minnesota passed the first charter school law to help its TPS system reform (Buckley & Schneider, 2007; Mintrom & Vergari, 1996, 1997, 1998; Renzulli & Roscigno, 2007 ; Schneider et al., 2000; Stoddard & Corcoran, 2007; Vergari, 1999, 2007; Wong & Langevin, 2005, 2007; Wong & Schen, 2002). decision to adopt the charter school statute has stimulated other states to question if educational services that the TPS system offers their customers are the best. Following this basic doubt, 42 jurisdictions 41 states plus Washington, D.C. have passed charter

PAGE 54

38 school legislations over the subsequent years through the end of 2012 (CER, 2 012). Their fast growth has been remarkable compared to other innovative education tools ( Kemerer 2009; Renzulli & Roscigno, 2007; Vergari, 2007; Wong & Klopott, 2009; Wong & Langevin, 2007) Additionally, Jefferson (2004) explains that charter schools ar e generally located in all areas rural, suburban, and urban communities across the USA. As of the end of 2011, more than 5,000 charter schools are established across the USA (CER, 2012). was not evenly as signed in all the states that enacted the charter school law. As shown in Table II. 2, CER (2012) indicates that Mississippi is the state with the strictest authorization requirement for charter schools. In contrast, based on the same criteria for the evalu ation, CER (2012) ranks Washington D.C. as number one among the 42 jurisdictions that have passed charter school law. Washington, D.C. is the jurisdiction where it is easiest to establish charter schools in the USA 19 Table II. 3 reveals that as of December 2011, there are 107 charter schools in Washington, D.C. Regarding the number of charter schools, the state operating the most charter schools is California. There are 1,008 charter schools in C al i fornia nd most charter schools by operating 539 charter schools. By the end of the 2011 2012 school year, the two states Maine and Mississippi d id not have any charter schools with in their borders 19 r school law rank, evaluating Washington D.C. as the easiest jurisdiction means that Washington D.C. has a flexible charter school act allowing charter school establishers to easily make and operate their charter school. Reversely, the strict charter schoo l law has a contrary meaning. In the USA, Mississippi is the state having the strictest charter school law.

PAGE 55

ted Year Grade Rank State Year Grade Rank State Year A 1 Washington D.C. 1996 C 22 New Mexico 1993 2 Minnesota 1991 23 Oregon 1999 3 Indiana 2001 24 New Jersey 1996 4 Arizona 1994 25 Nevada 1997 5 Michigan 1993 26 Oklahoma 1999 B 6 New York 1998 27 Maine 2011 7 California 1992 28 Texas 1995 8 Florida 1996 29 North Carolina 1996 9 Colorado 1993 30 Illinois 1996 10 Utah 1998 D 31 Arkansas 1995 11 Missouri 1998 32 Rhode Island 1995 12 Idaho 1998 33 New Hampshire 1995 13 Pennsylvania 1997 34 Connecticut 1996 14 Louisiana 1995 35 Wyoming 1995 15 Ohio 1997 36 Alaska 1995 C 16 Wisconsin 1993 37 Maryland 2003 17 South Carolina 1996 38 Hawaii 1994 18 Delaware 1995 F 39 Kansas 1994 19 Massachusetts 1993 40 Iowa 2002 20 Georgia 1993 41 Virginia 1998 21 Tennessee 2002 42 Mississippi 2010 Source: The Center for Education Reform (2012) 39

PAGE 56

40 Table II. 3 Number of Charter Schools in Each State in the 2011 2012 School Year Rank State 2011 12 Rank State 2011 12 1 Washington D.C. 107 22 New Mexico 85 2 Minnesota 162 23 Oregon 116 3 Indiana 63 24 New Jersey 87 4 Arizona 539 25 Nevada 34 5 Michigan 316 26 Oklahoma 19 6 New York 201 27 Maine 0 7 California 1,008 28 Texas 444 8 Florida 517 29 North Carolina 105 9 Colorado 185 30 Illinois 105 10 Utah 85 31 Arkansas 37 11 Missouri 52 32 Rhode Island 17 12 Idaho 43 33 New Hampshire 11 13 Pennsylvania 170 34 Connecticut 23 14 Louisiana 113 35 Wyoming 4 15 Ohio 368 36 Alaska 28 16 Wisconsin 256 37 Maryland 46 17 South Carolina 48 38 Hawaii 31 18 Delaware 22 39 Kansas 19 19 Massachusetts 76 40 Iowa 8 20 Georgia 125 41 Virginia 4 21 Tennessee 35 42 Mississippi 0 Total 5,714 Note: In the case of Colorado, there is the gap of seven data. CER (2012) reports that Colorado has 185 charter schools. On the other hand, CDE (2012) indicates dependent variables in the dissertation. Source: The Center for Education Reform (2012)

PAGE 57

41 Spread of Charter School Laws T he cumulative number and proportion of the jurisdictions that adopted charter school law each year f rom 1991 to 2011 is shown in Table II. 4 Each cumulative proportion is calculated by dividing the number of cumulative jurisdictions that adopts charter school law each year by the total of 51 jurisdictions (50 states plus Washington, D.C) On the basis of each cumulative number or proportion a typical S shaped normal curve is drawn in Figure II. 1 as well Several scholars (Berry & Berry, 2007; Gr a y, 1973; Rogers, 2003) indicate that the S shaped normal curve accounts for how policy actors such as individuals, organizatio ns, local governments, state governments, or countries adopt a new policy over time. The pattern of the typical S shaped normal curve display s that the number of policy actors that adopt a new policy are very few in the initial period, their number is abru ptly increased during the middle period, and their number is very low in the final period. This implies that the majority of policy actors adopt the new policy through initial policy actor experi ences and performance of the new policy. Explaining contents of innovative policy adoption, this study adds one more category of actors called potential policy actors to the five categories innovators, early adopters, early majority, late majority, and l aggards Rogers (2003) distinguishes individuals in the process of adopting innovations. exclude potential adopters, who have not adopted an innovation yet, in the population for his study. The author includes them and extend categories. The reason for this is because in the USA, the state charter school law adoption is still continuous. Namely, as described in Figure II. 1, the stable situation of the

PAGE 58

42 state charter school law adoption fr om 2003 to 2009 was broken in 2011 since Mississippi and Maine adopted their charter school law in 2010 and 2011 respectively. The graph in Figure II. 1 shows that in 2010 the trend started to increas e once more. T he remaining states that have not adopted charter school law up until now still have a potential to adopt the charter school law soon or in the future. Thus, it is reasonable to consider including and mentioning the nine states as potential policy actors in explaining the state charter school law adoption. Moreover, it will be more valuable to account for accounts for each proportion of the five adopter segments, using individuals who adopt an innovation among whol Using perfect sample size is a contribution of the dissertation in explaining proportion of adopter segments. Additionally and parallel to this view, applying case of the state c harter school law adoption, this study has changed the five original names as follows: Innovative policy actors, early policy actors, early major policy actors, late major policy actors, late policy actors, and potential policy actors. 20 As drawn i n Figure II. 1, there were no jurisdictions that adopted charter school law in the USA before Minnesota voters passed charter school legislation in 1991. This time, there were 51 potential policy actors Washington D.C. and 50 states that might adopt charter school legislation before 1991. The S shaped normal curve of the state charter school law adoption shows that 28 jurisdictions passed their own charter school legislation during the five years (1994 1999) after the eight leading states named 20 analysis are school districts while the terms Rogers (2003) uses fit cases studying individuals. In addition, th is study evaluates that the word policy actor is a more familiar term than the terms Rogers used in the public administration and p olicy field. Thus, this dissertation uses the ne w terms.

PAGE 59

43 innovative po licy actors and early policy actors first and rapidly passed their charter school law between 1991 and 199 3. This number of eight almost fits the percentage (16%) that Rogers (2003) calculates adopter number and percentage in the beginning adoption period that both innovator and early adopters appear in. The percentage of jurisdictions adopting charter school law from 1994 to 1999 reached nearly 70% among the 51 potential jurisdictions. The author regards these 28 jurisdictions as both early major policy ac tors and late major policy actors. 21 This explanation is illustrated in the section where the slope is steepest on the graph shown in Figure II.1. This abrupt adoption phenomenon can be explained with the logic of transaction costs. 22 The experiences of the first eight jurisdictions adopting charter school law serve as examples of the effe ctiveness and legitimacy of adopting charter school law to other jurisdictions considering the charter school law adoption. 23 These initial jurisdictions include both innova tive policy actors and early policy actors from 1991 to 1993. Their total percentage occupies approximately 16% in the whole sample size. The beneficial information obtained from the leading states functions to reduce other jurisdiction transaction costs in considering creating their own charter school law. Previous good experiences and information of the eight leading states playing roles as innovative policy actors and early policy actors help other jurisdictions adopt charter school law. 21 The period that both early major policy actors and late major policy actors show up almost matches the adopters in accepting an innovation ar e completed during the periods that both policy actors take an action. 22 Holzinger and Knill (2005) argue that The fact that many others apply a certain policy serves as information that this might be the best thing to do (p. 784). This information deliv ery and social legitimacy embedded in it play roles as factors helping other policy actors minimize transaction costs when they try to adopt or implement a policy. That is to say, leading policy actors help following policy actors avoid paying for high transaction costs by informi ng the specific policy s information or social legitimacy to the following policy actors. 23 These initial eight states include M N (1991), CA (1992), CO (1993), GA (1993), MA (1993), MI (1993), NM (1993), and WI (1993).

PAGE 60

44 After many jurisdictions adopt charter school law in the short term, the S shaped normal curve becomes almost a flat line and gradually turns into a lin e with an increasing slope in Figure II. 1 because the remaining late policy actors acting as laggard states are very few. In the charter school law adoption case, the six states that are regarded as late policy actors between 2000 and 2011 are Indiana, Iow a, Tennessee, Maryland, Mississippi, and Maine. The six states were shown in the flat line until Mississippi adopted charter school law in 2010 The graph line increases even more in 2011 because Maine adopted charter school law that year. Th is study puts these six states in the category. The value of the 16% is 6.72 (=51 jurisdictions 0.16). Since 2012, there remain nine jurisdictions that have a potential chance of joining the charter school state club in the future. They are Alabama, Kentucky, Montana, Nebraska, North Dakota, South Dakota, Vermont, Washington, and West Virginia. We can refer to them as potential policy actors among 51 jurisdictions. As of the 2011 20 12 school year, the value of the potential policy actors is 0 .176, which means that all of the jurisdictions have charter school law when the number one of the perfect proportion value is completed by adding 0 .176 of the value to 0 .824 of the current cum ulative proportion value in 2011 because the perfect number of proportion is number 1.

PAGE 61

45 Table II.4 Cumulative Proportion of Jurisdiction Charter School Legislation by Year Year Number of States Cumulative Number of States Cumulative Proportion 1991 1 1 .02 1992 1 2 .039 1993 6 8 .157 1994 3 11 .216 1995 8 19 .373 1996 7 26 .51 1997 3 29 .569 1998 5 34 .667 1999 2 36 .706 2000 0 36 .706 2001 1 37 .726 2002 2 39 .765 2003 1 40 .784 2004 0 40 .784 2005 0 40 .784 2006 0 40 .784 2007 0 40 .784 2008 0 40 .784 2009 0 40 .784 2010 1 41 .804 2011 1 42 .824 Note: The value of the cumulative proportion is obtained by dividing the number of cumulative jurisdictions that adopts charter school law each year by 51 jurisdictions. Source: The Center for Education Reform (2012)

PAGE 62

Note: Each cumulative proportion indicates the value dividing the number of cumulative jurisdictions that adopts charter school law each year by 51 jurisdictions Source: The Center for Education Reform (2012) Figure II. 1: Graph of Cumulative Proportion of 51 s Charter School Act ( 1991 to 2011 ) 0.02 0.039 0.157 0.216 0.373 0.51 0.569 0.667 0.706 0.706 0.726 0.765 0.784 0.784 0.784 0.784 0.784 0.784 0.784 0.804 0.824 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 46

PAGE 63

47 Growth of America s Charter Schools F igure II. 2 and Figure II. 3 indicate that as of the 2011 2012 school year, there are approximately 5 7 00 charter schools across the USA since St. Paul in Minnesota authorized the first charter school City Academy in 1992. About 2 million students enroll and study in these charter sch ools (CER, 2012). In the 1992 1993 school year, just one charter school existed across the USA. Since then, many charter schools had been nationally established up until the 2011 2012 school year. In the 2001 2002 school year, more than 2,000 charter schoo ls were operated and as of the 2011 2012 school year, there are more than 5, 7 00 charter schools across the USA That means that the number of the current charter schools across the USA ha s more than doubled, compared to the 2001 2002 school year Additionally, the number of students enrolling in charter schools was more than 166,000 in the 1997 1998 school year. About 15 years later, the number of students enrolled in American charter schools ha s increased more than ten times by the 2011 2012 schoo l year. 24 This fast and impressive growth of b oth entities explains that charter schools in the USA have successfully played their roles in facilitating the T PS system reform and supporting the school choice movement. The poll conducted by CER in 2013 reports that more than 73% of American adults support the charter school institution. This result is similar with CER s 2005 poll results, which showed that nearly 78% of American adults support the charter school institution (CER, 2013 a ). This po ll shows that half of Americans have steadily supported charter schools. 24 CER (2012b) describes that as o f 2011, there are 132,180 K 12 schools across the USA. Among them, conventional public schools are 98,817 and private schools are 33,370. Meanwhile, there are a total of 54,876,000 students in American K 12 schools, 49,484,181 students are in conventional public schools while 5,488,000 students are in private schools.

PAGE 64

48 According to the demographics and socioeconomic factors of the students who study in charter schools in the USA, CER (2010) reports that the average number of students enrolling in ch arter schools are less than the average number of students enrolling in conventional public schools. In charter schools, about 54% of students are from low income families and approximately 50% of students are from at risk families. on highlig h ts that charter schools are preferred to children from the poor families than conventional public schools. In addition, there are more non Caucasian students than Caucasian students (non white students 52% vs. white students 48%) in the total ch indicates that students in need prefer charter schools more than students in wealth do and shows that charter schools are more attractive to non Caucasian students than Caucasian students. Through th is e scription, we can at least know that charter schools have high popularity to both groups poor students who need assistance and minority students and gain a clue as to why many charter schools across the USA have long waiting lists as well. Moreover, the impressive increases of both the number of charter schools and s the USA show that charter schools have played positive system. According to the poll that CER conducted, more than 75% of respondents support establishing charter schools if charter schools hold This explain s that many American educational leaders, practitioners, scholars, and customers have considered a charter school as a successful educational innovation for changing the American TPS system and improving its educational circumstances.

PAGE 65

49 Therefore, we can predict that charter schools will continuously spread and w ill be more actively operated across the USA based on the aforementioned status quo of charter schools increases of states enacting charter school law, charter schools, and students enrolling in charter schools each year

PAGE 66

Source: The Center for Educati on Reform (2012) Figure II.2 Cumulative Number of Charter Schools (1992 1993 School Year to 2011 2012 School Year) 0 1000 2000 3000 4000 5000 6000 1992-93 1995-96 1998-99 2001-02 2004-05 2007-08 2011-12 1 269 1205 2357 2996 4147 5714 50

PAGE 67

Source: The Center for Education Reform (2012) Figure II.3 Growth of Students in Charter Schools (199 7 199 8 School Year to 2011 2012 School Year) 0 200000 400000 600000 800000 1000000 1200000 1400000 1600000 1800000 2000000 1997-98 2000-01 2003-04 2006-07 2009-10 2011-12 166627 433797 684495 1149986 1536099 1941831 51

PAGE 68

52 Colorado was one state that need ed new educational innovations for a better educational environment at the end of the 1980s and the beginning of the 1990s. During this period, makers and leaders recognized the reform of K 12 schooling to be an important educational policy issue (Hirsch, 2002). To change and improve Colorado K 12 education circumstances, Colorado legislators first consider ed adopting two new educational tools These were vouchers and a tax increase. However, these educational instruments were rejected in November, 1992 ( Lee & Kim, 2010; system made Colorado educational leaders Governor Roy Romer, Republican Senator Bill Owens, and Democratic Representative Peggy Kerns more enthusiastic to find new innovative education institutions in order to accomplish Colorado K reform (Griffin, 2013; Lee & Kim, 2010). educational innovations to change the Colorado K 12 schooling syst em. One year later, in 1993, the Colorado assembly enacted and Governor Roy Romer signed the charter school legislation to alienate the huge barriers that prevent ed charter schools to be created (Hirsch, 2002; Lee & Jeong, 2012; Lee & Kim, 2010; Ziebarth, 2005). 25 Along with the inception and support of the state charter school statute, the Douglas County RE 1 school district and the Pueblo County 70 school district became the pioneers of school districts by authorizing the first two charter schools (Griffin 2013). These two school districts 25 The Colorado H o use approved Senate Bill 183 (SB 183) on May 11, 1993. In this vote, 41 house representatives expressed Yes to SB 183 while 23 house representatives were against SB 183. In one hour later, the Colorado Senate respected House s decision with 23 pros and 11 cons. On June 3, 1993, G o vernor Roy Romer finally signed to let SB 183 become Colorado s charter school legislation (Benigno & Morin 2013).

PAGE 69

53 established and operated the Academic Charter School and the Connect Charter School in 1993 (Benigno & Morin, 2013; Carpenter & Kafer, 2013; Lee & Jeong, 201 2 ). In the same year, 1993, five o ther states Georgia, Michigan Massachusetts, New Mexico, and Wisconsin passed their own charter school law as well (Miron & Nelson, 2002). 26 Growth of Colorado s Charter Schools Since former Governor Romer signed the Colorado Charter Schools Act in 1993, charter schools have been impressively and widely growing in Colorado (Carpenter & Kafer, 2009; Hirsch, 2002; Ziebarth, 2005). They were authorized by either their own school dist rict or the Colorado Charter School Institute (CSI). Both government entities supervise and support charter schools as well (Griffin, 2013). 27 As described in Figure II. 4, as of the 2011 2012 school year, 178 charter schools have flourish ed across Colorado. 28 s been about 90 times and two times compared to the 1993 1994 school year operating the two charter schools and the 2002 2003 school year running the 92 charter schools respectively The 178 current charter school number takes up nearly 10% of K 12 public schools in Colorado. 29 Among the 178 charter schools, 17 charter schools are schools offering their students all K 12 school grade levels while the other 161 charter schools are schools p roviding their students with partial K 12 school grade levels (Carpenter & Kafer, 2013). 26 ategories for the innovation adoption stages, these states are included in the early policy actors. 27 As of the 2011 2012 school year, 25 charter schools among 178 charter schools are authorized, supervised, and supported by the Colorado Charter School Institute (CSI). 28 Carpenter and Kafer (2013) report that 1 8 charter schools in Colorado have been closed between the 1997 1998 school year and the 2011 2012 school year. 29 As of the 2011 2012 school year, Colorado has 1,773 K 12 public sc hools

PAGE 70

54 As shown in Table II.5 and Table II.6 there are three school districts very actively operating charter schools, which means that school district very strongly impleme nts the state charter school law. They are Jefferson County School District R 1 (16 charter schools), Douglas County School District RE 1 (12 charter schools), and Denver County School District 1 (31 charter schools). The US Census Bureau (2012) describes that as of 2012, the total population of these three school districts is 1,478,435. This population occupies approximately 30% total population 30 The p average per capita income is higher than $31,039 which is average per capita income. Their population density is higher than 48.5 which is the density. According to charter school location, Finn Mano, and Vanourek (2001) and Jefferson (2004) point out that charter schools are located in all styles of communities cities, suburbs, and rural areas across America Likewise, Table II.5 indicates that charter schools in the 45 Colorado school districts, which have at least one charter school in their own territory, are geographically situated in the rural, suburban, and urban areas across Colorado (Ziebarth, 2005). 31 In addition Table II.7 illustrates that among the 178 charter schools, the 30 charter schools (about 21%) provide more than 600 students with education services while the 16 charter schools (about 9%) do fewer than 100 students 30 As of 2012, the total population of Colorado is 5,189,458 (The US Census Bureau, 2012). 31 suburban, and urban areas. This means that these school d average income, education level, or race are very diverse. Therefore, it is very difficult to define dist ricts have a charter school or do not have a charter school in their territory.

PAGE 71

55 with education services. students As drawn in Figure II.5, Carpenter and Kafer (2013) illustrate that as of the 2011 20 1 harter schools serve more than 80,000 students. When the two charter schools were operated in the 1993 1994 school year across C olorado, the number of students was just 187 (Griffin, 2013). In roughly two decades, the students enrolling in Colorado s chart er schools have dramatically increased by more than 400 times. population of Colorado charter schools represents approximately 10% of the 854,265 total student population in Col 12 public schools in the 2011 2012 school year.

PAGE 72

Source: Data from the Center for Education Reform (1993 2001 School Year ) and the Colorado Department of Education (2002 2012 S chool Y ear) Figure II.4 Growth of Colorad Charter Schools (1993 1994 S chool Y e ar to 2011 2012 S chool Y ear) 0 20 40 60 80 100 120 140 160 180 1993-94 1996-97 1999-00 2002-03 2005-06 2008-09 2011-12 2 29 64 92 123 148 178 56

PAGE 73

2012 School Year) School District # of Charter Schools School District # of Charter Schools 1 Academy School District 20 4 26 Lamar RE 2 1 2 Adams 12 Five Star Schools 6 27 Lewis Palmer 38 1 3 Adams Arapahoe 28J 6 28 Littleton 6 2 4 Aspen 1 1 29 Mapleton 1 1 5 Bennett 29J 1 30 Mesa County Valley 51 2 6 Boulder Valley RE 2 5 31 Moffat 2 1 7 Brighton 27J 5 32 Montezuma Cortez RE 1 2 8 Canon City RE 1 1 33 Montrose County RE 1J 2 9 Cherry Creek 5 1 34 Park County RE 2 2 10 Cheyenne Mountain 12 1 35 Poudre R 1 2 11 Clear Creek RE 1 1 36 Pueblo City 60 3 12 Colorado Springs 11 7 37 Pueblo County Rural 70 3 13 Denver County 1 31 38 Roaring Fork RE 1 1 14 Douglas County RE 1 11 39 St. Vrain Valley RE 1J 6 15 Eagle County RE 50 2 40 Steamboat Springs RE 2 1 16 East Grand 2 1 41 Strasburg 31 J 1 17 Elizabeth C 1 1 42 Thompson R 2J 2 18 Falcon 49 4 43 West End RE 2 1 19 Greeley 6 4 44 Westminster 50 1 20 Gunnison Watershed RE1J 1 45 Widefield 3 1 21 Harrison 2 4 46 Windsor RE 4 1 22 Jefferson County R 1 16 Colorado Charter School Institute 25 24 Johnstown Milliken RE 5J 1 Total 178 25 Keenesburg RE 3J 1 Source: The Colorado Department of Education ( 2012) 57

PAGE 74

Table II.6 Location of Charter Schools in Colorado (2011 2012 School Year) Authorizer, Supervisor, and Supporter Number of Charter Schools Location Douglas County RE 1 12 Suburban Jefferson County R 1 16 Suburban Denver County 1 31 Urban School Districts with 1 to 10 charter school s 94 Rural, suburban, & urban Total 153 Note: The 25 charter schools that the Colorado Charter School Institute (CSI) authorizes, supervises, and supports are excluded in the table. Source: The Colorado Department of Education ( 2012) Table II.7 Charter School Size in Colorado (2011 2012 School Year) Size of Charter Schools Number of Char ter Schools Percentage of Charter Schools More than 600 students 38 21% 500 to 599 students 18 10 % 400 to 4 99 students 28 16 % 300 to 3 99 students 23 1 3 % 200 to 2 99 students 25 14 % 101 to 199 students 30 17% Fewer than 100 students 16 9% Source: Carpenter and Kafer (2013) 58

PAGE 75

59 Demographics of Colorado s Charter Schools In regard to the demographics of charter schools, at the nationa l level, minority students enroll more often in charter schools than Caucasian students do. However, Table II.8 illustrates that demographics in total Colorado charter schools are different from the national level. In Colorado Caucasian students compose a majority of student groups in the state 2012 school year, th e total percentage of minority students enrolled in charter schools in Colorado is 42.83% while the total percentage of minority students enrolled in public schools is 43.99% (CLCS, 2013). Since Colorado passed its charter school law i n 1993, the percentage of minority students enrolling in conventional public schools ha s been higher than the percentage of minority students enrolling in charter schools up until now. However, Carpenter and ent rate of the ethnic minority students has been increas ing per year in Colorado. They found that as of the 2011 2012 school year, the enrollment rate of minority students is twice as high as the 2000 2001 school year. The percentage of minority students in charter schools across Colorado has increased from 27% to 43.5%.

PAGE 76

Source: Data from Carpenter & Kafer (2009, 2013), CER (1993 2001 School Year ) and C DE (2002 2012 S chool Y ear) & Griffin (2013) Figure II.5 Changes of Registered (1993 1994 S chool Year to 2011 2012 S chool Y ear) 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 1993-94 1996-97 1999-00 2002-03 2004-05 2007-08 2009-10 2011-12 187 6675 17119 28782 36658 56188 63799 83478 60

PAGE 77

Table II.8: Comparison of Races between Charter Schools and Conventional Public Schools in Colorado (2011 2012 School Year) Charter S chools Conventional Public Schools Notes Caucasians 57.17% 56.01% Minorities (one race) 39.92% 40.93% Alaskan Native, American Indian, Asian, Black Hispanic, & Native Hawaiian Minorities (two or more races) 2.91% 3.06% Total 100% 100% S ource: Charter school facts provided by the Colorado League of Charter Schools (CLCS) in 2013. 61

PAGE 78

62 Performance of Colorado s Charter Schools McBeath et al. (2008) explain that all kinds of schools and school districts strive to improve their the United States House of Representatives and Senate enacted the No Child Left Behind (NCLB) act in 200 1 in school c hoice programs (p.1). That means that their research does not confirm if student achievements in school choice programs are better or worse compared to student achievements in conventional publ ic schools. Establishing charter schools, charter organizers are generally responsible for matching their student achievements with authority government standards (Henig & Sugarman, 1999 ; Lee & Jeong, 2012 ). Colorado voters created the Colorado Charter Schools Act in 1993. This act emphasizes that all public schools in Colorado have been responsible for taking the Colorado Student Assessmen t Program (CSAP) curre nt ly named the Transitional Colorado Assessment Program (TCAP). 32 Thus, all charter schools must have all their Ziebarth, 2005 charter schools must take TCAP becau se there is a strong desire of the two authorizing school districts and CSI. They allow charter school organizers or establishers to run charter schools in order to improve the accountability of the K 12 public school body 32 TCAP evaluates student performance with the four levels: Advanced (student performance is over the standards), proficient (student performance reaches the standards), partial ly proficient (student performance does not reach the standards), and unsatisfactory (student performance is far away from the standards).

PAGE 79

63 Along with the main ordinance clearly expressed in the state charter school act, TCAP tests are conducted to students from grade 3 to grade 10 in the three academic areas school performance with conventional public school performance The most recent report charter schools is generally higher than student s schools in the 2011 2 012 school year Table II.9 addresse s that in the 2011 2012 TCAP results, charter schools h ave students who are more advanced or proficient in almost all of the school grades from grade 3 to grade 10 compared to conventional public schools. They just have a lower percentage of advanced or proficient students in the grade 9 and 10 s math and writing results and in After Ziebar th (2005) analyze d level between 1996 and 2004, he highlighted evements of schools at the high school level are not higher than other conventi on al public school achievements. Thus, charter schools in Colorado because of their achiev ements, have given Coloradans a good image.

PAGE 80

Table II.9 Comparison of A dvanced or Proficient Students in C harter S chools and C onventional P ublic S chools (2011 2012 School Y ear) Charter schools Conventional public schools Grade % of Math % of Reading % of Writing % of Math % of Reading % of Writing 3 75.7% 78.2% 55.3% 70.6% 73.4% 52.5% 4 76.1% 74.1% 55.7% 71.2% 66.4% 48.8% 5 67.3% 74.8% 62.5% 64.4% 69.4% 58.2% 6 64.1% 77.1% 61.1% 61.5% 72.9% 55.7% 7 56.7% 73.1% 67.9% 53.0% 68.1% 61.4% 8 53.9% 71.8% 62.5% 51.6% 67.1% 54.6% 9 35.9% 68.9% 50.6% 38.3% 68.4% 52.0% 10 28.3% 66.4% 46.1% 33.7% 70.2% 49.7% Note: An underline means that conventional public schools have higher percentage of advanced or proficient students than charter schools. In the 2011 So urce: Carpenter and Kafer (2013) 64

PAGE 81

Research Question s e vidence for why charter schools have been recognized as a successful education innovation in Colorado (Carpenter & Kafer, 2013; Lee & Jeong, 2012; Ziebarth, 2005) That is, after Coloradan voters passed the Colorado charter school legislation in 1993, the aforementioned two factors play ed ro les as indicators showing charter school s ides in Colorado. However this charter school growth has created an inequality in providing Coloradans with public education services. In other words, their growth has gradually brought charter school there are school district s strongly, weakly, or not at all offering their education customers charter school service s Rewording this varied charter school service phenomenon with the policy implementation lens, we can re expres s this situation as follows: t here is a broad variati on among the 178 school districts in respect to implementa tion of the state charter school law. And, academically paraphrasing this sentence, t he following research question is proposed : W hy does the variation in implementing the state charter school polic This dissertation constructs each chapter to empirically explore causes and reasons of variation shown charter school education services to their cu stomers. Previous empirical studies (Lee & Kim, 2010; Lee & Jeong, 2012; Mintrom 1997, 2000; Mintrom & Vergari, 1998; Renzulli, 200 5 ; Renzulli & Roscigno, 2005; Stoddard & Corcoran, 2007; Wong & Langevin, 2005; Wong & Langevin, 2007; Wong & Shen, 2002) indicate that several theoretical framew orks are utilized to develop hypoth e ses that help e mpirically test the research

PAGE 82

66 question. Along with previous similar research tendency, th is dissertation utilizes four theoretical perspectives that can support expla natory factors that seem to have relationships with the uneven implementation of the Colorado charter school policy which addresses the services to residents. They are the policy diffusion policy entrepreneur, policy interest group, and policy network models. These t heoretical approaches have main ly been used to analyze mechanism s observed in the policy cycle. Contents of each theoretical approach are introduced in the next chapter and are extended to support the development of h ypotheses testing the research question.

PAGE 83

67 CHAPTER III THEORETICAL APPROACHES I n troduction As one of the educational innovations t hat lead to change tem the concept of the charter school was delivered to Coloradans about two decades ago (Griffin, 2013) two charter schools, which were established in 1993, are more than 20 years old During this period, there have been many changes related to the number of charter schools, the number of students enrolled in them, and their achievements. These physical changes in Colorado charter school s h ave provided scholars and practitio ne rs with many research sources ( Carpenter & Kafer, 2013; Ziebarth, 2005) However, scholars have not studied a research area exploring (Lee & Jeong, 2012 ; Lee & Kim, 2010). Along with this intellectual insight, th is disser t ation chose a research question : W hat factors lead variation in the implementation of charter school policy among This research question style is a typical style of empirical research oriented to questions. Thus, this study primarily covers the four theoretical approaches and introduces hypoth eses made through their contents in this chapter. In general, empirical research studying mechanisms of the policy process is supported by theoretic a l approaches named theory, model, and framework (Ostrom, 2007; Sab a tier, 2007). Their roles provide scholar s with both logic and grounds for hypotheses development that might result in answers for an empirical study.

PAGE 84

68 Among these three theoretical approach styles theory, model, and fram e work th is research selects the four theoretical models policy diffusion, policy entrepreneur, policy interest group, and policy network model s because they are generally utilized in studying the first two stages policy formulation and policy implementation stages in the pol icy process. Moreover, the literature review indicates that many scholars (Chubb & Moe, 1990; Mintrom & Vergari, 1998; Renzulli & Roscigno, 2005; Schneider et al., 2000; Teske & Williamson, 2006) who are interested in the charter school topic have used the se four theoretical models in empirically examining mechanisms of the charter school policy process Therefore, it is reasonable to use these four theoretical models in generating hypotheses, which provide this empirical study with potential answers. In t he following sections, this dissertation covers contents of the four primary theoretical models, starting with the policy diffusion model. In addition, various hypotheses expressing school district characteristics are introduced in the later sections. On t he other h a nd, as Table III.l describes, t h is dissertation keeps an academic emphasis of Holcomb and Nigh t ingale in mind. Holcomb and Nightingale (2003) declare that scholars studying policy implementation analysis must consider that theoretical approaches supporting policy implementation analysis are from not a single discipline but multiple disciplines. Therefore, r eflect ing Holcomb and Nightingale s declaration to this dissertation the author strives to apply voices of scholars from diverse academic dis ciplines into explainin g theoretical approach es and creating hypotheses. to better explain unevenly implement the state charter school policy. As previously mentioned, this dissertation pursues an empirical research using a quantitative method. For this, th is

PAGE 85

69 empirical study creates the seven main constructs policy diffusion, policy entrepreneur, policy interest group, policy network, demographic factor, school related factor, and socioeconomic factor. This dissertation designates the s e seven construct names based on names of the four primary theoretical models and segments of school district characteristics. This means that several hypotheses inv olved in each construct are introduced in this chapter as well. Thus, the roles of the four theoretical models help hypotheses in each construct fit the research question, lending their main ideas to each hypothesis. Meanwhile, constructs related to school district characteristics have their hypotheses with previous empirical proofs found through the literature review.

PAGE 86

Table III. 1 Theoretical C ontributions of M ultiple A cademic D isciplines in P olicy I mplementation A nalysis Academic Fields Theoretical Approaches and Methods 1. Anthropology Ethnography, cultural dynamics 2. Economics Microeconomic behavior an d theory macroeconomic theory, labor economics, economic development theory, industrial relations 3. Evaluation Evaluability assessment, individual impact analysis, program outcomes analysis 4. Law Constitutional law, labor law, civil rights, administrative law, legislative development 5. Operations research Systems analysis, decision theory, simulation analysis, grou p dynamics 6. Organizational development Organizational behavior, management theory, network theory 7. Political science Implementation theory, political behavior, political culture, governance theory 8. Psychology Cognitive development, communication, relationships 9. Public administration Public management theory, performance analysis, program budgeting/accounting 10. Public finance Cost benefit analysis, fiscal federalism 11. Public policy Policy process, policy analysis, simulation, public management, program evaluation 12. Sociology Social network theory, bureaucratic behavior, group dynamics/behavior, phenomenology Note: Th is table was originally created by Holcomb & Nightngale (2003). The original table was reorganized and edited to fit the dissertation s research topic 70

PAGE 87

71 Factors Affecting Innovation Adoption a nd Implementation Before exploring specific contents of the four theoretical models and hypotheses derived from the models in the dissertation, it is good to know about general factors that drive policy actors individuals, organizations, jurisdictions, and countries to consider accepting and carryin g out innovative institutions ( instruments ) policies, programs, rules, or services because this work can help us draw a broad layout in understanding mechanisms of the general policy process. for satisfying th e intent. Regarding processes of general innov ations, Klingner (2006) explains that innovations need processes of adaptation, anticipation and openness to change (p. 61 ). His explanation shows that innova tion occurrence in organizations starts when policy actors feel necessity of their own change. If so, we wonder what factors motivate policy actors to seek innovations. As answers to this question, Klingner (2006) neatly arranges v arious factors that lead policy actors to accept and carry out innovations in Table III.2 A s illustrated in T able III. 2 multiple factors influence policy actors to adopt and implement new innovations for their own change and reform Th is dissertation picked the four factors neighboring jurisdictions, entrepreneurs, interest groups, and linkage mechanisms under t hese six characteristic categor ies in Table III.2 Their contents are more specifically explained and treated in the following session s with the diffusion, policy entrepreneur, policy interest group, and policy network models. Four Theoretical Models Bri e fly explain ed t h e policy d iffusion model developed and articulated by Roger s (1962) and Walker (1969) has help ed scholars test why policy actors individuals,

PAGE 88

72 organizations, and jurisdi c tions accept innovations 33 Among various factors introduced and explained by the diffusion model, this dissertation focuses on the regional diffusion, which is one of the representative factors that the diffusion model employs in account ing for mechanisms of the policy process The primary content of this regional diff usion is to emphasize a phenomenon that a policy actor emulates neighboring policy actors. Th is study expects this emulation effect to explain reasons of variation shown in the case of implementation On the othe r hand, Kingdon (2003) and Mintrom ( 1997, 2000 ) emphasize that policy entrepreneurs help decision makers or public servants, who want to deliver better public services to their customers, adopt innovations that policy entrepreneurs bring by introducing and explaining innovations to decision makers. If there are many policy entrepreneurs working for education innovations in each school district, this might mean that customers in school districts receive more innovative education serv ic es. Thus, with this aca demic view, this dissertation treats the policy entrepreneur model as a primary theoretical approach. The third model of the policy interest groups is important because interest groups that want to receive some benefits on a specific innovative polic y always strongly affect the process that governments accept and carry out the policy. Scholars ( Schneider et al., 33 The stages heuristic called the terms policy cycle and policy stages is composed of initiation, estimation, selecti on, implementation, evaluation and termination (Brewer & deLeon, 1983; deLeon, 1999). The stages heuristic scholars explain that each stage has a sequent linearity (Walt, Shiffman, Schneider, Murray, Brugha, & Gilson, 2008). Based on the sequent character of the stages heuristic, we expect that the policy is implemented after a policy is formulated. In another words, the completion of the policy formulation stage leads to the next stage of policy implementation, in which an innovation is delivered to the p ublic Thus, this dissertation predicts that the regional diffusion factor can also play a role as a determinant explaining why a school district more actively implements the state charter school law (Of course, this study also agrees with Brewer and deLeon implemented after they are formulated). Moreover, a pluralistic approach, where a scholar analyzes a social phenomenon with multiple academic perspectives, supports that it is possible or valuable for this study to apply the diffusion model mainly used in the policy formulation stage to a study analyzing the policy implementation stage. This trial is an academic contribution of this dissertation.

PAGE 89

73 2000 ; Teske, 199 0, 1991 ) explain that interest groups are main policy actors strongly asking governments about public services that give policy actors more benefits. Decision makers and public servants cannot ignore their requests in the case of formulating and implementing innovative policies Therefore, interest groups that are strongly affected by charter school policy are expected to influenc policy implementation in Colorado. Regarding the policy network model Putnam (1993, 1995, 2000) defines policy networks as part of social capitals. Many studies (Heclo, 1974; Milward & Provan, 2000 ; 2012) argue that networks of policy actors are facilitators in helping specific policies to be formulated and carried out. Th is dissertation predicts that networks of organizations which work for the school choice movement with each school district, can facilitate each school district to deliver more charter school services to their residents On the basis of these aforementioned theoretical approach s summar ies the dissertation in the following sections, looks at the content of each theoretical model and develops hypotheses that help test the research question.

PAGE 90

Table III.2 Categories of Potential Factors for Innova s Adoption and Implementation Political Organizational Social/ Interaction Technical Human Spatial 1) Internal interest groups a 2) External interest groups a 3) Unions a 4) Form s of Government d 5 ) Policy entreprenurs e 1) Risk taking a 2) Attitudes toward change a 3) Focus on users needs b 4) Users context b 5) W o rk relevance b 6) Policy relevance b 7) Federal/state agency b 8) Number of employees b 1) Adaptation of products b 2) Acquisition efforts b 3) Linkage mechanisms b 1) Goal orientation a 2) Information a 3) Resources a 4) Qualitative products b 5) Quantitative products b 6)Theoretical products b 7) Focus on advancement of scholarly knowledge b 1) Graduate studies b 2) Function of position b 3) Decision making style c 1) Neighboring Jurisdictions f Source: Klingner (2006) originally cited this table and Lee (2014) also intr oduced this table in his article The author added the three factors form s of government, policy entrepreneurs, and spatial factor to Note: a: Julnes and Holzer, 2001; b: Landry Lamari, and Amara 2003; c: Web b er, 1987 ; d: Benton, 2002; e: Kingdon, 2003; f: Berry and Berry, 1999 & 2007 74

PAGE 91

75 Policy Diffusion Model To comprehend mechanisms of the policy process, several scholars (Rogers, 1962, 2003; Balla, 2001; Berry & Berry, 1990; Maks e & Volden, 2011; Renzulli & Roscigno, 2005; Walker, 1969 1981 ; Wong & Langevin, 2007 ) have used diffusion logic In their studies, they highlight that emulation is a primary concept in explaining the diffusion logic. E mulation means that a policy actor i ndividual, organization, jurisdiction, or country resembles or imitates other policy actors with previous experience. Among popular scholars who are dedicated to researching diffused n process consists of the modeling and imitation by potential adopters of their network partners inherently embraces that policy actors accept and deliver innovative policies and services through a process emulating other policy actors that have already experienced more positive performance of new policies. R eal actions that policy actors formulate and implement innovative policies or programs appear when policy actors have their imitating attitudes and lea rning minds that follow other policy actors that have already implemented policies successfully (Berry & Berry, 1994, 199 9, 2007; Lee & Jeong, 2012; Lee & Kim, 2010 ) To examine diffused phenomena, Walker (1969) posed a classic research question: how do t hese new forms of service or regulations spread among the America n which can be expressed as the term policies, spread among states in the USA. To seek answers of Walker s research question, scholars have used various diffusion styles proposed by Berry and Berry (1990, 2007). They extend the policy diffusion model by

PAGE 92

76 sharing diverse styles such as the national interaction, regional diffusion isomorphism, leader laggard, and vertical infl uence models. Among them this dissertation tries to connect and integrate the regional diffusion model and vertical influence model with the mimetic isomorphism and coercive isomorphism respectively 34 The studies of DiMaggio and Powell ( 1983 ) and Powell and DiMaggio ( 1991) explain the concept of isomorphism with the three categories such as mimetic, coercive, and normative isomorphism; T he mimetic isomorphism views that organizations are changed by imitating other organizations that have successfully experienced specific policies or services; the coercive isomorphism explains that organizations are evolved by responding to orders or pressure s from upper organizations; the normative isomorphism supports that organizations are changed by ada pting ethical norms or cultures shown in other organizations. Rivera and deLeon (2004) and Rivera, deLeon, and Koerber (2006) highlight that the core reason that an organization resembles other organizations is because other organizations have already and concretely possessed social legitimacy, which is formed by beliefs, norms, values, or definitions, through their previous experience and learning. Social legitimacy has a positive effect that reduces transaction costs when an organization chooses or implem ents new innovative policies (Lee, 2014). In short organizations already having social legitimacy in general become the main targets that an organization wants to resemble and to receive help. This dissertation treats the regional diffusion model a nd the mimetic isomorphism together to know the effect of neighboring jurisdictions on the local charter school policy implementation. First of all, the regional diffusion model supports that jurisdictions 34 This study brought DiM o this diffusion model section in order to offer more concrete logic in creating hypotheses based on the regional diffusion model and vertical influence model.

PAGE 93

77 emulate other jurisdictions that are geographicall y proximate (Berry & Berry, 1999, 2007 ; Lee, Chang, & Berry, 2011 ) Diffused phenomena usually happen due to neighboring jurisdictions stimulating a jurisdiction to learn and implement their successful policies and programs. 35 Namely, the regional d iffusion model particularly emphasizes that a neighboring jurisdiction is a determinant that stimulates a jurisdiction to have and carry out specific innovations through functions of competition or imitation (Lee & Jeong, 2012; Renzulli & Roscigno, 2005). Therefore, neighboring jurisdictions that have already experienced the same or similar policies are key factors in explaining mechanisms of processes that jurisdictions accept and implement innovative policies and programs. somorphism represented as contagious and mimic effects explains that a jurisdiction with near jurisdictions adopting and implementing new policies is more prone to espouse to new policies (Renzulli & R oscigno 2005). Thus, the transmission and spread of ne w policies can be influenced by the existence of neighboring jurisdictions choosing and implementing new policies already. Th is study uses a new term mimetic diffusion policy phenomenon This view possibly provides an intellectual insigh t in to why C school districts as policy actors implement the state charter school policy. We can test that the probability that school districts implement charter school policy will increase if neighboring school districts hav e already operate d th e state charter school policy Therefore, it is hypothesized that Colorado s school districts with many neigh boring 35 In the dissertation, the term neighboring jurisdictions indicate a jurisdiction sharing a bo rder line with a specific jurisdiction. A neighboring jurisdiction can be interchangeably used with an abutting or proximate jurisdiction as well.

PAGE 94

78 school districts that implement the state charter school policy are more likely to implement the state charter school law To measure this mimetic diffusion variable, this study uses secondary data from the Colorado Department of Education ( CDE ) counts neighboring school districts that already implement the state charter school policy and creates a ra te by divid ing the number of neighboring school districts that have implemented the state charter school policy already by total number of neighboring school districts. Thus, the operational definition of this variable is a rat e of neighboring school districts having already implemented charter school policy This means that this mimetic diffusion independent variable is expected to have a positive direction for the dependent variable. Among several diffusion mod els, Berry and Berry (1999, 2007) point out that the vertical influence model provide s studies that look at mechanisms of the policy process with a good explanatory logic as well. Its core content is that institutional actions of a lower government level are influenced by government pressures at the higher level. In other words, policymaki ng actions of state governments are affected by federal government pressures (Daley & Garand, 2005). Thus, the idea of the vertical influence model is similar with the top In the same vein, DiMaggio and Powell (1983) use the term coe rcive isomorphism to explain why an organization is changed. They stress that a keyword of the coercive isomorphism is power. A weak organization resembles a strong organization can see that an organization is changed by following mandatory standards, sanctions, or ordinances that governments create. In addition, Rivera and deLeon (2004) and Rivera,

PAGE 95

79 deLeon, and Koerber (2006) apply the logic of the coercive isomorphism to their s tudy empirically examining why an organization participate in a specific program. Their study indicates that there is the statistically significant relationship between higher participation in the Sustainable Slopes Program (SSP) and coercive isomorphic pr essure. If the concept rule is SSP and policy actors that accept and participate in SSP are western USA ski resorts. That is their study demonstrates that western USA ski resorts are more likely to participate in SSP when federal government environmental oversight or state environmental pressures are higher. In their stud ies federal and state governments are described as stronger and upper organizations that influence western USA ski resorts. This phenomenon can appear among states as well ; S tates that depend more on federal financial support are more likely to adopt or carry out federal laws (Daley & Garand, 2005) In a similar context, Portz (1996) demonstrates that financial support from upper governments is a key factor in leading governments at the lower level to select educational innovations. DiMaggio & Powell (1983) emphasize that resource dependence logic is impo which organization A will change isomorphically to resemble the organizations on which it depends for re 154). On basis of these empirical study results, we can posit that local governments accept and carry out policies due to state pressures at the upper level that provides funding or incentives. On the other hand, the literature review shows th at Colorado

PAGE 96

80 charter school law is highly ranked. 36 This means that Colorado support s charter schools and a school district in Colorado will tend to follow the is dissertation views that this is a reason that a school district in Colorad o has a charter school in their territory. Based on contents of both vertical influence model and coercive isomorphism, th is study hypothesizes that if financial conditions of school districts depend more on financial support from a state, school district s are more likely to implement the state charter school law. In this dissertation, a new term of coercive diffusion is used to consolidate different terminologies vertical influence model and coercive isomorphism that both public administration and polic y scholars and sociologists have use d respectively To estimate this variable, this study examines what percentage of state supported financial aid constitutes the whole budget of each school district. Thus, it is hypothesized that there is a positive rela tionship between this coercive diffusion independent variable and the dependent variable. Policy Entrepreneur Model This dissertation focuses on examin in g why dynamic mechanisms appear in the r school policy To answer this research question, t h is dissertation borrows logic from the policy entrepreneur model. Many scholars (Kingdon, 2003; Mintrom, 2000; Roberts & King, 1996, Schneider Teske, & Mintrom 1995) demonstrate impacts of policy entrepreneurs in the policy process. In general, policy entrepreneurs are important key policy actors in delivering new ideas to decision makers before public organizations deliver innovative policies to the public. In 36 As of 2012, Colorado is the 9 th state with the strongest charter school law across the USA (CER, 2012).

PAGE 97

81 the policy process Roberts & King (1996) highlight that policy entrepreneurs are pivotal actors in both policy formulation and policy implementation stages. In their book, they neatly define policy entrepreneurs and illustrate their roles as follows: Policy entrepreneurs a The term entrepreneurs is introduced in various academic areas, such as, sociology, political science, public administration, education, economics, an d business administration (Moon, 1999). Scholars put different words in front of the term entrepreneurs to clearly show their study area (Berry & Flowers, 1999) For example, entrepreneurs are called in the public affair s area as follows: Bureaucratic en trepreneurs (Baez & Abolafia, 2002), educational entrepreneurs (Hess, 2008; Teske & Williamson, 2006), policy entrepreneurs (Mintrom, 2000), and public entrepreneurs (Schneider et al., 1995). These different names of entrepreneurs mean that areas that entr epreneurs work for are very diverse. D iverse organizations expect entrepreneurs to help themselves change and reform by bringing innovative ideas to them. Scholars (Baumgartner & Jones, 1993; Kingdon, 2003; Mintrom, 2000; Polsby, 1984) in the public polic y area have proved that roles of policy entrepreneurs are to raise specific social issues and problems to help public organization s choose some innovative polices to resolve them by providing public organizations with information. P olicy entrepreneurs play roles of facilitators who help public organizations deliver innovative policies that serve as solutions to the public as well This study entrepreneurs is supported by Roberts and King (1991, 1996). Focusing on policy

PAGE 98

82 they reduced the traditional five policy process stages to four stages creation stage, design stage, implementation stage and institutionalization stage. Roberts and King (1991, 1996) policy process stages cover the first two stages and finally reach the third stage. About more than 15 years ago, the classic book, Public entrepreneurs written by Schneider and his co autho rs in 1995 descri change by being alert to opportunities for profits (broadly defined) that emerge in the sentence shows that policy entrepreneurs stimulate local governments to conduct innovative policies that are beneficial to each local government. Policy entrepreneurs are closely related to the condition that residents can receiv e better or innovative public services. In the policy process, policy entrepreneurs are driven factors that lead public organizations or deci si on makers to recognize what social issues are and bring innovative ideas that can resolve social issues. ell their policy ideas and, in so 423) and they problems and are responsible not only for prompting important people to pay attention, but also for coupling solutions to problems and for coupling b oth problems and solutions policy actors who accelerate innovative policies to be introduced to public entities including governments or decision makers. Their advice and explanations to public

PAGE 99

83 entities are good standards for decision makers to decide whether they accept innovative polices or not for their residents. Regarding characteristics of policy entrepreneurs, Fisher and Koch (2008) explain that policy entrepren eurs are positive, optimistic, and visionary risk takers compared to risk averse ordinary people. Thus, it is possible for them to invest their time and money and willingly offer their innovative ideas to public organizations or decision makers that pursue delivering better public services to their residents. Further more, policy entrepreneurs have exact timing sense that helps them know when a policy window opens. Namely, they know when they jump to solve social issues and to meet decision makers (Kingdon, 2003). Policy entrepreneurs can be expressed as risk takers who possess enough capabilities in introducing and explaining new innovations to governments. Their help aids governments in providing their residents with better public services (deLeon, 1996; Os borne & Gaebler, 1992; Roberts & King, 1991; Schneider et al., 1995). That is, policy entrepreneurs are smart initiators and helpers who provide public organizations with prop osed solutions when public organizations or decision makers need innovative polic ies. A ltruistic and sagacious policy entrepreneurs with creative ideas can be seen as diverse areas. Their job positions are not the same such as: Either scholars, researchers, journalists, and consultants working outside of the government or presidents, p residential staff, and political appointees working on the inside of government (Kingdon, 2003). Along with the aforementioned definitions, policy entrepreneurs explained as follows: Policy entrepreneurs 1) recognize social issues and problem s, 2) develop innovative ideas and plans 3) sell the ir ideas to public organizations and

PAGE 100

84 decision makers and 4) help public organizations implement their innovative ideas in the real world (Boyett, 1997; Roberts & King, 1991, 1996) In the public educati o n area, scholars (Hess, 2008; Teske & Williamson, 2006) call policy entrepreneurs educational entrepreneurs. Teske and Williamson (2006) define educational entrepreneurs as individuals seeking to instigate change in the public education system that will disrupt, transform, or radically alter the way education is provided (p. 46). Educational entrepreneurs like entrepreneurs in other areas are facilitators who try to make the traditional education system better by introducing innovative ideas. They want t o offer education customers students and their parents better educational innovations that help students performance to be improve d ( Gonring, Teske, & Jupp 2007; Smith & Petersen, 2008). Educational entrepreneurs have enough positive characteristic s to accomplish their goals and roles. Leisey and Lavaroni (2000) describe that educational entrepreneurs are entities with tenacious, optimistic, creative, courageous persistent, willing to take risks, resourceful independent, opportunistic, and thought ful characteristics (p. 28) Additionally, they describe educational entrepreneurs as entities who always think of how they can develop education better These characteristics of educational entrepreneurs are strong factors that make educational entrepren eurs work hard for jurisdictions they are involved with and bring innovative ideas to them (Finn, 2008) Education al entrepreneurs can be defined as entities willingly using their time and money for the education reform of jurisdictions they are interested in or are involved in. Many studies emphasize that educational entrepreneurs with g oal oriented characteristics must be key factors in the case of the education innovation and reform.

PAGE 101

85 Mintrom (1997, 2000) emphasize s that educational entrepreneurs are significantly related to the state charter school law adoption. Charter schools were innovative when they were first introduced in the USA more than two decades ago The concept of the charter sc hool was not familiar to many Americans. The two educational entrepreneurs Ray Budde and Albert Shanker played roles of educational entrepreneurs introducing charter schools as innovative tools to the USA ( Kolderie, 2005; Lee & Jeong, 2012). Their actions and efforts had been sho wn from about 10 years ago before Minnesota was the first state adopting charter school law in 1991 They have been the key players in the case of innovative education tools of charter schools spread ing across the USA through their articles, interviews, an d lectures etc. Thus, Budde and Shanker can be seen as pathfinders and educational entrepreneurs in the American charter school case. In the case of Colorado, the literature shows that there were the four political leaders in the case of Colorado s chart er school act. Griffin (2013) points out that they were Peggy Kerns, Barbara O'Brien Bill Owens, and R oy Romer They played roles as both decision makers and educational entrepreneurs in the process of Colorado s charter school policy (Lee & Jeong, 2012). Their efforts were key factors to Colorado pass ing charter school law in 1993. Teske and Williamson (2006) mention that educational entrepreneurs are in a critical position to accomplish educational innovations or resolve educational issues The four Colorado political leaders were typical samples of educational entrepreneurs in a critical position. They were leaders of decision makers and their positions were high enough to influence other decision makers. They felt that 12 publ ic school system needed a change.

PAGE 102

86 T his logic appears in the case s that school districts implement state charter school law as well S lower leaders or decision makers always exist in organizations. Thus, if there are knowledgeable policy entrepreneurs who often meet decision makers and actively deliver their information to them, an opportunity to reduce transaction costs, which can be estimated by the amount of time consumption between acceptance of innovations and their delivery to residents, increases ( Alston & Gillespie, 1989; Anderson & Parker, 2013 ; Barzel, 198 7) 37 That is to say, policy entrepreneurs help policy makers or implementers by introducing innovative strategies that reduce transaction costs, which occur in the exchange phrase of policy ch oice or implementation. If this logic is applied to the case of local charter school policy, this following research scenario may appear: 1) charter school organizers ask authorizers to allow their charter schools to be run, 2) authorizers need time to kn ow about the contents of charter schools that organizers want to establish, and 3) authorizers allow charter school organizers to operate the charter schools if the charter schools fit all of the conditions of state charter school law well. In these proces ses, i f school districts have many educational entrepreneurs introducing charter schools and providing authorizers in school districts with information relevant to charter schools, transaction costs that might occur in each step can be reduced. Therefore, if there are many educational entrepreneurs who work for the school choice movement in school districts, we can expect that th ese school districts will have more charter schools To operationalize this policy entrepreneur independent variable, th is study u ses the number of educational entrepreneurs in each school district 37 from trade, hires other inputs to carry out his idea, and finds ways to capture the returns associated with they point out that an entrepreneur is not a person who drives transaction costs to zero but a professional who just lowers transaction costs.

PAGE 103

87 acteristics, it is hypothesized that there is a positive relationship between this independent variable and the dependent variable. Policy Interest Group Model The intent of the research question raised in this section is to know if interest groups influence the uneven local charter school policy impl ementation Interest groups are primary political entities leading governments to adopt or carry out a specific policy (Cahn, 1995) After Alexis de Tocqueville looked at the American democratic society about two centuries ago, he reported that a group sty le connecting with each other leads American democracy (Cigler & Joslyn, 2002). That is to say this means that existence freedom and democracy. Freedom can be described as div s and democracy can be expressed as more p a r ticipation by many political entities individuals, groups, and organizations. In another word s i nterest groups in the USA can exist because the American government (country) basically allo ws and respect s individual or organization benefits and opinions. The USA is one of representative countries allowing individuals or groups to participate in processes making institutions rules and polic i es governance. It is bec ause institutions are so changeable and malleable that various ideas help institution makers shape the best institution (Boatright, 2011). Ideas and voices of multiple groups are necessary and sufficient conditions in shaping institutions although governme nts are the main actors. I nterest groups will exist forever in the USA if Americans do not lose their founding spirits. In general, many interest groups are

PAGE 104

88 involved in the public policy process as well (Birkland, 2001) However, they are usually located in either the expansive side as supporters or the defensive side as opponents for a specific policy. The author wants to labe l interest groups on the expansive side pro interest groups and name interest groups on the defensive side anti interest gr oups. 38 Scholars ( Baumgartner & Leech, 1998; Teske, 1991) in the public administration and policy area emphasize political acti vism and ro l e s of interest groups in examining mechanisms of the policy process shown in multiple policies. They underscore that policies are inclined to be designed and implemented according to benefits and desires. In reality, decision makers cannot ig n ore their requests due to their political power i.e. financial support or votes (Lee & Choi, 2011; Teske, 1991). Therefore, many p olicies voices In the public administration and policy area, interest groups are policy joiners who lead both policy makers and implementers to make or carry out policies that are beneficial to interest groups. M eanwhile, interest groups are political entities sharing essential information to accomplish the same goals and directions. Interest groups are recognized as critical political groups that truly influence the formulation and implementation of public polic ies related to their benefits (Andlovic & Lehmann, 2014) They are policy actors actively expressing their preferences about contents and directions of public policies. Baumgartner and Leech (1998) define interest groups as any associations of individuals or corporate entities that attempt to affect the policy process. Teske (1991 2004 ) indicates that interest groups do not hesitate to lobby 38 Some scholars (Hawkwins, 2011; Jeong, 2006; Lubell, Feiock, & Ramirez, 2005) distinguish interest groups as an ti interest groups or pro interest groups. For example, in the case of growth management policy, if interest groups are against growth management policy, they are called anti growth management policy interest groups. Otherwise, they are named pro growth ma nagement policy interest groups.

PAGE 105

89 legislators and bureaucrats so that policies that interest groups prefer are adopted or implemented if their benefits and profits are positively influenced by appearance and implementation of a specific policy. He explains the most active interest groups as ones both with the strongest benefit relation f or a specific policy and with the most impressive capabilities to organize each other to make a policy favorable to them formulated and implemented. His view stress e s that interest groups are often self interested and well organized political entities. Thu s, there is no doubt that interest groups have steadily increased in the USA strongly emphasizing both capitalism and democracy. M ore than 10 years ago, Nownes (2001) agreed with the aforementioned statement s point ing out that over 200,000 interest group s exist in the USA and they have also been increasing at the state level and local level. Their influence and size have been getting stronger and bigger in the USA that when scholars analyze mechanisms of th e policy process, it is very basic and important to understand how diverse interest groups play roles in the policy process and what their thoughts and reactions for a specific policy are. Interest groups are keywords for specifically explaining mechanisms of the policy process. Interest groups exist in the educational policy area as well. Mazzoni (1995) policy systems had become congested with individuals and group s trying to set agendas and shape decisions. The mainline K 12 groups representing teachers, administrators, and boards had been joined over the decades by a myriad of other organized interest in education. In addition, noneducation group s other than busin ess for example, parent, civic, urban, labor, farm, and foundation groups wanted to have a crack at changing

PAGE 106

90 involved in education policy formulation and implementati on and they want to actively join in changing schools and creating education policies. Mawhinney and Lugg (2001) indicate that such an expansion of interest groups has been seen in the public educational area since the 1970s so as to accomplish effectiv e public school performance. They might be student groups, parent groups, teacher s unions, for profit companies, and governments. Wirt and Kirst (1997) explain that interest groups in the public educational area are representative of a diverse spectrum suc h as political, economic, social, cultural, and religious sides. For example, conservative Christian groups in some states have greatly influenced public educational policies. These Christian conservative groups have a tendency to try to stop sexual educat ion programs from expanding, and block the implementation of comprehensive sexuality programs (Vergari, 2000). In Utah, Christian groups play an important role of providing phonics instruction (Osguthorpe, 2002). Furthermore, Fuhrman (2001) found that stat e officials in traditional interest groups have influenced a policy formation requiring stricter teacher certification processes. Regarding the school choice mo vement (SCM), Hill and Jochim (2009) highlight that many interest groups are influenced by SCM and its tools. They are public school employees, parents, and teachers unions. They are always for (pro) or against (anti) school choice movement tools according to existence or non existence of benefits for themselves. Interest groups pressure elected de cision makers such as school board members to accept educational tools that they prefer or are beneficial to them (Chubb & Moe, 1990). For instance, pro SCM interest groups want policy makers to adopt alternative education tools

PAGE 107

91 beyond the monopolistic TPS system constructed on the top down system, arguing that the American education system must be more competitive for its better performance and education services. If the logic of pro SCM interest groups is more reasonable, decision makers consider deliveri ng education tools related to the school choice movement to their residents. In regard to charter schools, Jefferson (2004) and Molnar (199 6 ) emphasize that there are several charter school sponsors and opponents. They might be state education governments school districts, parents, teachers, or other community organizations. 39 Among these interest groups in volved in charter schools, this dissertation is interested in interest groups that oppose charter schools. S everal scholars (Anrig, 2008; Finn, 2013; Gi llard, 2011; Urban & Wagoner, 2009) describe that teachers unions National Education Association (NEA) and American Federation of Teachers (AFT) are usually against SCM. It is because teachers unions argue and benefit s. Cooper and Sureau (2008) highlight that a primary role of teachers unions is Bierlein (1993) observe that state teachers unions lobby legislators to escape disadvantageous education policies to them. Mintrom (1997, 2000) and Mintrom and Vergari (1998) indicate that the primary interest groups that oppose it are teachers unions by p roving that state teachers unions have not usually agreed with charter school laws. Overall, scholars agree that teachers unions are anti SCM interest groups working against charter schools. A s Frumkin et al. 39 Charter school proponents argue that charter schools make the TPS system more competitive. Schools in the TPS system cannot be lazy in improving themselves and providing their customers with better education services. Thu s, existence of charter schools across the USA helps the total qualification of American education services to be better and its performance to be improved. Based on this blueprint, pro charter school interest groups support a charter school. Kirst (2007) introduces parent groups dissatisfied with local public school services, local business associations, real estate and developer associations, and fa i th based organizations as pro charter school interest.

PAGE 108

92 (2011) indicate, teachers unions like other opp onents for charter schools play roles of impeding charter school expansion by lobbying governments for undermining laws supporting charter schools and for reducing funding allocated to charter schools. Therefore, if school districts have many teachers who participate in teac h ers unions as anti interest groups for charter schools, the likelihood of charter school existence in their territory will be decreased. In sum, the interest groups model proposes that interest groups can account for the local charter school policy implementation or its variation. That is, interest groups can be pivotal factors in determining uneven implementation of the state charter school policy. In the dissertation t o know if interest groups influence the local charter school polic y implementation this study tests if school districts with strong er teachers unions take the position of opposing the local charter school policy implementation. As shown in the literature review, teachers unions are generally recognized as anti interest groups for charter school services To operationalize this interest group independent variable, the author uses the L ikert scale that indicates the percentage of teachers in each school district who participate in the National Education Association (NEA) o r the American Federa tion of Teachers (AFT). T h erefore, this independent variable is expected to have a negative effect on the uneven local charter school policy implementation Policy Network Model As shown in Table III.2 Klingner (2006) accounts for why innovation acceptance and delivery are affected by various factors. His intellectual effort emphasizes that factors influencing the process of innovative policies appear at all levels micro, meso, and macro research has been blossomed by studying relations between factors

PAGE 109

93 and both formulation and implementation areas of innovative policies. Among the multiple factors that affect the innovative policy formulation and implementation stages, the linkage mechanis (individuals or organizations) frequent connections meetings, e mails, or phone calls facilitating and stimulating their social networks are critical to explain why policy actors accept a nd implement innovative institutions. In se ve ral disciplines, the term networks has been used in accounting for inter individual or organizational arrangements. T he networks mean that a policy actor is connected with other policy actors. They co exist and cooperate with each other. In the world of networks, p olicy actors are not lonely entities. The concept of networks has been a remarkable term in describing policy a than four decades. Their roles shed lights o n explaining why policy actors produce better outcomes in public service deliveries. Fi r st of all, networks improve policy among them. H igher levels of trust play role s of lowering transaction costs and solving collective action problems that might happen at each st a ge of the policy process (Burt, 2000; Lubell Scholz, & Mete 2002). Additionally, some scholars emphasize that policy actors can m ore easily get information that they want to know about through networks (Borgatti & Foster, 2003; Granovetter, 1973 1982 ). Sociological scientists support the idea that interactions of multiple actors individuals or organizations are important factors to explain social phenomena (Aldrich, 1979 2008 ; Aldirich & Whetten, 1981; Freeman, 2004; Granovetter, 1973 1982 ) The network approach in sociology importantly treats interacti ve processes among interdependent organizations (Aldrich, 1979 2008 ). The star ting point of the inter

PAGE 110

94 o r ganizational approach is that organizations cannot ignore other organizations on the basis of the open system F or an organization to survive and develop it needs to require and obtain resources from other organizations. O rganizations naturally engage in exchange relations with each other, and network s formulated by mutually dependent organizations play role s as main stimulator s facilitat ing their resource exchanges. That is, the substantial attention of inter organizationa l theory is links (ties) among organizations which influence the exchange processes of resources and affect the processes sharing information (Aldrich & Whetten, 1981). T he view of inter organizational theory is that resource dependence perspective of int er organizations supports rationale of organization networks. W ith these networked phenomena, we can get clues or answers explaining 1) under what circumstances organizations depend on each other well and 2) how they obtain resources that are needed to sur vive or develop themselves. On the other hand, focusing on networks at the individual level, Granovetter (1973) shows that networked relationships are important for the job search process. He interview ed 100 males in Newto n MA from several job areas in order to understand and confirm the role of social connections when peopl e get job s He finds that job searchers benefit from using networks. His argument is that networks are sources of new information because they play a role serving as connections or bridges to informants in other disconnected and unfamiliar work areas. The literature r eview clearly portrays that sociolog i sts agree that networks significantly influence policy actors growth and survival at both individual and organizational levels. Scholars in the public administration and policy field became more familiar with the term networks after Heclo (1974) used the term issue network s His issue network s

PAGE 111

95 stimulate d political scientists to recognize the importance of social networks in analy zing mechanisms of the policy process. The issue network s focus on the patterns of close interaction s among several policy actors bureaucrats, political decision makers, policy consultants, researchers, etc The concept of issue network s especially emphasi zes the open system rather than the closed system of iron triangles fundamental concept. Its mechanisms are more diverse and complicated than iron triangles although both forms recognize that connections and commitment of participants are primary factors in explaining the decisionmaking process (Heclo, 1995) Iron triangl es are just composed of the three main po l icy actors interest groups, agencies, and legislative committees involved in a specific policy Therefore, a number of policy actors shown in iron triangles are less than a number of policy actors in the issue netw ork perspective and they do not have any networks with other groups except themselves In the case of diversity of information, iron triangles are weaker than networked structures. The major functions of issue networks are to help policy actors 1) define a social issue, 2) find solutions or alternative options, and 3) decide the most ideal policy that fits the social issue (Heclo, 197 4 19 95 ). In issue networks, each policy actor is an entity who knows about content s of a specific issue, considers solution s for the issue, and shares its information with other policy actors. Thus, policy actors in the category of issue networks know about information and knowledge for a specific issue through their networks. Since the term issue networks was introduced in th e public administration and policy area, scholars in this academic area have currently called them policy networks (deLeon & Varda, 2009; deLeon & Vogenbeck 2007 ; Mintrom, 2000) or have also named them policy communities (Kingdon, 2003 ; Walker, 1981).

PAGE 112

96 S everal scholars ( Agranoff, 2004; Bressers & O Toole, 1998; Howlett, 2002; Pro va n & Milward 2001 ) in the public administration and policy area have demonstrated that joint actions of policy actors individuals and organizations lead to facilitate public service deliveries and their performance. They support that a n organization that connects with other organizations pursuing a similar goal can more easily deliver public services it wants to carry out to their residents than an isolate d organization can do. Thus, a networked organizational setting fosters public organizations to deliver their services to residents. This dissertation focuses on testing if a networked arrangement, which is formed by se v er al organizations that work for the school choice movement in Colorado and each To know about the effect of policy networks on the local charter school policy implementation, this dissertation fi r st of all n eeds to find empirical research proving that networks play important roles in delivering charter school services to customers Th ere are several scholars who proved that a networked organizational setting is a facilitator in adopting education services. F tested if networks formed among policy entrepreneurs affect the state charter school policy adoption. In their study, they categorized network styles with the two forms external networks and i nternal networks. The term e xternal policy networks is interchangeable with the term inter policy networks, which indicates networks that policy actors working at different level s local, state, and federal levels collaborate with each policy actor. The te rm i nternal policy networks can be expressed as the term intra networks, which

PAGE 113

97 indicate s that policy actors at the same level cooperate with each policy actor. Their empirical study proves that internal (intra) policy networks facilitate both agenda setting and approval stages and external (inter) policy networks influence the agenda setting stage. These results p rovide this dissertation with a hypothesis ; Stronger policy networks create more chances in delivering innovative policies to their residents. This dissertation focuses on external (inter) policy networks rather than internal (intra) policy networks because the literature review reveals that school districts and thirteen organizations at the state level are main policy actors that play primary roles for the Colorado s school choice movement and education reform. Many scholars (Cibulka, 2001; Meier & O Toole, 2001; Torenvlied et al. 2012) indicate that there is a positive association between strong policy networks among organizations and educational innovation deliveries. Torenvlied et al. (2012) describe that the development of local A merican education has been affected by interdependent organizational arrangements. Cibulka (2001) explain s that the development of policy networks stimulates the growth of discourse and several demands for public school innovation s. Therefore, it is propos ed that the a main factor facilitating public educational innovation s Among organizations relevant to public education policy, school districts are very important entities because they decide to keep supporting or to stop almost all education services provided this is possible because school districts have their own taxing power in collecting their revenues. However, the two scholars point out that school districts have not been effectively operated by themselves in accepting and implementing education services. Their

PAGE 114

98 operation is totally involved with other organizations. School districts basically need to cooperate with other organizations to effectively implement education policies and get y argue that school districts are naturally familiar with networked organizational settings. It is usual for school districts to have a chance to cooperate with other policy actors state governments, state legislatures, private organizations, and other school districts for effective policy implementation and its outcomes. Thus, in many ways, school districts collaborate with other organizations at the state and local levels when they have education policies that they want to deliver to their residents. research topic. The mate there are the most active twelve movement. They are the organizations at the state level. Through the literature review, they are as f ollows: BEST Board, Boards of Cooperative Educational Services, Colorado Association of School Executives, Colorado Education Association, Colorado Charter School Institute, Colorado Department of Education, Colorado Department of Higher Education, Colorado Childr Colorado Legacy Foundation, Colorado State Board of Education, and Education Leadership Council. These public and non profit organizations are substantive organizations education ref orm. Each organization has its own unique work to school performance better. Commonly, o ne of them is to provide C s school districts with information legal support, financial support, or facility

PAGE 115

99 support related to the schoo l choice movement. For example, the Colorado State Board of Education is a deeply involved organization in enacting school choice movement. The Colorado Department of Education is a public organization offering a school district with administrative services for better public eduation. The Colorado Legacy Foundat ion plays the role of introduci ng a strategy to improve public education to a school district The financial assistance in improving their facilities. 40 This dissertation aims to test if a school district that has denser networks with o ther organizations delivers more charter school s ervices t o their residents. This study ol choice movement. Scott (2000) categorizes networks with the following three concepts density, centrality, and centralization. He defines each network style as follows: 1) network density is to measure how many connections actors in a social networked st ructure make, 2) network centrality is to measure how influential a prominent actor (ego) within a social networked structure is, and 3) network centralization is to measure collectively, how well connected actors in a networked structure are. Among thes e styles of networks, the dissertation uses the concept of network density to estimate inter (external) policy networks because this study targets to know if frequent ( many ) connections between a school district and the other organizations stimulate a scho ol district to more actively implement state charter school policy. Scott (2000) calculates the degree of network density by dividing the number of actual links 40 These public and non profit organizations are educ 12 12 public school reform.

PAGE 116

100 that actors in a networked structure make by the maximum possible number of links they can make This means that the degree of network density exists between zero and one because its calculation style is a proper fraction. Using the degree of each school this dissertation proposes that if a s chool district with stronger network density is more likely to deliver c harter school services to their residents. Thus, this network density independent variable will have a positive direction for the dependent variable. School District Characteristics Although the dissertation is interested in the empirical test for the explanatory power of policy diffusion, policy entrepreneurs, policy interest groups, and policy networks on the uneven implementation of the state charter school policy, it is fundamental for the study to test the impetus of other factors that could possibly affect the dependent variable. In fact, in the public administration and policy ac ademic area, there are many scholars that emphasiz e the importance of test ing jurisdiction characteristics to explain mechanisms of the policy process. Particularly public administration and policy scholars, who mainly use the I nstitutional Analysis and D evelopment (IAD) framework and policy diffusion theory for their empirical research, have discovered that jurisdiction characteristics strongly influence the policy process that appeared in several policy areas (Lee, 2014; Lee & Jeong, 2012). One of the a cademic core roles of the I AD framework leads scholars to know how critical jurisdiction characteristics are in accounting for dy namic mechanisms of the policy process. Using a flowchart of the policy process, Ostrom (2007, 2011), one of main scholars with in the IAD framework, explains that scholars must not miss testing

PAGE 117

101 roles of jurisdiction characteristics to certainly understand phenomena of the policy action arena. 41 Likewise, Berry and Berry (1990, 2007), who are representative scholars studying public policy diffusion mechanisms, highlight that examining impacts of neighboring jurisdictions is critical in accounting for dynamic mechanisms of the policy process (Tyran & Sausgruber, 2005) Along with this academic insight, t hey contribute to categorizing jurisdiction internal determinants into the three factors economic, political and social factors Up until now, they have consistently shown that jurisdiction charac teristics are important explanatory factors in explicating mechanisms of the policy process through their studies Following the contents that Ostrom and Berry and Berry emphasize, this dissertation classifies several school district characteristics into t he following three categories demographic, school related, and socioeconomic factors to complete the independent variables that are parts of the two final equation model s Demographic F actors Demograp h ic factors are proper in expressing school district characteristics. Future Americans will be getting more culturally or ethnically diverse older, and more educated (Coughlin & Tompkins, 2009) Thus, continuously chang ing These changeable and e volving d emographics become fac tors that influence the policy process (Lee & Jeong, 2012) That is, governments must consider changes in age, race, or population size in delivering public services to their residents. For instance, a white community supports sustainable policies more tha n a non white community does (Lee & Choi, 2011). This means that jurisdictions with more 41 policy process flowchart, the policy action in the dissertation indicates that school districts implement charter school law

PAGE 118

102 white resident s will tend to adopt and implement pro environmental policies rather than pro developmental policies. Demographics include factors relevant to ethnici ty gender, employment status, age, and location. A mong several demographic factors, this dissertation focuses on studying if both ethnicity and location factors are influential on the uneven implementation of the state charter school policy Ethnicity The USA is a country that is composed of Caucasians and diverse minorities. This situation explains that in the USA, an ethnic group is a critical demographic factor that government s must consider when they formulat e and implement public policies because t he formulation and implementation of several public policies might be directly related with each ethnic group s advantages or disadvantage s There is no doubt that e ducational policies are the public policy area that minorities are interested in (Good & Br aden, 2000; Levy, 2010). Its main reason is because currently, minorit ies social position in the USA is not quite high or strong. Educational success has been a good tool to improve people s social position in all of the countries (Lee, 2014) S uccessful education achievement s bring economic stability, wealth, fame, and power to people In general, success through e ducation al achievements has been recognized as a favorable way for the ascent of status In the USA, many minorities live at the bottom part of the social pyramid compared to Caucasians. When considering how it is human nature to pursue a better life, we can understand that minorities want to improve their life and want their children to live in better circumstances in the futu re. Educa tion can be a good tool that helps them accomplish their goals Thus, t his perspective makes an assumption that s chool districts with m any non white people are more likely to deliver more education services to their residents. In

PAGE 119

103 addition, s ome s cholars (Good & Braden, 2000; Lee & Kim, 2010) point out that non white famil ies are not satisfied by conventional public school services Minority parents want their children to receive better educational services through more public education service options. The study conducted by Stoddard and Corcoran (2007) demonstrates that school districts with many African Americans residents provide their residents with more charter school services. On the other hand some scholars ( Walker, 1969; Lee & Kim, 201 0) explain that population size is an influential factor in explaining why policies are formulated and implemented because there are many demands for public services as much as there are many residents. In the same vein, a school district with a larger stu dent population size has more demands from diverse education customers. This situation leads a school district to seek more innovative education services that Charter school s are regarded as new educational innovation s whi ch reduce the burdens o n educational organizations and officials generated by demand s ( Schneider et al., 2000). Thus, the contents synthesizing minority s and population size perspective s help t his dissertation posit if a school district with more minority residents is more likely to implement the state charter school policy. This independent variable is operationalized by the number of minority students in each school district. Therefore, the direction of this minority ind ependent variable will be positive for the dependent variable. Location The term l ocation in demographics can be broadly categorized into the following three areas urban, suburban, and rural areas These areas can be named other areas by population increa se or decrease in the long term. For example, a rural area can

PAGE 120

104 be changed to an urban area due to regional development and incoming population (local immigration) increase. This logic implies that jurisdiction location helps researchers know what kind of p olicies strongly or weakly run in a particular jurisdiction. In the USA, the majority of c harter schools are usually located in urban areas (Ziebarth, 2005). Urban areas have more population than rural areas do The National Charter School Resource Center (2010) reports that as of 2008, about 85% of charter schools are typically found in urban areas while approximately 15% of them have been established in rural areas across America. This situation can be accoun ted for by the law of demand and supply as well, applying the logic of the aforementioned population size to this independent variable together. Governments in urban areas are requ ired to provide their residents with more public services because there are more customers who demand multiple public services in their territory A government with more demanders is highly likely to supply more public services. Th is logic will similarly be shown in the charter school case. A s chool district in urban areas, where more education customers reside will be pressured by more education serv ic e demands from their residents. Therefore, it can be hypothesized that a school district located in urban areas is more likely to implement the state charter school law. Usually, population density is a good estimator informing where a city or a town is located. Thus, t his location independent variable is operationalized by each school o f each schoo l district. Based on both this hypothesis and this location variable s data, this study expects that this independent variable will ha ve a positive direction for the dependent variable.

PAGE 121

105 School related F actors School related factors embrace the following f our factors educational s ervice s capacity fiscal situation, market Among them, the three factors describe characteristics of school districts as jurisdictions and the one factor explains characteristics o f school districts as governments. The fiscal situation variable falls down the latter case other variables are used in ex plaining general characteristics of school district jurisdictions. Meanwhile, t he literature review supporting the fiscal situation variable is completed through scholars empirical research from both public administration and policy and sociology areas. Educational s ervice s capacity School districts try to adopt and implement innovative educational policies when their current education services are not sufficient enough to satisfy their customers (Lee & Jeong, 2012). S chool districts that provide their residents with poor education services more actively conduct policy action s that complement their poor education services by applying other innovative education log ic more than four decades ago He explains that jurisdictions with insufficient service insuffi cient education service capacities of school districts impact the uneven implementation of the local charter school policy.

PAGE 122

106 One empirical research conducted by Wong and Langevin (2007) supports logic that poor educational service capacities of ju risdictions become an influential factor that stimulate s a jurisdiction to find other innovative edu c ation tools. They prove that states education al service capacities play roles of impacting state government s to have their own charter school law. In thei r study, the ratio of students to teachers in each state is us A state where a teacher teaches more students is more likely to support the charter school law adoption Based on Wong and Langevin s finding, t h is study applies this logic to Colorado s school districts. This study hypothesize s that a s chool district with poor educational service capacities is more likely to implement the state charter school policy This independent variable is opera tiona lized by a ratio of dividing the number of total students by the number of total teacher s in each school district A h igh value of the ratio means that a school district has poor educational service capa bilities Thus, this independent variable will have a posi tive direction for the dependent variable. Fiscal situation situations. Scholars recognize that the pros and cons for the fiscal situation variable exist. First of all, some scholars agree jurisdictions attempt innovative tools. S cholars (Pfeffer & Salancik, 1974, 1978; Renzulli 2005 ) emphasizing resource dependency theory recognize that organizations with enough resources easily challen ge applying innovative tools to themselves. Namely, plentiful budgets or human resources of jurisdictions stimulate jurisdictions to adopt and implement innovative tools in delivering public services to their residents. A f inancial

PAGE 123

107 situation of a local gov ernment is a basic factor that lets a local government decide if it enforces its educational reform (Lee, 2014). Mintrom and Vergari (1998) agree with this logic for They support that high spending states consider formu lating innovative education policies because their strong financial situations can help themselves explore several innovative educational policies. This means that school districts with healthy financial sit u at ion s are more likely to try to adopt and imple ment innovative educational policies because their healthy financial situation s can cover high costs in conducting their educational reform. Based on this logic, this dissertation hypothesizes that a school district with healthy finan cial situations is mor e likely to implement the state charter school policy. On the other hand there is an opposite perspective for the aforementioned hypothesis. Wong & Langevin (2007) suggest that u nhealthy financial situation s of j urisdictions stimulate governments to find and try innovative tools because residents living in jurisdictions with an unhealthy financial situation know that they cannot receive better public services or more public service options due to the jurisdictions unhealthy financial situ at ions. Residents, who know that current education services will not be better, ask governments to find innovative tools bearing low costs and high benefits (Lee, 2014). Therefore, a school district with bad financial situations is more likely to carry out other innovative education tools. The study conducted by Wong and Langevin (2007) empirically argues for this logic. Reflecting both side perspectives, this study synthetically make s a hypothesis as follows : Jurisdiction fi nancial situation s motivate school districts to implement the state charter school policy.

PAGE 124

108 condition, this study uses an annual average of teacher salary as the operational definition. Therefore, this fisca l situation independent variable is expected to have a positive or negative direction for the dependent variable. Market based education tool School districts have other market based education tools supporting the school choice movement. This dissertation wonders if other market based education tools affect the implementation of the local charter school policy. Rationales for this research question can be found in some theories population ecology, resource dependence, or sociological neo institutionalism i ncluding the open natural system perspective. They illustrate that policy actors individuals or organizations are more inclined to consider accepting or implementing a new institution when they have been already embedded in or familiar with the new institu tion through the effects or legitimacy of other similar innovations (Lee & Jeong, 2012; Hannan & Carroll, 199 2 ). The concept of cognitive legitimacy means that a policy actor possesses information and knowledge about a new innovative policy through previo us experiences to the new innovative policy. This can be interpreted as follows: C ognitive legitimacy for the new innovative policy can lower transaction costs that occur in processes that a policy actor adopts and carries out the new innovative policy. Finally, cognitive legitimacy helps a policy actor accept and deliver the new innova tive policy. Therefore, policy actors more easily or actively accept or implement the new innovative policy if they have strong cognitive legitimacy for the new innovative policy through their previous experiences for other similar institutions. In sum, th

PAGE 125

109 legitimacy for a specific policy is a very important factor when a policy actor accepts and implements it, and other similar policies can be facilitators that help cognitive legitimacy for the specific pol icy to be formed in a policy actor. Renzulli (2005) applies this logic to her study and demonstrates that this logic is correct at the state level. Her article demonstrates that states adopting open enrollment laws are more likely to pass their own charter school law. Lee and Jeong (2012) concretely verify that the existence of similar public educational innovations increases the likelihood that school districts implement the state charter school policy as well. Based on the two empirical results, the disse rtation offers a hypothesis : I f a school district has other educational tools that support the school choice movement, it is more likely to implement the state charter school policy. The operational definition of this variable is the number of private scho ols that each school district possesses. Therefore, this independent variable is expected to have a positive association with the dependent variable. Pe r formance. L ow student performance is a main cause for school districts to consider adopting o r implementing education innovations (Lee & Kim, 2010). Education leaders or decision makers in school districts know that parents in their performance is low. Mintrom (2 000) supports this perspective by proving the hypothesis that states with lower average Scholastic Assessment Test (SAT) scores are more likely to pass charter school law. Namely, a state is inclined to accept other innovative education policies if its stu Corcoran (2007) discover that school districts with a high dropout rate among high

PAGE 126

110 schools more actively provide charter school services with their residents. On the basis of result, this study tests a hypothesis: A school district is more likely to implement the state charter school policy if its s This study uses the percentage of student dropouts of each school district to operationalize this ind e pendent variable. Therefore, it is hypothesized that this independent variable will have a positive association with the dependent variable. Socioeconomic F actors Some studies (Lee & Kim, 2010; Teske Fitzpatrick, & Kaplan 2006) emphasize that there is a the degree of information acquisition relevant to schools. P arents with high socioeconomic status obtain school information more than parents with low socioeconomic status do. Generally, jurisdi ctions with a higher economic level of residents cannot escape multiple demands from their residents (Berry & Berry, 1999 2007 ). Chubb and Moe (1990) argue that parents and residents who have better information regarding schools can ask public educational governments for more school innovations. Schneider et al. (2000) indicate that parents with higher socioeconomic status can get better information about schools by formulating and managing their networks. Therefore, their explanation shows that it is poss ible for school districts having residents with higher socioeconomic status to meet more requests for school innovations from customers. This study uses the two char level. Their hypotheses are as follows:

PAGE 127

111 ducational level Some scholars (Stoddard & Corcoran, 2007; Lee & Kim, 2012) demonstrate that a school district with more highly educated residents more actively implement the state carter school policy. Based on their empirical results, this study hypothesizes that s chool districts having a greater percentage of residents with a implement the state charter school policy. To operationalize this independent variable, this study utilizes the percentage of people is study expects that this independent variable will have a positive association with the dependent variable. ncome In reality, Schneider et al. (2000) address that parents earning high income usually can get more information about schools. This situation might stimulate them to ask for more education services to their school districts. Thus, this study hypothesizes that s chool districts having residents with high income are more likely to implement the state charter school policy To operationalize this independent variable, the dissertation uses the logged average per capita income in each school district. Finally, this independent variable is expected to have a positive association with the dependent variable.

PAGE 128

112 CHAPTER IV RESEARCH METH O DOLOGY Overview Scholars in the public administration and policy area mainly find answers to public policy or social phenomena with two epistemological approaches positivism and postpositivism ( Foster, McBeth, & Clemons, 2010; Witte, 2000). Positivistic research is based on a value free approac h expressed as the fact value dichotomy or value neutral as well (Fischer, 1998, 2003). It totally counts on scientific methods and statistical techniques in finding answers to a research question (Steward et al., 2008). Po sitivistic scholars argue that on ly empirical results which are tested and found by sci entific methods, can be knowledge. Namely, p ositivists argue that knowledge is obtained by analyzing generalized, objective, and observable data with scient ific methods. Foster et al. (2010) describe t hat positivistic research is totally oriented in quantitative methods, microeconomics, and rational approach. This means that the positivis principle is that diverse opinions and thoughts are strictly excluded in epistemological processes that knowledge is formed On the other hand, postpositivistic research is constructed on a goal value approach ( Dryzek, 2002; Fischer, 1998, 2003 ; Foster et al., 2010 ). Scholars oriented to postpositivism stress that we cannot reach absolute knowledge (or truth ) throug h generalized data and their scientific analyses because k nowledge is relative to everybody (Alvesson & Skoldberg, 2009; Khakee, 2003). F o r instance, specific knowledge can be acceptable to someone but can not be acceptable to others. In postpositiv i sm, at least, knowledge cannot be called knowledge if there is not an epistemological effort accepting

PAGE 129

113 and reconciling diverse opinions and multiple thoughts. Po stpositivists highlight that generalized data are not enough to explain social phenomena including mechanisms of both public administration procedures and public policy processes and their scientific methods have a limitation in seeking and completing knowledge that reflects or reconcile s several opinions and perspectives Therefore, postpos itivists hesitate to agree that a scientific discovery is the best way to form knowledge Generally speaking, the p ostpositivistic approach emphasizes the importance of both diversity and subjectivity in epistemological processes of acquiring knowledge. Li kewise, Lynn (1999) presents that it is fundamental to acknowledge that there are diverse views and subjective opinions in analyzing a policy. Therefore, p ostpositivists regard conversations or debates as excellent and pivotal tools in acquiring and formin g knowledge. Foster and his colleagues (2010) support that this postpositivistic approach helps scholars analyze nonrational social phenomena which are composed of metaphysical elements such as value, conflict, ambiguity or subjectivity. Witte (2000) co ncisely summarizes general research steps of each intellectual perspective as follows: P ositivism with a question and then searches for evidence and data to prove or disprove the proposition behind the question while postpositivism with an answer and then works back to evidence and the construction of a supportive local argument (p. xiii). The main research inquiry of t he dissertation is as follows: W state charter school policy? Based on Witte s intellectual efforts, we can say that t his research question is a typical question oriented research style that positivistic scholars prefer. Meanwhile, t his study follow s a typical cycle of positivistic studies as follows;

PAGE 130

114 positivistic scholars 1) start their studies with a research question, 2) search propositions or build up hypotheses that can help scholars find answers to the research question, 3) collect data related to propositions or hypotheses and analyze data with scientific methods 42 4) obtain empirical answers by confirming if propositions or hypotheses are correct, and 5) consider or regard a proved empirical result as knowledge explaining a specific social phenomenon. Following this cycle of positivistic studies this dissertation constructs each chapter to obtain generalizable and empirical research question. As mentioned in Chapter 1, this research question is a very unexplored topic because up until now few scholars have examined mechanisms of policy implementation with the charter school policy topic. This fact gives this study an academic motivation. Meanwhile, this dissertation takes a stand that finding empirical answers that can account for a social phenomenon is a st arting point that leads scholars to attempt finding goal value answers because empiri cal answers and data or information supporting them are good sources for post postpositivism does not exclude both scientific methods and empirical results in conducting a deliberative approach. He demonstrates that an empirical result is a good source supporting discursive processes and situational context that form a social construction ciling perspective to this dissertation, the author expects that empirical results, which are analyzed in this dissertation, become cornerstones or facili t ators that support a deliberative circumstance that postpositivists have construct ed for better Color ado charter school services 42 In this research ste p, positivistic scholars using qualitative methods discover evidence that support propositions or hypotheses through observations.

PAGE 131

115 In the same context, empirical answers of positivistic research are found on the basis of objective information and data relevant to the social phenomenon. Many scholars (Alvesson & Skoldberg, 2009; Khakee, 2003; Roth & Mehta, 2002; St eward et al., 2008) agree that objective data are a prerequisite for positivistic research. Thus, we can say that its final results can be clear cut. P ositivistic research first needs to accumulate unbiased data and analyze them with scientific met hods. Finally, objective data and empirical results can provide value oriented scholars with objective data necessary for their studies. Th erefore as an academic innovator or pathfinder of this research topic area, it is valuable for this dissertation to be conducted with a positivistic research approach focusing on a quantitative method. Through this positivistic approach th is study expects to find objective answers explaining me chanisms of the current Colorado charter school policy implementation well. Furthermore, the author wants these objective findings to play roles of pr o v i ding several actors de cision makers, practitioners, and scholars who services with meaningful and excellent information. Along with these whole intellectual perspectives and academic goal s, th is dissertation conducts scientific methods to analyze generalized data. T h e author investigates t research question with the multiple ordinary least square s (OLS) regression analysis under the multivariate analysis category. The mult iple OLS regression analysis implies that answers to th research question are obtained by analyz ing the associations between the dependent variable and several independent variables. A primary purpose of the multiple OLS regression analysis is to accou n t for variation in one dependent variable with a combination of multiple

PAGE 132

116 independent variables (Morgan, Leech, Gloechkner, & Barrett, 2013; Remler & Van Ryzin, 2011). In the multipl e OLS regression analysis, a de pendent variable must be a cont inuous (interval ratio and continuous ordinal ) variable while an independent variable is required to be a continuous or a dummy variable (Babbie, Halley, Wagner, & Zaino, 2013; Wagner, 2013). Thus, it is proper to employ the multiple OLS regression analy sis to this dissertation because the two dependent variables are continuous and all of the independent variables are continuous as well. Hypotheses As highlighted in the previous chapters the dependent variable in the dissertation is the uneven charter school policy implementation that appears among Colorado school districts. Potential answers for the dependent variable are completed by testing 13 hypotheses introduced in Chapter 3. Thirteen hypotheses were created on the basis of f our main theoretical approaches policy diffusion, policy entrepreneur policy interest group, and policy network models and school district characteristics. Their predicted directions betwe en the dependent variable and 13 independent variables are precisely d rawn in Table IV. 1 on page 11 9 The author paraphrases each hypothesis with their null hypotheses as follows: H 1 (mimetic diffusion MIDF) : School districts with many neighboring school districts that have already implemented the state charter school poli cy are more likely to implement the state charter school policy. H 1 ull hypothesis : School district s mimetic diffusion factor is not related to the uneven implementation of Colorado s local charter school policy H 2 (coercive diffusion CODF) : School districts whose state supported aid occupies a higher percentage of their own whole budget are more likely to implement the state charter school policy.

PAGE 133

117 H 2 ull hypothesis : School district s coercive diffusion factor is not related to the uneven impleme ntation of Colorado s local charter school policy. H 3 (policy entrepreneurs POEN) : S chool districts having many educational entrepreneurs are more likely to implement the state charter school policy. H 3 ull hypothesis : School district s policy entrepreneur factor is not related to the uneven implementation of Colorado s local charter school policy. H 4 (policy interest groups POIG) : School districts having stronger teachers unions are less likely to implement the state charter school policy. H 4 ull hypothesis : School district s teachers unions do not influence the uneven implementation of Colorado s local charter school policy. H 5 (policy networks PONW) : School districts having a strong density of policy networks between their own school d istrict and other organizations, which are related to public educational innovation and reform, are more likely to implement the state charter school policy. H 5 ull hypothesis : School district s network density is not related to the uneven implementation of Colorado s local charter school policy. H 6 (ethnicity ETHN) : School districts having more minority students are more likely to implement the state charter school policy. H 6 ull hypothesis : School district s minority students are not related to the uneven implementation of Colorado s local charter school policy. H 7 (location LOCA): School districts having a higher population density are more likely to implement the state charter school policy H 7 ull hypothesis : School district s population density is not related to the uneven implementation of Colorado s local charter school policy. H 8 EDSC) : School districts having more students per teacher are more likely to implement the state charter school policy. H 8 ull hypothesis : School district s educational service s capacity is not related to the uneven implementation of Colorado s local charter school policy. H 9 (fiscal situation FISN) : School districts having a sufficient and healthy fiscal situation are more likely to implement the state charter school policy or school districts having an i nsufficient and unhealthy fiscal situation are more likely to implement the state charter school policy

PAGE 134

118 H 9 ull hypothesis : School district s fiscal situation is not related to the uneven implementation of Colorado s local charter school policy. H 10 (market based education tools MBET) : School districts having more private schools are more likely to implement the state charter school policy. H 10 ull hypothesis : School district s private schools are not related to the uneven implementation of Colorado s local charter school policy. H 11 (stud STPE) : School districts having higher percentage of student dropouts are more likely to implement the state charter school policy. H 11 ull hypothesis : School district s student dropouts are not related to the uneven implementation of Colorado s local charter school policy. H 12 REEL) : School districts having more residents state charter school policy. H 12 ull hypothesis : School district s resident educational level is not related to the uneven implementation of Colorado s local charter school policy. H 13 REIN) : School districts having more residents with a higher income are more likely to implement the state charter school policy. H 13 ull hypothesis : School district s residen t income s are not related to the uneven implementation of Colorado s local charter school policy.

PAGE 135

119 Table IV.1 Predicted Directions of All Regressors on the Dependent Variable Independent Variables Predicted Directions Primary Explanatory Variables 1. Mimetic diffusion (MIDF) 2. Coercive diffusion (CODF) 3. Policy entrepreneurs (POEN) 4. Policy interest groups (POIG) 5. Policy networks (PONW) School District Characteristic Variables Demographic factors 6. Ethnicity (ETHN) 7. Location (LOCA) School related factors 8. Education al (EDSC) 9. Fiscal situation (FISN) 10. Market based education tool (MBET) 11. (STPE) Socioeconomic factors ducational level ( REEL ) ncome (REIN) positive positive positive negative positive positive positive positive negative or positive positive positive positive positive Note: MIDF=mimetic diffusion; CODF=coercive diffusion; POEN=policy entrepreneurs; POIG=policy interest groups; PONW=policy network s ; ETHN=ethnicity; LOCA=location; capacity; FISN=fiscal situation; MBET =market

PAGE 136

120 Data Collection The main research question of the dissertation asks what determinants affect the empirically explicate this research topic, this dissertation looks at the relationships between t he degrees of the local charter school policy implementation and the independent variables originated from the main theoretical models and school district characteristics. Among 13 total independent variables, there are five independent variables related t o the main theoretical approaches policy diffusion, policy entrepreneur, policy interest group, and policy network model s The rest of the independent variables are involved in school district characteristics. Among five independent variables related to t he main theoretical approaches, t h is research found the secondary data that help measure (operationalize) two independent variables mimetic diffusion and coercive diffusion under the policy diffusion category through the Colorado Department of Education an d F ind TheD ata website s respectively However, this research could not find data sources offering secondary data that help measure the other three independent variables relevant to policy entrepreneurs, policy intere s t groups, and policy networks. Survey Th e lack of secondary data for these three independent variables led to the conduct of two surveys. The first survey was composed of questions related to policy entrepreneurs, interest groups, and policy networks while the second survey was composed of quest ions only related to policy networks. Dr. Paul Teske the dissertation chair and educational professional, gave the author priceless comments and corrections

PAGE 137

121 about all of the survey questions in the first and second questionnair e His supervision appeared in the contents of the consent letter sent to all of the respondents as well. The target of the first survey was a superintendent in each school district while the second survey targeted representatives in 1 2 public and non profi t organizations at the state level. That is why superintendents are one of the most influential decision makers in selecting and implementing public educational policies in their own school dis trict, and representatives in 12 organizations are educational leaders who know about contents of relevance in mind, which means that recipients must be related to the target population. livan, Rassel, and Taliaferro (2011) emphasize that the respondent rate of a survey could be low when recipients think that they are not involved in the target population. su perintendent in each school district and a representative in 1 2 organizations are the best choice s for these two surveys. With thi s confidence, the author distri bu ted the first survey to a superintendent in each school district by mail and mailed the secon d survey to a representative in 1 2 organizations after receiving the university s approval. All of the formal survey procedures have followed the guidelines of the o rado Multiple Institutional Review Board (COMIRB). Bef ore entering in to a survey review, the author needed to take a qualification exam, which covers contents that a researcher must know in conducting a survey. The author passed the exam. Its passing date was at the beginning of January,

PAGE 138

122 2012. The university announces that passing the exam means that a researcher is well inf ormed of conditions that a quest ion nai re and its questions must involve or not. approval from COMIRB and tu rned them in to COMIRB on Feb ru a ry 1 3 2012. The author requested the expedited review instead of the full board review because the questionnaires do not include special tests and treatments to respondents. After turning to the application materials, COMIR B gave the protocol number (12 0218) on the submitted materials. Almost 3 days later, COMIRB confirmed to the author that all of the survey question na i r screened and evaluated each quest ion in the survey. COMIRB sent the author their approval letter named Certificate of Exemption by e mail (see the appendix A ). 43 This means that the university agrees and allows that the author proceeds and conducts the survey as a qualified and principal investigator. After receiving the approval letter from COMIRB, the author collected weeks. Respondents directories were obtained from the official homepages of each school district, all 1 2 organization s and CDE The population frame of this study consists of 178 school districts in Colorado and 1 2 organizations including public and non profit organizations at state level. T welve public and non profit organizations are as follows: Best Board, Boards of Cooperative Educational Services, Colorado Association of School Executives, Colorado Education Association, Colorado Charter School Insti tute, Colorado Department of Education, 43 14, 2012. Namely, COMIRB anticipates that this project is completed by February 14, 2015.

PAGE 139

123 League of Charter Schools, Colorado Legacy Foundation, Colorado State Board of Education, and Education Leadership Council. The total survey collection was completed through five rounds by mail. The surve ys in all of the survey rounds had been delivered to the targeted populati ons (178 school districts and 12 public and non profit organizations) by U.S. mail because a traditional paper survey has a higher response rate than an e mail survey or a web survey (Fraze, Hardin, Brashears, Haygood, & Smith, 2003). The author, on average, needed one month to complete all procedures of each survey distributing the survey to respondents, receiving answers from them, and re mailing to non respondents. The first survey was distributed to targeted respondents on March 3. The first responses were returned to the author in 3 weeks. Its response rate is about 30% among 178 school districts and about 2 5 % among 1 2 organizations. The final survey was ended in the first week of August, 2012. The final survey result indicated that 1 28 school districts among 178 school districts of the whole population cooperated with responding to this survey Among 50 school districts that did not respond to the survey, three school districts notified the author that they could not respond to the survey because the two superintendents were too busy and the superintendent was in an inte rim position This dissertation treats these 50 school districts as the missing case s. Among them, almost of all these school districts have very low population density (45 school districts have fewer than 100 people per square miles ). 39

PAGE 140

124 school districts are located in the rural area while 11 school districts are located in the urban or suburban area. 44 T able IV.2 demonstrates that this dissertation does not have a serious non response bias issue because b oth response and non response rat es of the survey h ave similar values on the basis of each school district s ( respondent s ) location 45 The two cases rural and suburban have the difference of 5% while the other two cases small rural and urban have the difference of 3%. This means that the respondents to the survey are not very different from the non respondents. Thus, we can expect that statistical results analyzing these samples are similar to or are representative of statistical results analyzing the total population ( McClendon, 2004; Remler & Van Ryzin, 2011) Table IV.2 Comparison between Responding School Districts and Non responding School Districts Location Description 128 Respondents (100%) 50 Non Respondents (100%) Small Rural 76 (59%) 28 (56%) Rural 37 (29%) 12 (24%) Suburban 9 (7%) 6 (12%) Urban 6 (5%) 4 (8%) Source: The Colorado Department of Education (2013) 44 Amo ng the 50 school districts that did not respond to the survey, there are the three school districts having more than 5 charter schools in their own territory. They are Colorado Springs School District 11 (7 charter schools), Jefferson County School Distric t R 1 (16 charter schools), and St. Vrain Valley School District No. Re 1J (6 charter schools). The website USA.com describes that as of 2009, the population .31 per square mile). As of 2010, the population rank of each school district is 5 (227,196 residents), 2 (537,431 residents), and 10 (151,981 residents) respectively. 45 Based on the student number of each school district, CDE (2013) classifies 178 sch ool districts into the four areas small rurual (fewer than 1,000), rural (1,001 to 6,500), suburban (6,501 to 25,000), and urban (more than 25,000).

PAGE 141

125 Meanwhile, seven organizations out of 12 organization s at the state level answered this survey A m ong six organizations that did not respond to the survey, two organizations rejected answering this survey while four organizations did not send any responses to the author Based on the final responses to the survey, t he final response rate shows that about 7 2 % of 178 school districts and approximately 5 8 % of 1 2 organizations responded t o the survey. Secondary Data Regarding the data collection for eight independent variables relevant to school he author searched multiple websites and finally found three websites providing s econdary data that support measur ing eight independent variables They are CDE, USA.COM, and EdNewsColorado websites. Among them, the CDE website offers the most secondary data by covering five independent variables ethnicity SC), fiscal situation (FISN), m arket based Table IV. 3 neatly arranges and draws how to measure each independent variable and what their data sources are All of the survey question s were completed on the basis of the literature review. For instance, t he author writes down a survey question related to the policy entrepreneur hypothesis in T able IV. 3 When making a survey question, it is very important to exactly define a core word related to a variable because it helps reduce the risk of receiving answers that do not fit a survey question ( Remler & Van Ryzin, 2011) A defining a word has a similar effect in making a concise survey question.

PAGE 142

Table IV. 3 Measurement of All Independent Variables I.V.s Data Sources Measurement / Operationalization Main Explanatory Variables 1. MIDF CDE The rat e of the number of neighboring school districts already having charter schools divided by the number of total neighboring school districts. 2. CODF FindTheData The rate of state supported financ ial aid in the whole budget of each school district. 3. POEN Survey The number of educational entrepreneurs in each school district. Teske & Williamson (2006) define l 4. POIG Survey The L ikert scale indicating the percentage of teachers in each school district who participate in the Colorado Education Association (CEA) or the American Federa tion of Teachers (AFT). 5. PONW Survey The degree of network density among a school district and the seven organizations at the state level. Other Explanatory Variables 6. ETHN CDE The log of minority students number of each school district. 7. LOCA USA.COM The number of people per mile of each school district. 8. EDSC CDE The ratio of the number of total students divided by the number of total teachers in each school district. 9. FISN CDE The log of annual average teacher salary for each school district. 10. MBET CDE The number of private schools in each school district. 11. STPE CDE The percentage of student dropout rate for each school district. 12. REEL EdNewsColorado The percentage of residents with in each school district. 13. REIN USA.COM The log of average per capita income in each school district. Note: CDE = The Colorado Department of Education; FindTheData, USA.COM, and EdNewsColorado are website names. MIDF=mimetic diffusion; CODF=coercive diffusion; POEN=policy entrepreneurs; POIG=policy interest groups; PONW=policy network s ; ETHN=ethnicity; =market based education tool; STPE 126

PAGE 143

127 Data Analysis The author utilizes two analytic tools UCINET and SPSS to conduct the social network analysis (SNA) and the multiple OLS regression analysis. T heir latest version is UCINET 6 and IBM SPSS 21 respectively. Borgatti, Everett, and Freeman (2002) point out that UCINET 6 is the newest and most general among SNA programs, and George and Mallery ( 2013) emphasize that IBM SPSS 21 is the most convenient and popular program among several statistical programs used in social sciences. UCINET 6 Firs t of all, UCINET 6 is employed to measure degrees of the policy network independent variable. As previously explained, th is research focuses on the concept of ne obtained from survey responses. The survey asked a superintendent in each school district and a representative in the organizations to score their connections relevant to the school choice movement For example, in the first survey whose respondent is a superintendent, t he main question of the survey ask ed a respondent to choose an organization that shares information for the school c hoice movement with her/his school district. And, a respondent w as asked to mark a number from 1 to 5 to any organizations that have connections to share information relevant to school choice tools O therwise a respondent was asked to mark a number zero ( see the appendix). After that, for data manipulation, the author scored a case that receives a number from 1 to 5 with number 1 while getting a case that receives number 0 unchanged. This data manipulation produces a network density matrix drawn in Figure IV.1 and Figure IV.2 Finally, the author

PAGE 144

128 created 126 network density matrices, using valid survey responses from 126 school districts and seven organizations. UCINET 6 calculated the value of each matrix and obtained degrees of network density among each school district and seven organizations. The final analyzed results indicate which school district forms the strong network density with seven organizations and which school district has weak network density with them. For instance, Figure IV.1 and IV.2 de scribe examples showing the matrix with the strongest and weakest network density respectively. Figure IV. 1 indicates that the Mesa County school district has 0.839 of the strongest network density while Figure IV. 2 says that the Byers 32J school district has 0.625 of the weakest network density. With these examples, we can assume that the Mesa County school district is more likely to implement the state charter school law due to their strong network density, compared to the Byers 32J school district.

PAGE 145

BEST C O EA CCSI CDE D CCC M CLCS ELC O MESA BEST 0 1 1 1 1 0 1 C O EA 1 0 1 1 0 1 1 CCSI 1 1 1 1 1 0 0 CDE D 1 1 1 1 1 1 1 CCC M 1 1 1 1 1 1 1 CLCS 1 1 1 1 1 0 0 ELC O 1 1 0 1 1 1 1 MESA 1 1 1 1 1 1 1 Note : BEST Board (BEST), Colorado Education Association (COEA), Colorado Charter School Institute (CCSI), Colorado Department of Colorado League of Charter Schools (CLCS), Education Leadership Council (ELC O), and Mesa County school district (MESA). Figure IV.1 A S ample of the M atrix for the Strongest Density of P olicy N etwork s BEST C O EA CCSI CDE D CCC M CLCS ELC O BYER BEST 0 1 1 1 1 0 0 C O EA 1 0 1 1 0 1 0 CCSI 1 1 1 1 1 0 0 CDE D 1 1 1 1 1 1 0 CCC M 1 1 1 1 1 1 0 CLCS 1 1 1 1 1 0 0 ELC O 1 1 0 1 1 1 0 BYER 0 0 0 0 0 0 0 Note : BEST Board (BEST), Colorado Education Association (COEA), Colorado Charter School Institute (CCSI), Colorado Department of Colorado League of Charter Schools (CLCS), and Education Leadership Council (ELCO), and Byers 32J S chool D istrict ( BYER ) Figure IV.2 A S ample of the M atrix for the Weakest Density of P olicy N etwork s 129

PAGE 146

130 M ultivariate A nalysis jurisdictions. To test the 13 hypotheses that could be answer s for the research question, this research employs the multiple ordinary least square s (OLS) regression analysis as a statistical technique. Employing the multiple OLS regression analysis first o f all means that the dependent variable, the uneven implementation of the state charter school policy, has a continuous variable characteristic. In the case that the dependent variable is a continuous variable and other independent variables are continuous or dummy variables, several scholars (Gujarati, 2003; Leech Barrett, & Morgan 20 11 ; Remler & Van Ryzin, 2011) indicate that the multiple OLS regression analysis is useful for predicting a relationship between the dependent variable and one independent v ariable when other independent variables are held constant. Two Equation Models Meanwhile, to obtain more accurate final results, this study uses two continuous variables to estimate the dependent variable. The first dependent variable is estimated by per centage of charter schools among whole K 12 public schools in each school district while the second dependent variable is measured by the raw number of charter schools in each school district. However, 13 independent variables explaining both dependent var iables are the same. The author calls them Model A and Model B respectively. Many scholars in the public administration and policy area primarily utilize the two ways in measuring a degree of policy implementation when they study performance as part of

PAGE 147

131 pol icy implementation. 46 On the basis of the two dependent variables, this dissertation creates the two equation models variable is the percentage of charter schools among K 12 public schools, and Model B has th e dependent variable measured by the raw number of charter schools in each school district. Model A and Model B have their own equation model as follows: among K 12 pub lic schools in each school district in each school district (Note: U ICSP = Uneven Implementation of the charter school policy; MIDF=mimetic diffusion; CODF=coercive diffusion; POEN=policy entrepreneurs; POIG=policy interest groups; PONW=policy network s MBET =mar ket ) 46 Public administration and public management conferences usually have performance sessions where schola rs present their performance topics related to policy outputs or outcomes. This means that policy implementation covers both policy outputs and outcomes as well. In the same vein, Stewart et al. (2008) indicate that scholars in the public policy area broad ly use the definition of policy implementation including a process, an output, and an outcome. The term process means that a lower government makes a law or policy corresponding with an upper law. The term output implies the extent to which a government su pports a new policy to accomplish its intents and goals. The term outcome denotes the extent to which a new policy virtually completes its goals. Therefore, in this study, the term policy implementation outcome ecause the uneven implementation of the state charter school policy, which is the dependent variable in this dissertation, indicates how well (actively) or poorly each school district complies with the state charter school law.

PAGE 148

132 In the two equation models, is the intercept (constant), which denotes the value of the dependent variable when all ind 2011). is the error term known as residual or remainder, which appears when an equation model does not exactly represent the actual assoc i ations between a dependent variable and indepen dent variables (Gujarati, 2003). T he error term indicates the total values in the equation when the multiple OLS regression analysis is conducted. Thus, zero in the err or term means that there is not a difference between the equation model and real world application. A relationship between a dependent variable and an independent variable will be explained by interpreting the coefficient of the indep en dent variable when other independent variables are held constant. These analyzed results are specifically shown and interpreted in Chapter 5 with descriptive results.

PAGE 149

133 CHAPTER V FINDINGS AND RESULTS This chapter covers the main statistical analysis tools used in this dissertation and accounts for their analyzed results. To conduct an empirical study orien ted to quantitative methods, this dissertation focuses on descriptive statistics, multicollinearity diagnosis tools, and the multiple OLS regression analysis. As mentioned in the previous chapters, between the two dependent variables and 13 independent variables, which compose Model A and Model B. In this c hapter, the descriptive statistics and multicollinearity parts handle both models together while the multiple OLS regression analysis part Descriptive Statistics The main function of d escriptive statistics is to offer researchers general contents and information of each variable in data (Babbie et al., 2013; Leech et al., 20 11 ; Morgan et al. 2013; Wagner, 2013). Usually, descriptive statistics are conducted to know about approximate ch aracteristics of each variable prior to operating inferential statistics Descriptive statistics possess several analysis tools number of observations, range (minimum and maximum), mean, median, mode etc. and multiple graphical skills ba r, line, pie, 3 di mension bar etc. to express their analyzed results (Babbie et al., 2013; Leech et al., 2011 ; et al., 2011 ). Among several descriptive statistical tools, th is research picks and utilizes four main descriptive tools the number of valid cases cal led observations, maximum, minimum, and mean. Table V. 1 describes their analyzed results for all of the variables,

PAGE 150

134 which are involved in the two final equation models. Meanwhile, in the dissertation, lained in company with contents of its codebook ( questionnaire ). T his collaborative interpretation trial leads researchers to have more detailed descriptions and information in looking at First of all, the first two variables shown in Table V. 1 are the dependent variables UICSP 1 and UICSP 2 which are employed to estimate variation of each the UICSP 1 dependent variable report that the range of the local charter school policy implementation percentage in Colorado is from 0 to 40%. Forty percent means that schools of 4 0% among total K 12 public schools in a school district are charter schools. The dataset shows that the South Park RE 2 school district in Park County has the value of 40%. O ther data relevant to this dependent variable indicate that two schools among its total five K 12 public schools are charter schools. Meanwhi le, the other dependent variable UICSP 2 draws that the range of charter school numbers in school districts is from 0 to 31. The dataset indicates that there are 31 charter schools in Denver school district. Thus, it is the Denver school distric t that has the maximum value of 31. Table V.1 informs that the valid cases ( observations or samples) of both dependent variables are 178. That is to say, all of the 178 observations in both dependent variables have no missing data. Regarding the descript ive statistics of 13 independent variables, th is research first looks at the descriptive results of five independent variables related to the main theoretical approaches policy diffusion, policy entrepreneur, policy interest group, and

PAGE 151

135 policy network model s The mimetic diffusion (MIDF) variable of the first independent variable is measured by the diffusion rate, which is calculated by dividing the number of neighboring school districts that have previously had charter schools by the number of total neighbo ring school districts. 47 The mean of the mimetic diffusion variable is about 0.06. Its range is from 0 to 1. The dataset shows that the school district with 1 of its maximum value is the Pueblo City 60 school district This means that a neighboring school district has already operated charter schools before a charter school appeared in the Pueblo City 60 school district The Pueblo City 60 school district has only one neighboring school district, the Pueblo County Rural 70 school district. A map that CDE distributes shows that the Pueblo County Rural 70 school district surrounds the Pueblo City 60 school district. 48 Therefore, the mimetic diffusion rate of 1 illustrates that the Pueblo County 70 school district ha s already operated a charter school before a charter school is established in the Pueblo City 60 school district The descriptive results of the mimetic diffusion variable are obtained by calculating 178 valid observations. Based on 178 valid observations, t he descriptive statistics of the coercive diffusion (CODF) variable, which is operationalized by the rate of state supported financial aid in whole budget of a school district, describe that its range is from 0.03 to 0.93. Its maximum value of 0.93 explains that a financial support from a state co mprises approximately 93% of the whole budget of a school occupies about 93% of the 47 A neighboring school distr ict means an adjoining school district sharing a border line with an observed school district. 48 Pueblo county has two school districts the Pueblo City 60 school district and Pueblo County Rur a l 70 school district in its territory.

PAGE 152

136 The descriptive statistics of the policy entrepreneur (POEN) va riable indicates that 107 school districts provided valid answers to this survey question. Th e mean of educational entrepreneur numbers obtained from 1 07 valid observations is 1.81. The dataset displays that the Steamboat school district is the school distric t that ha s the most educational entrepreneurs. This school district has 30 educational entrepreneurs. The policy interest group (POIG) variable is operationalized by the 5 Likert scale indicating the percentage of teachers in each school district wh o participate in the Colorado Education Association or the American Federation of Teachers (AFT). 49 119 respondents gave valid responses to this survey question. The dataset shows that among 119 valid observations, eight school districts are under the fifth category. Its mean is 2.26. The policy network (PONW) variable is estimated by the density of policy networks, whose dataset was completed by the cooperation of both 12 6 school districts and 7 public and non profit organizations. Its mean is 0.69435 and i ts range is from 0.625 to 0.839. The dataset reveals that the two school districts with the highest score of the policy network density are the Mes a County Valley 51school district and the Pueblo City 60 school district Both school districts receive 0.839 school districts are given 0.625 of the lowest score of the policy network density. On the other hand, there are eight independent variables that describe and explain school district characteristics in the disse rtation. First of all, the descriptive statistics of two independent variables ethnicity (ETHN) and location (LOCA) related to the demographic factors show that all of the 178 school districts have valid info rmation about 49 In the original survey, the Likert scale for this POIG variable is composed of 10 categories as follows: 1) 0 10% 2) 11 20%, 3) 21 30%, 4) 31 40%, 5) 41 50%, 6) 51 60%, 7) 61 70%, 8) 71 80%, 9) 81 90%, 10) 91 100%. The author manipulates this data by redu cing this10 Likert scale to 5 Likert scale as follows: 1) 0 20%, 2) 21 40%, 3) 41 60%, 4) 61 80%, 5) 81 100%.

PAGE 153

137 two variables. The ethnicity (ETHN ) variable is measured by the number of minority students. The mean of the minority students of 178 school districts is approximately 2 076. Its range is 0 to 64 384. The maximum value of 64 384 belongs to the Denver school district. It denotes that the De nver school district among 178 school districts has the most minority students. Meanwhile, the dataset informs that Agate School District 300 does not have any minority students. On the basis of 178 valid observations, t he descriptive statistics of the loc ation (LOCA) variable, which is estimated by population density, show that its range is from 0.2 to 4746.2. The dataset displays that the Englewood school district has the most population density of 4746.2 while two school districts Branson Reorganized S ch ool D istrict No. 82 and Kim Reorganized S chool D istrict No. 88 have the lowest population density of 0.2. This dissertation utilizes four independent variables regarding the school related factors. All of these four independent variables have 178 valid obser va tions. First, the is measured T on average, one teacher teaches nearly 14 students across Colorado. This variable s range is from 5.3 to 33.4. The dataset indicates that a teacher in Kim Reorganized School District No. 88 teaches the fewest students ( about 5 ) while a teacher in Julesburg School District No. Re 1 covers the most students ( about 33 ) in Colorado Secondly, the descriptive statistics of the fiscal situation (FISN) variable show that the average salary of the Cherry Creek school district is the highest one of $62,086 while the average

PAGE 154

138 Thirdly, the descriptive statistics of the market based education tool (MBET) variable indicate that on average there are about two private schools across Colorado and the Denver school district among 178 school districts has the most private schools of 43. is operationalized by the percentage of student dropouts from each school district. Its mean is 1.6% and its range is from 0 to 14.6%. The dataset shows that the dropout percentage of 14.6%. The final independent variable category embraces socioeconomic factors. Th e approximately 24% of residents possess a bachelor or higher degrees across Colorado The range of this independent variable is from 7.3% to 62.2%. The dataset reveal s that 62.2% of residents over 25 year old living in the Boulder Valley school district earn a bachelor or higher degrees while 7.3% of adults living in Adams County School District 14 have a bachelor or higher degrees come (REIN) variable describes that its mean is approximately $25,519 and its range is from $13,261 to $70,665. The dataset points out that residents living in Aspen School District No. 1 have the highest average per capita income of $70,665 while resident s living in Granada School District RE 1 have the lowest average per capita income of $13,261.

PAGE 155

Table V.1 Descriptive Results of All Variables Variables N Minimum Maximum Mean UICSP1 (D.V. 1) 178 0 40 3.583 UICSP2 (D.V. 2) 178 0 31 .86 MIDF (I.V. 1) 178 0 1 .0633 CODF (I.V. 2) 178 .03 .93 .5028 POEN (I.V. 3) 107 0 30 1.81 POIG (I.V. 4) 119 1 5 2.26 PONW (I.V. 5) 126 .625 .839 .69435 ETHN (I.V. 6) 178 0 64384 2075.88 LOCA (I.V. 7) 178 .2 4746.2 266.787 ED SC (I.V. 8) 178 5.3 33.4 14.008 FISN (I.V. 9) 178 26598 62086 40384.61 MBET (I.V. 10) 178 0 43 1.56 STPE (I.V. 11) 178 0 14.6 1.564 REEL (I.V. 12) 178 7.3 62.2 24.356 REIN (I.V. 13) 178 13261 70665 25519.16 Note: Dependent variables: UICSP, the acronym of the dependent variable, stands for the uneven implementation of the charter school policy. UICSP1 is measured by the percentage of charter schools among K 12 public schools in each school district while UICS P 2 is estimated by the raw number of charter schools in each school district. I ndepend en t variable s: MIDF=mimetic diffusion; CODF=coercive diffusion; POEN=policy entrepreneurs; POIG=policy interest groups; PONW=policy network s ; =market 139

PAGE 156

140 Multicollinearity After interpreting the desc riptive statistics of each variable, the next step looks at if there are multicollinearity issues among independent variables to obtain more precise answers from the multiple OLS regression analysis. The term multicollinearity means that there is a high de gree of correlation among two or more independent variables involved in multivariate analysis me thods (Hair Black, Babin, & Anderson 2010; Lee & Kim, 2010; Leech et al., 2011; Re lmer & Van Ryzin, 2011 ; Wheelan, 2011 ). M ulticollin e arity denotes that two or more independent variables in the same multivariate regression analysis have the same or overlapping information or characteristics. Thus, matc hes with other independent variable information (Relmer & Van Ryzin, 2011). In such a case, the independent variable does not play a role of offering researchers any unique information that other independent variables in the same multivariate regression an alysis do not have. Thus, s cholars ( Gujarati, 2003; Hair et al., 2010; Lee & Kim, 2010) emphasize that it is very hard to estimate exact coefficients of all independent variables with a multicollinearity problem. Eventually, it is the best way to reduce al l independent variables with the same information to obtain more accurate results from the multivariate regression analysis. One pivotal assumption of the multiple OLS regression analysis is that independent variables entered into the multiple OLS r egression analysis are not completely correlated with other independent variable s (Gujarati, 2003; Leech et al., 20 11 ; Lee & Jeong, 2012; Remler & Van Ryzin, 2011). If there is an independent variable that is highly correlated with other independent variab les, each coefficient has

PAGE 157

141 higher standard errors and the overall fit of the equation model might not be precise ( ; Hair et al, 2010, Remler & Van Ryzin, 2011). Namel y the final statistical results and interpretations become inaccurate if high intercorrelations among independent variables exist (Bord et al., 2000 ; Lee & Kim, 2010) Therefore, it is very important to detect if there is a multicoll i nearity problem among independent variables with a multicollinearity diagnosis to ol before researchers conduct several multivariate analysis methods. In diagnosing a mu l ticollinearity problem scholars (Hair et al., 2010; Lee & Jeong, 2012; Leech et al., 2011) mainly use the following three analysis tools correlation tolerance level, and variance inflation factor (VIF). Therefore, we first check if there are high correlations among all pairs of independent variables. If the ir correlation coefficients are higher than 0. 5 w e need to consider that a multicollinearity issue might exist i n a pair of independent variables because it means that there is a high possibility that the two independent variables have the same information (Lee & Jeong, 2012; Leech et al., 20 11 ). And then, we carefully test a multicollinearity of the independent variables having high correlation coefficients with the two analysis tools VIF and tolerance level. Following this procedure of the multicollinearity diagnosis, this study checked multicollinearity is sues of all of the independent variable s included in the two final equation model s A fter analyzing correlation coefficients, tolerance levels, and VIFs of all of the independent variables i n the two final equation model s this study brings and summarizes their analyzed results into Table V.2 50 Table V.2 synthetically draws the 50 This dissertation has the two final equation models Model A and Model B. Both equation models have the same independent variables although each equation model has the different dependent variable. Thus, this study conducted the multicoll i nearity diagnosis for the same 13 independent variables in both equation models one time.

PAGE 158

142 results of the three statistical f unctions that are used to check an existence of multicoll i nearity among 13 independent vari a bles in Model A and Model B. C orrelation C oefficients We first need to check a correlation of a pair of independent variables to know if an independent variable in an equation model has a multicollinearity issue with other independent variables by detecting correlation coefficients of all pair s of independent variables. As previously mentioned, a scholar has to doubt the possibility of a multicoll i nearity issue if the value of the correlation coeffic i ent of a pair of independent variables is higher than 0.5 (Lee & Jeong, 2012; Lee & Kim, 2010 ; Leech et al., 20 11 ). We consider that the two independent variables have the conceptually same (at least, similar) information if they have high (large) correlation coefficients (Leech et al., 20 11 ; Remler & Van Ryzin, 2011 ). 51 The analyzed results show that among 78 to t al independent variable pair cases in th e two e quation model s the independent variables under 11 correl a tion cases mig h t have the risk of a multicollinearity issue because coefficients are higher than 0.5. These results inform that th is research must be careful that the independent variables under 11 pairs of independent vari a bles might have a multicollinearity issue. However, scholars (Lee & Jeong, 2012; Leech et al., 20 11 ) highlight that a pair of independent variables having a correlation coefficient that is higher than 0.5 do es not always mean to have a multicollinearity issue. Namely, t he role of this correlation coefficient in detecting a mulitcollinearity issue just gives researchers a notice o r warning that a pairwise independent variable case that has a correlation 51 Scholars (Lee & Jeong, 2012; Leech et al., 2011) usually regard tha t a pair of independent variables has a high correlation if a correlation coefficient is higher than 0.5. But, it is not clear cut.

PAGE 159

143 coefficient with higher than 0.5 might have a multicollinearity issue. Th erefore we need to check VIF or tolerance level of each independent variable to know if an independent vari able highly correlated to another one meets a mul t i coll i stresses that both VIF and tolerance are widely applied statistical tools to estimate degrees of multicollinearity of an independent variable with the other ones in a mu ltivariate regression analysis. V ariance I nflation F actor (VIF) and T olerance L evel VIF assesses how much the variance of the estimated coefficients is increased ; Stine, 1995 ). If an i ndependent variable in a multiple regression analysis does not have any relations hip s with other independent variables, the value of its VIF becomes 1. The value of VIF is obtained from 1/Tolerance (Leech et al., 20 11 52 Thus, each of their results is related. If an independent variable has a value of VIFgreater than 10 and a value of tolerance smaller than 0.1, this result supports that th is independent variable ha s a serious mulitcollinearity issue with other independen t variables (Bord et al., 2000; 53 If there is a serious multicollin e arity issue, Leech and 52 where denotes the multiple for the regression of the targeted j th independent variable on the other independent variables in the multiple regression analysis. Meanwhile, 53 Scholars say that if an independent variable has a high value of VIF and a low value of tolerance, this independent variable might meet a multicollinearity issue. However, there is not a clear cut rule for the value of VIF and tolerance. Scholars do not decide what a large (or small) VIF and a small (or large) tolerance are. Namely, scholars have a different standard for VIF and tolerance. Some scholars (Lee & Jeong, 2012; Leech et al., 20 11 ) argue that when a value of VIF is greater than 5 and a value of tolerance is smaller than 0. 5 it is possible for a serious multicollinearity issue to exist. In the dissertation, the author follows Bord et al . T h ey say that t here is no multicollinearity issue if the value of VIF is smaller than 10 and the value of tolerance is higher than 0.1.

PAGE 160

144 her colleagues ( 20 11 ) suggest dropping one independent variable among independent variables with the same information. The values of e V.2 These results reveal that there are no independent variables with a severe smaller than 10 and all of the toleranc e higher than 0.1. In another words, the same 13 independent variables that consists of Model A and Model B are in the ab sence of complete (full or p e r fect) multicollinearity. Therefore, the final results of these two multicollinearity diagnosis tools VIF and tolerance concretely supports that all of the independent variables in the two final equation models do not have any multicollinearity issues. Based on these multicollinearity analysis results, the final results obtained by the multiple OLS regression analysis will be more accurate answers because any pairs of independent variables do not have a pe r fe c t (exact) linear relation among the same 13 independent variables in each equation model 54 54 1 or 1, the two independent variables have a perfect (exact) linear rel ation (or relationship).

PAGE 161

Table V.2 Analyzed R esults for Multicollinearity of All Independent Variables MIDF CODF POEN POIG PONW ETHN LOCA ED S C FISN MBET STPE REEL REIN Tolerance VIF MIDF 1.000 .65 .05 0 .242*** .488*** .447*** .341*** .351*** .346*** .249*** .156** .15 7 ** .161** .540 1.851 CODF 1.000 .080 .149 .277*** .137* .035 .026 .373*** .161** .106 .480*** .52*** .559 1.787 POEN 1.000 .271*** .190* .303*** .312*** .231** .36*** .374*** .108 .178* .058 .701 1.426 POIG 1.000 .379*** .563*** .280*** .441*** .491*** .377*** .069 .124 .018 .531 1.883 PONW 1.000 .52*** .294*** .307*** .506*** .433*** .269*** .359*** .264*** .483 2.069 ETHN 1.000 .529*** .754*** .713*** .56*** .232*** .258*** .108 .239 4.191 LOCA 1.000 .268*** .507*** .46*** .226*** .097 .035 .607 1.647 ED S C 1.00 .542*** .279*** .178** .132* .058 .457 2.189 FISN 1.00 .474*** .138* .484*** .445*** .272 3.672 MBET 1.00 .096 .332*** .241*** .432 2.315 STPE 1.00 .165** 194*** .777 1.287 REEL 1.00 76*** .330 3.032 REIN 1.00 .299 3.348 No te: * significant at .01 level (2 tailed) ; * significant at .05 level (2 tailed); significant at .1 level (2 tailed) I.V.s: MIDF=mimetic diffusion; CODF=coercive diffusion; POEN=policy entrepreneurs; POIG=policy interest groups; PONW=policy network s ; =market 145

PAGE 162

146 Multiple OLS Regression Analysis The next step of statistical analysis explicates the two equation models Model A and Model B with the multiple OLS regression analysis among generalized linear modelling techniques. The main function of t he multiple OLS regression analysis is to explain or predict a variation of a dependent variable, utilizing two or more independent variables (Leech et al., 20 11 ; Remler & Van Ryzin, 2011; Wagner, 2013). It demonstrates if associations between one dependent variable and several independent variabl es have positive or negative directions as well ( Morgan et al., 2013) In the multiple OLS regression analysis, a dependent variable must be a continuous variable style indicating a ratio or interval variable while it is good for independent variables to b e either continuous variables ratio or interval variables or dummy variables among categorical variables (Babbie et al., 2013; Wagner, 2013). In the equation models below, the two dependent variables UICSP 1 and UICSP 2 are continuous variables. 13 other i ndependent variables are continuous variables as well. Therefore, the multiple OLS regression analysis is a proper statistical technique to analyze the following two equation models: UICSP 1 or UICSP 2 = (Note: U ICSP = uneven Implementation of the charter school policy; MIDF=mimetic diffusion; CODF=coercive diffusion; POEN=policy entrepreneurs; POIG=policy interest groups; PONW=policy network s ation; MBET =market )

PAGE 163

147 The final results of the two equation models are drawn in Table V.3 and Table V.4 respectively. In treating or manipulating the data for each variable, th is research has logged the data of the three independent variables before operating the multiple OLS regression analysis to help its final results bear more statistically significant independent variables. The log transformed ind ependent variables in the two final equation model s are variables. Th is dissertation fixes the standard level of significance at the 0.1 level. Thus, the term statisti cally significant or statistical significance in th is dissertation is used when a probability value, which is usually called p value, is less than 0.1. Usually, scholars in social sciences employ three p values 0.1, 0.05, and 0.01 levels on the basis of a critical value (Leech et al., 20 11 ) The critical value is named the alpha level as well and plays a role as a standard or cutoff value determining if a scholar accepts or rejects a null hypothesis. That is to s ay, the critical value is a border line to determine if the likelihood of Type I error the likelihood of rejecting an actually true hypothesis occurs (Morgan et al., 2013). In the same vein, if we choose the 0.1 level as a critical value, we express that t his is more liberal in rejecting a null hypothesis. Meanwhile, if we select the 0.01 level as a critical value, we say that this is more conservative in rejecting a null hypothesis ( Leech et al., 20 11 ; Morgan et al., 2013). Thus, the analyzed results in th is dissertation are more liberal in rejecting a null hypothesis shown in all cases. For instance, in interpreting the slope coefficient of an independent variable in the multiple OLS regression analysis, an independent variable with p value that is less th an 0.1 will be explained as follows: its null hypothesis is treated as an unlikely hypothesis and its

PAGE 164

148 alternative hypothesis is regarded as a likely hypothesis because its p value is less than 0.1 level of the critical value. Concisely speaking, its null h ypothesis is rejected and its alternative hypothesis is accepted at the 0.1 level. Table V.3 and Table V.4 respectively. First of all, both analyzed results present that there are 100 valid samples (cases) while there are 78 missing ( i nval i d) samples in running the multiple OLS regression analysis. This is why as described in the descriptive statistics, there are four independent variables policy entrepreneur (POEN), policy interest group (POIG), policy netw ork (PONW), and ethnicity (ETHN) with missing samples although the two dependent variables and other nine independent variables completely have 178 valid samples. 55 Among four independent vari a bles with missing samples, the policy entrepreneur (POEN) variab le first has 71 missing samples, the policy interest group (POIG) variable has 59 missing samples, policy network (PONW) variable has 52 missing samples, and the ethnicity (ETHN) variable has one missing sample respectively The final 78 missing samples oc cur by the combination of missing samples in the four aforementioned independent variables. Meanwhile, the dataset also confirms that the final results of the multiple OLS regression analysis are obtained by analyzing 100 valid observations. lyzed Results Table V.3 results obtaine d by the multiple OLS regression analysis show that its result of F statistic ( F =5.886, 55 In ETHN independent variable, Agate S chool D istrict 300 does not have any minority students. When the author logged 0 of its minority students, its log transformed form is log 0. The value of log0 is undefined. Th erefore the author treated the log transformed minority students of Agate S chool D istrict 300 as a missing value.

PAGE 165

149 d.f.=13, 87) is statistically significant at 0.01 level. Therefore, we can reject its null hypothesis that there is no relationship between the dependent variable and the set of 13 independent variables. This F statistic result means that the overall regression model fits the data well, and 13 independent variables significantly combine together to explain or predict the dependent variable. That is to say, th e combination of 13 independent variables significantly explains or predicts the dependent variable the variation of licy implementation. Meanwhile, the adjusted R squared value of Model A is .388, which indicates that 13 independent variables in Model A account for about 39% of the variation in the UICSP 1 dependent variable. Among 13 independent variables in Model A the three independent variables are statistically significant at the different levels 0.1 level and 0.01 level while the rest of the independent variables are not statistically significant The three statistically significant independent variables are the mimetic diffusion (MIDF), policy entrepreneur (POEN), and policy network (PONW) independent variables. The first two independent variables are statistically significant at the 0.1 level and the policy network (PONW) variable statistically significant at th e 0.01 level. In interpreting the analyzed multiple a statistically significant independent variable accounts for a dependent variable, a schol a r generally focuses on explaining unstandardized c oefficients (or slopes) of statisti cally significant independent v ariables w ith all other independent variables held constant (Remler & Van Ryzin, 2011). First of all, the result of the mimetic diffusion (MIDF) independent variable supports that its null hypothesis is rejected at the 0.1 le vel. That is, we can reject the null

PAGE 166

150 hypothesis that the UICSP 1 dependent variable is not associated with the MIDF independent variable as follows: Reject the MIDF variable ull hypothesis: School district s mimetic diffusion factor is not related to the uneven implementation of Colorado s local charter school policy Accept School districts with many neighboring school districts that have already implemented the state charter school policy are more likely to implement the state charter school policy. coefficient means that one unit increase of the MIDF independent variable results in an expected increase in 7.715 of the UICSP 1 depe ndent variable w ith all other independent variables held constant. This interpretation indicates that there is a positive relationship between the MIDF independent variable and the UICSP 1 dependent variable as well Broadly speaking, we can interpret that a likelihood that a school district is more likely to implement the state charter school policy is increased when there are many neighboring school districts that have already implemented the state charter school policy. The policy entrepreneur (POEN) v ariable is statistically significant at the 0.1 level as well. Thus, we can reject the null hypothesis that the UICSP 1 dependent variable is not associated with the POEN independent variable as follows: Reject the POEN variable ull hypothesis : School d istrict s policy entrepreneur factor is not related to the uneven implementation of Colorado s local charter school policy. Accept S chool districts having many educational entrepreneurs are more likely to implement the state charter school policy. As predicted, its direction is positive to the dependent variable. Its coefficient for the UICSP 1 dependent variable is 0.25, which indicates that the im plementation of

PAGE 167

151 Colorado school district charter school policy increases by about 0.25 for one unit increase of educational entrepreneurs w ith all other independent variables held constant. Briefly speaking, a school district with more policy entrepreneu rs in its territory is more likely to implement the state charter school policy than a school district with fewer policy entrepreneurs Finally, the policy network (PONW) independent variable is statistically significant at the 0.01 level. Thus, the resul t of the PONW independent variable supports that its null hypothesis is rejected at the 0.01 level. That is, we can reject the null hypothesis that the UICSP 1 dependent variable is not associated with the PONW independent variable as follows: Reject th e PONW variable ull hypothesis : School district s network density is not related to the uneven implementation of Colorado s local charter school policy. Accept School districts having a strong density of policy networks between their own school district and other organizations, which are related to public educational innovation and reform, are more likely to implement the state charter school policy. This indep 38.945 for each unit increase variable w ith all other independ ent variables held constant. Its direction shows that there is the positive relationship between this independent variable and the UICSP 1 dependent variable. That is to say, a school district with a stronger policy network density is more likely to implem ent the state charter school po l i c y than a school district with a weaker policy network density.

PAGE 168

152 Based on the standardized coefficients of the statistically significant independent variables in Model A, the results reveal that the policy network (PONE) ind ependent variable has the strongest explanatory impact in accounting for the variation in the significant independent variables. The independent variable having the second strongest explanatory impact for the dependent variable is the mimetic diffusion (MIDF). The policy entrepreneur (POEN) variable has the third strongest explanatory impact for the dependent variable 56 56 The main function of standardized coefficients, which is called beta weights, is to inform us how powerfully an independent variable accounts fo r the deviation of a dependent variable. That is to say, an independent variable with the largest standardized coefficient is the most powerful (important) independent variable in explaining the deviation of a dependent variable compared to the other indep endent variables in multiple OLS regression analysis ( Lee & Jeong, 2012; Leech et al., 2011; Remler & Van Ryzin, 2011).

PAGE 169

153 Table V.3 Multiple OLS Regression Analysis Results Unstandardized Coefficients Standardized Coefficients B S.E. Beta t Sig. 1. MIDF* 7.715 4.153 .198 1.858 .067 2. CODF 1.488 3.616 .043 .412 .682 3. POEN* .250 .132 .177 1.899 .061 4. POIG .919 .557 .177 1.651 .102 5. PONW*** 38.945 14.548 .301 2.677 .009 6. ETHN .516 .587 .141 .879 .382 7. LOCA .001 .001 .042 .420 .676 8. ED SC .065 .200 .038 .326 .746 9 FISN 8.850 7.964 .167 1.111 .270 10. MBET .436 .442 .117 .987 .326 11. STPE .010 .346 .003 .029 .977 12. REEL .107 .075 .196 1.436 .155 13. REIN 4.172 3.537 .169 1.18 .241 Constant 19.593 75.885 .258 .797 N 100 F statistic (13, 87 ) 5.886*** Adjusted R square .388*** Note: *** significant at .01 level; ** significant at .05 level; significant at .1 level In Model A, D V the percentage of charter schools among K 12 public schools in each school district I.V.s: MIDF=mimetic diffusion; CODF=coercive diffusion; POEN=policy entrepreneurs; POIG=policy interest groups; PONW=policy network s capacity; FISN=fiscal situation; MBET =market income.

PAGE 170

154 Model B is constructed to analyze relationships between the UICSP 2 dependent variable which is estimated by the raw number of charter schools in each school district and 13 independent variables The results of Model B s multiple OLS regression analysis are summarized in T able V4 Like Model A, there are 100 valid observations used to analyze Model B. Its result of F statistic ( F =21.504, d.f.=13, 87) is s tatistical ly significant at the 0.01 level. This statistical result helps us reject its null hypothesis that the dependent variable does not have a relationship with the set of 13 independent variables. This means that th e combination of 13 independent variables sig nificantly predicts the uneven policy well Meanwhile, the value of adjusted R 2 is about 0.7 3 This indicates that approximately 7 3 % of the vari ation in the UICSP 2 dependent variable is accounted for by all of 13 independent variables that a mong 13 independent variables there are seven independent variables that are statistically significant at different levels First of all, there a re four statistically significant independent variables the policy entrepreneur (POEN) variable, policy network (PONW) variable, ethnicity (ETHN) variable, and location (LOCA) va riable at the 0.01 level. The two independent variables the mimetic diffusion (MIDF) variable and market based education tool (MBET) variable are statistically significant at the

PAGE 171

155 0 .05 level. Finally, the fiscal situation (FISN) variable is statistically si gnificant at the 0.1 level. 57 Among four independent variables that are statistically significant at the 0.01 level, f irst, 0 039 of the slope for the policy entrepreneur (POEN) independent variable means that the implementation of Colorado school districts charter school policy increases by approximately 0.039 for each number increase of policy entrepreneurs w hen all other independent variables are held constant A school district with more policy entrepreneurs is more likely to implement the charter school policy than school districts with fewer policy entrepreneurs. S econd, the policy network (PON W ) independent me ans that the implementation of Colorado school districts charter school policy increases by 4.175 f or each unit increase of policy network density w hen all other independent variables are held constant This interpretation means that a school district with more dense policy networks is more likely to implement the state charter school policy with other indepe ndent variables held constant. In Model B, two other independent variables under the demographic factor category are statistically significant at the 0.01 level as well. First, the statistical result of the ethnicity (ETHN) independent variable supports t hat its null hypothesis is rejected at Reject School d istrict s minority students are not related to the uneven implementation of Colorado s local charter school policy. 57 In Model B, the processes rejecting the null hypotheses of the three statistically significant independent variables MIDF, POEN, and P ONW are skipped because their process contents are the same with the cases in Model A.

PAGE 172

156 Accept School districts having more minority students are more likely to implement the state c harter school policy. T the implementation of Colorado school 0.00191 for one percent increase of minority students w hen all other independent variables are held constant 58 This interpretation indicates that a school district with more minority students is more like ly to implement the state charter school policy. Finally, the analyzed result of the location (LOCA) independent va riable supports that its null hypothesis is rejected at the 0.01 level. That is, we can reject the null hypothesis that the UICSP 2 dependent variable is not related with the LOCA independent variable as follows: Reject School district s population density is not related to the uneven implementation of Colorado s local charter school policy. Accept School districts having a higher population density are more likely to implemen t the state charter school policy Its coefficient of 0.001 indicates that the implementation of Colorado school districts c harter school policy increases by 0.001 for every one unit of increase in population density w hen all other independent variable s are held constant This interpretation means that a school district with a higher population density rate is more likely to implement the state charter school policy. 58 A way to interpret an equation model of Y = a + b ln X + e with a level dependent variable and a log ged independent variable is as follows: T he change in Y is obtained by multiplying (or dividing) the coefficient of b by 0.01 (or 100) when a percentage in X increases For instance, if b is 500, the change in Y is obtained by multiplying 500 by 0.01. T h us, the change in Y becomes 5 for a percentage increase in X ( Guj arati, 2 003 p. 182 )

PAGE 173

157 There are the two independent variables that are statistically significant at the 0.05 means that the implementation of Colorado school districts charter school policy increases by 0.909 for one unit increase of mimetic diffusion rate w hen all other inde pendent variables are held constant Namely, this interpretation indicates that a school district with more neighboring school districts that have already carried out the state charter school policy is more likely to implement the state charter school poli cy. The analyzed result of the market based education tool (MBET) variable advocates that its null hypothesis is rejected at the 0.05 level. Namely, we can reject the null hypothesis that the UICSP 2 dependent variable is not related with the MBET independ ent variable as follows: Reject School district s private schools are not related to the uneven implementation of Colorado s local charter school policy. Accept School distri cts having more private schools are more likely to implement the state charter school policy. Its coefficient of 0.098 means that the implementation of Colorado school districts charter school policy increases by 0.098 for each number increase of private schools w hen all other independent variables are held constant Namely, there is the positive relationship between this MBET independent variable and the UICSP 2 dependent variable. This interpretation broadly supports that a school district with more pr ivate schools is more likely to implement the state charter school policy. There is one independent variable that is statistically significant at the 0.1 level. It i s the fiscal situation (FISN) independent variable. Thus, we can reject its null hypothesi s

PAGE 174

158 that the UICSP 2 dependent variable is not related with the FISN independent variable at the 0.1 level as follows: Reject School district s fiscal situation is not related to the uneven implementation of Colorado s lo cal charter school policy. Accept S chool districts having an insufficient and unhealthy fiscal situation are more likely to implement the state charter school policy. First of all, this analyzed result shows that the direction between the UICSP 2 dependent variable and this FISN independent variable is negative. The literature review propose s that this FISN independent variable might have a positive or a n egative direct ion with the dependent variable. This empirical result demonstrates that there is a negative relationship between this FISN independent variable and the UICSP 2 depen dent variable. I n the Colorado school district case, a school district with an insufficien t fiscal situation is more likely to deliver more charter school services to their residents. Meanwhile, its coefficient indicates 1. 624. Most carefully, like the interpretation of the ethnicity (ETHN) independent variable, interpreting this FISN independ 1 .624 of this FISN independent s coefficient is interpreted as follows: T he implementatio n of Colorado school districts charter school policy decreases by approximately 0.01624 for one percent increase of the average teacher salary w hen all other independent variables are held constant Namely, this result shows that a school district with a weak financial situation is more likely to implement the state charter school policy.

PAGE 175

159 Among the statistically significant seven independent variables in Model B, their standardized coefficients indicate that the location (LOCA) independent variable has the most explanatory impact in accounting for the dependent variable. Its standardized coefficient is 0.346 which is higher than the standardized coefficients of the other statistically significant independent variables. The ethnici ty (ETHN) independent v ariable ha s the second strongest explanatory power for the dependent variable. The policy network (PONW) independent variable has the third strongest explanatory power for the dependent variable. Meanwhile, the standardized coefficient of the mimetic diffu sion (MIDF) independent variable demonstrates that th is independent variable has the weakest explanatory power for the dependent variable.

PAGE 176

160 Table V.4 Multiple OLS Regression Analysis Results Unstandardized Coefficients Standardized Coefficients B S.E. Beta t Sig. 1. MIDF* .909 .447 .144 2.032 .045 2. CODF .417 .389 .075 1.071 .287 3. POEN* * .039 .014 .170 2.726 .008 4. POIG .023 .060 .027 .375 .708 5. PONW*** 4.175 1.567 .200 2.665 .009 6. ETHN*** .191 .063 .323 3.018 .003 7. LOCA*** .001 .000 .346 5.154 .000 8. ED SC .019 .022 .067 .866 .389 9 FISN 1.624 .858 .189 1.893 .062 10. MBET** .098 .048 .163 2.054 .043 11. STPE .003 .037 .004 .071 .944 12. REEL .004 .008 .042 .466 .642 13. REIN .604 .381 .151 1.584 .117 Constant 7.247 8.174 .887 .378 N 100 F statistic (13, 87 ) 21.504*** Adjusted R square .727*** Note: *** significant at .01 level; ** significant at .05 level; significant at .10 level In Model B, D the raw number of charter schools in each school district I.V.s: MIDF=mimetic d iffusion; CODF=coercive diffusion; POEN=policy entrepreneurs; POIG=policy interest groups; PONW=policy network s capacity; FISN=fiscal situation; MBET =market formance; income

PAGE 177

161 CHAPTER VI CONCLUSIONS AND FUTURE STUDIES Conclusions Scholars (Mintrom, 2000; Mintrom & Vergari, 1998; Wong & Langevin, 2007; Wong & Shen, 2002) in the public administration and policy area have focused greatly on studying mechanisms of the charter school policy formulation at the state level while they have paid little attention to studying mechanisms of the charter school policy implementation at the local level. Up until now they have not conducted many studies empirically analyzing under what conditions school districts deliver more charter school service s to their customers. Therefore, th is disser t ation empirically explicating mechanisms of th e local charter school policy implementation plays a role as an academic pathfinder that lead s scholars to extend the scope to the policy implementation stage. For this, th is dissertation school districts in which the state charter school policy is practically implemented. Most importantly, th is dissertation determines this study to school district jurisdictions. Scholars (Schneider et al, 2000; Carlson, Lavery, & Witte, 2011; Witte, Schlomer, & Shober, 2007) in the public administration and policy area usually use this academic analysis standard for their studies. This analytical perspective helps scholars utilize more explanatory factors for their research. For instance, if researche rs fix school district governments instead of school district jurisdictions as units of analysis, they cannot use several social characteristics e, educational level, and race which are not related to characteristics of school district gove rnments. Therefore, u sing school district jurisdictions as th is disser t ation s units of analysis allows

PAGE 178

162 us to use more multiple explanatory factors relevant to characteristics of school district governments as well as characteristics of school district jurisdictions in analyzing the variation of the local charter school policy implementation. In the same context, th is dissertation expects that de termining this study s units of analysis as school district jurisdictions helps this dissertation cover a broader research scope and obtain more fruitful answers to the research question. On the other hand, t he Colorado school districts that this disser tation targets are the unknown research area in the case of the charter school policy implementation research (Lee & Jeong, 2012; Lee & Kim 2010). This current academic situation supports bution. The refore, educational policy scholars need to scrutinize mechanisms shown in the charter school policy implementation stage in the educational innovation process. Along with this view, t his dissertation plays a role of a trigger motivating and stimulating scholars to be interested in Colorado s school districts in exploring mechanisms of an innovative education policy. To supplement the lack of this academic research shown in the charter school research world, this dissertation creates a resea rch question: U nder what conditions the uneven charter school policy implementation appears among C s school districts ? Namely, t he primary goal of this dissertation is to find determi n ants that lead or cause the variation of the local charter schoo school district case. To seek answers to this research question, th is dissertation employs the four theoretical approaches the policy diffusion, policy entrepreneur, policy interest group, and policy network models to address this research question. And th is

PAGE 179

163 dissertation design s the empirical study focusing on quantitative methods. This empirical study is conducted by analyzing the two equation models Model A and Model B. Their difference depends on how to measure the dependent variable the variation in This means that all of the independent variables in both equation models are the same although operationalization of the two dependent varia bles is different. As described in Chapter 4 and 5, t h is dissertation tests both equation models with demonstrate that the three independent variables are statistically significant for the dependent variable while Model show that the seven independent variables are statistical ly significant for the dependent variable. The final statistical results of both equation models commonly reveal that the three independent variables, which are invol ved in the policy diffusion model, policy entrepreneur model, and policy network model, are influential factors in implementation. Namely, t he three statistically signific ant independent variables in both equation models suggest clear answers and evidence that as existence of neighboring school districts with previous experience for charter schools, policy entrepreneurs, and policy networks increases, the likelihood that Co lorado s school districts implement the state charter school policy increases as well In this final chapter, focusing on the three significant independent variables that the analyzed results of both equation models commonly support th is dissertation con cisely rearranges the empirical findings, interprets what they imply, and proposes what future studies must handle for better academic contributions.

PAGE 180

164 Findings T he empirical results for both equation models commonly d emonstrate that the three independent variables mimetic diffusion, policy entrepreneur, and policy network have the explanatory power in ex amining the variation of the local charter school policy implementation shown across Colorado T heir standardized coefficients show that the po licy network (PONW) independent variable is the most critical in answering the research question by indicating that the value of its standardized coefficients is higher than the other two variable standardized coeffic i ents That is to say, based on the results of standardized coeffic i ents, the policy network (PONW) independent variable is the most important and powerful explanatory variable among the three independent variable s in accounti ng for why the uneven implementation of charter school policy Policy Diffusion As usually known, the policy diffusion model has been utilized as a primary theoretical approach to account for mechanisms of the policy formulation stage. esearch topic, th is study expected this diffusion model to address the two academic hypotheses: 1) a neighboring school district affects a school district to deliver charter school services to their residents, and 2) a state at the upper level influences a school district to implement the state charter school law. The analyzed results of both Model A and Model B consistently indicate that a neigh bo ring school district is a pivotal factor in explaining why a school district more actively delivers charter sch ool service s to education customers in their territory However, the final result of the coercive diffusion variable demonstrates

PAGE 181

165 school policy. Over all, the statistical result of the mimetic diffusion factor, which is usually expressed as the regional diffusion in the public administration and policy area, supports that a neighboring school district work s as a n important factor in understand ing why a school district more actively implement s the state charter school policy compared to other school districts. Many previous studies (Berry & Ber ry, 1990; Mintrom, 2000; Renzulli & Roscigno, 2005; Wong, Langevin, 2007) have used the regional diffusion factor to account for mechanisms of multiple policy adoptions at the sta t e level. Their many empirical results strongly demonstrate that a neighborin g state having p revious experience with a similar policy influences a state t hat pursues adopting a new policy. Th is dissertation applies this regional diffusion logic to the policy implementation stage. This empirical finding reveals that the effects of a neighboring jurisdiction work well in explaining mechanisms of the policy implementation stage As mentioned in C hapter 3, Theoretical Approaches, this empirical result also demonstrates that it is possible to use contents f or the policy implementation stage Policy Entrepreneurs Model A and Model B consistently reveal that an educational entrepreneur who works for the school choice movement of each school district is a significant factor in implementation. In this disser t a tion the educational entrepreneur variable is measured by the number of educational entrepreneurs in each school district. The statistical result that test s the association between the policy entrepreneur (POEN) independent variable and

PAGE 182

166 the dependent variable supports that sc hool districts having many educational entrepreneurs more actively implement the state charter school policy. In other words, the existence of many educational entrepreneurs critically influences school districts to provide their customers with more charte r school services. Therefore, school districts that want to deliver more charter school services to their education customers need to receive advice and help of educational entrepreneurs by contacting educational entrepreneurs working for charter schools m ore often. As mentioned in the previous chapters, Budde and Shanker were famous educational entrepreneurs when for the first time a charter school appeared in the USA. In the case of Colorado, the literature review indicates that Bill Owen and Peggy Kerns played roles as policy entrepreneurs when Colorado voters passed the state charter school law in 1993. T independent variable demonstrates that as shown in both federal and state levels a policy entrepreneur is important as well charter school services to their residents. Policy Networks Over the past 2 decades, articles and books published in the public administration and policy area show that policy networks have been more important in understanding why policy actors formulate and carry out innovative institutions (Howlett, 2002; Meier The st yle of governance that del ivers public services to customers generally reaches either the lonely governmental governance style or the networked organizational governance style. The former governance style has been called hierarchical governance while the latter

PAGE 183

167 governance style has been named collaborative governance. Scholars in the public administration and policy area have been interested in exploring their roles or which one is better or more effective i n the policy process and public policy outcomes; that is to say, studies analyzing the role of the governance style on the policy process have been fixed as an interesting research topic up until now. Focusing on the role of the collaborative governance t his dissertation has applied this research style to the local charter school p o licy implementation. Public education policies are formulated by either national or state governments and then are implemented by school districts of local governments. To find explanatory factors that can account for the variation of the local charter school policy implementation, this dissertation target s Colorado s school districts as the units of analysis and utilize s policy network theory for creating the main hypothes e s T he final statistical result as expected, concretely demonstrates that the role of policy networks is very critical in explaining why there is the variation in the charter school policy implementation among Colorado s school districts. T he result shows tha t local governments highly collaborating with other organizations deliver more charter school policy services to their customers than local governments weakly collaborating with other organizations. This conclusion can be expressed as follows: A charter sc hool policy service is better delivered in the jointed or collaborative governance structure that several organizations create together than in the single or hierarch ical government structure. direct and top down style to the indirect and bottom up style, which is closer to the democratic society. Many leaders and scholars ( deLeon & deLeon, 2002; deLeon & Weible, 2010; Meier &

PAGE 184

168 Weible, Heikkila, deLeon, & Sabatier, 2012) recommend that governments pursue the collaborative governance through conversations and cooperation with other organizations. Therefore, these days, seeking the best ways to accomplish inter organizational arran gements well is a primary work o f governments, compare d to the past. In the indirect and bottom up style, collaboration with other organizations is a main governing tool to treat both social issues that local governments must solve and multiple upper government policies that they must carry out. However, unt il recently, it was not clear whether networked organizational settings matter in facilitating the local charter school policy implementation. In this dissertation the empirical finding for policy networks reemphasize s the importance of collaborative gove rnance by demonstrating that dense policy networks among each school district and the seven other organizations make the local charter school policy implementation more active. L ocal governments having close relationships and frequent connections with othe more actively implement the state charter school policy. This academic finding means that the successful charter school policy implementation depends on collaborative governance with other organ izations. Based on this empirical finding, we can more concretely support the claim that democratic institutional actions which work with other organizations are pivotal tools in completing successful charter school policy implementation, which means to deliver bett er or more charter school services to residents.

PAGE 185

169 Implications Scholars in the positivistic research world typically explore answers to their research questions through empirical studies (W i tte, 2000) They need theoretical approaches support ing their academic logic more than postpositivists who already know about specific answers or directions Thus, positivists strive to seek ways to make their logic and arguments more exact. As indicated in the previous chapters, this dissertation is orient ed to wards the positivistic research style In conducting positivistic research, th is dissertation first of all, proposes that scholars need to have an interdisciplinary open mind that receives similar contents that other disciplines cover when they build up academic logic supporting hypotheses in their studies. T his interdisciplinary approach provides positivistic scholars with fruitful logic from a greater array of academic disciplines. Overall, the interdisciplinary approach is pivotal in mak ing hypotheses more concrete. This disser tation has kept an interdisciplinary attitude in constructing hypotheses. Especially the mimetic diffusion model and policy network model in this dissertation become examples for the interdisciplinary approach. With t he interdisciplinary approach, th is dissertation strived to combine and integrate similar contents from both regional diffusion model and mimetic isomorphism model to construct the hypothesis relevant to this mimetic diffusion explanatory variable. Th rough this trial, the mimetic diffusion hypothesis could be more concise and in depth by receiving academic support from both political sciences and sociology Meanwhile, t h is dissertation has shown this interdisciplinary attitude in the policy network hypothes is as well. Ci ting the concept of networks that sociologists define in explaining policy networks, t his study tr ies s introduced in sociology into our

PAGE 186

170 academic field to make the hypothesis relevant to policy networks more concrete Currently the term networks ha s been very popular in multiple academic areas. This situation suggests that s cholars in the public administration and policy field need to have a pluralis tic attitude, which is based on an interdisciplinary view to obtain the rational e or logic of networks. I f we adhere to using one discipline to l ook at our research questions, we might lose good opportunities to obtain useful theoretical logic and rational e provided from other academic fields. Therefore, it is r easonable and advisable for us to have the pluralistic perspective when we conduct empirical studies that analyze mechanisms of the policy process. In the same vein, using the term multi ple application perspective, th is dissertation propose s that scholars must not limit range in using theoretical approaches as well. Namely we need to have a mind that we can apply a theory to any policy stages if needed. In fact, many scholars very stubbornly employ a specific theory to analyze a specific policy stage. They implicitly assert that the specific theory must be used in analyzing only the specific policy stage. Proceeding w ith this dissertation the author has been against this academic attitude. Concretely this disser t ation s empirical results support the multi ple application perspective The gives a good instance for th is Scholars in the public administration and policy field immerse themselves in using the regional diffusion model to explore mechanisms of the policy formulation stag e. However, t that the contents of the regional diffusion model can be used in accounting for mechanisms of the policy implementation stage as well

PAGE 187

171 the final interpretation of its statistical result, while the final results of both policy entrepreneurs and policy networks have fully interpreted what they mean. With this insight th is dissertation tries to more specifically interpret and catch what the mimetic diffusion s result implies applying the concept of social legitimacy to the final result of the mimetic diffusion. The final result of the mimetic diffusion varia ble demonstrates that a neighboring school district, which has previous experience in providing their residents with charter school services, is an influential factor in explaining why the variation of the local charter school policy implementation appears across Colorado. T o more deeply look at what the s final result implies it needs t o revisit Chapter 2 because the S shaped normal curve of the state charter school law adoption can offer an academic clue in conjecturing what the mimetic diffusion factor implies As indicated in Cha pter 2, the S shaped normal curve graph drawn in Figure II. 1 proposes one potential cause for why American states have adopted charter school law. I ts rationale is because a transaction cost that each state has is reduced when it adopts charter school law. In this case, a low transaction cost happens due to beneficial information, high efficiency, or strong trust that a state having previous experience for charter schools offer. As shown in Figure II. 1 in Chapter 2, the 28 other jurisdictions including Washington D.C. abruptly adopt ed their own cha r ter school law during 1994 1999 after the eight leading states pass ed cha r ter school law during 1991 1993 Th is dissertation supports the rationale of this phenomenon with the logic of transa c tion costs. In Chapter 2, the eight leading states play a role as a yardstick showing the performance and

PAGE 188

172 effectiveness of charter school services. 59 Namely, the 28 other jurisdictions adopting charter school law later ar e directly or indirectly affected by the eight leading states experience and outcome s for charter school services. The eight leading states provide the 28 other jurisdictions with beneficial information and values that help the 28 jurisdictions have their social legitimacy for charter schools. Therefore, the 28 jurisdictions adopted their charter school law when their transaction cost was relatively lower due to information and trust of charter schools obtained from the eight leading states. In the same context school districts that do not experience charter schools are concerned that charter school policy can be efficient or successful in their territory because they do not have enough information or trust for charter schools. Thus, t he analyzed result of this study supports that neighboring school districts, having previous experience operate charter schools in their territory. Generally speaking, new sc hool districts providing their residents with charter school services more actively implement the state charter school law when or because they can feel confident with effects of charter schools through experience of other school districts that have previo usly had charter schools As one strategic management that facilitates charter school services, this mimetic government officials or charter school organizers with guidelines showing how they foster char ter school services in their territory 60 Usually, scholars explain that when actors have strong legitimacy for an institution, the institution can be accepted by them well (Hannan & Carroll, 199 2; Lee, 59 Among states a d opting charter school law, the author calls the eight states the leading states. T h ey are Minnesota, California, Colorado, Georgia, Massachusetts, Michigan, New Mexico, and Wisconsin, and their charter school law adoption was accomplished during the initial period (1991 1993). 60 In this dissertation, the term policy implementers means policy actors who comply with the state charter school law.

PAGE 189

173 2014). P olicy implementers must consider how to strate gically increase legitimacy for an institution that they want to carry out. Lee (2014) shows tools fitting this approach, using South Korea as an example. He proposes that when citizens or education customers are not familiar with charter schools and do no t have social legitimacy for charter schools, policy implementers need to strategically utilize public meetings such as public hearings, forums, or seminars. These public meetings will facilitate education customers to increase their belief for charter sc hools because education customers can know what a charter school is and what its strengths are, compared to conventional public schools, through these public meetings. Some scholars (Aldrich & Fiol, 1994; Lee, 2014) point out that citizens are policy entit ies who influence governments to create, accept, or carry out a specific policy. If citizens or education customers hold social legitimacy for charter schools through public meetings, they will be pro charter school entities who push governments to authori ze and support charter schools. Thus, policy implementers who want to facilitate, vitalize, or dynamize charter school services need to often hold public meetings, which help citizens increase their social legitimacy for charter schools, as a strategy to accomplish their goals Future Studies In this final se ct ion, th is dissertation proposes some complementary ways for future studies. Based on the results of this e mpirical study, there may be some limitations in the empirical design and research methods. Although they are not big issues that affect more conc r e te and stronger answers for a study accounting for mechanisms of the local charter school policy

PAGE 190

174 implementation if these complementary wa y s are applied to make up for some limitations. For this, th is dissertation mentions three complementary ways. The first is to make the validity of the market based education tool (MBET) independent variable higher. This dissert a tion utilizes the raw numb er of each school s private schools to measure this MBET variable. In reality, many education policy scholars have been using a private school as a proxy to operationalize school choice movement tools because a private school virtually competes wi th other schools in the TPS system. T he validity power of this MBET variable can be increased if other tools in the category of the school choice movement are used as proxies for assessing this MBET variable. Charter schools and private schools are essent ially different. The former are public while the latter are private although both school entities make the American education s ystem more competitive. T his difference between the two school entities the extent to which a measurement (or proxy) exactly gauges what it is supposed to measure ( Babbie, 2010; Kumar, 1996 ; McClendon, 2004) Thus, this characteristic difference of the two school entities can To reduce this valid ity problem in the dissertation the author proposes that future studies should use other school entities under the school choice movement as proxies estimating this MBET variable For this, it is good to use a m agnet school as a proxy because like charter schools, magnet schools are public schools and are representative school entities supporting the school choice movement. Therefore, to improve and

PAGE 191

175 MBET variable in fu ture studies. The second suggestion is that future studies need to use a variable explaining factors are not tested. However, the literature review highlights th at a political tendency is an important factor in examining mechanisms of the policy process. Fusarelli (2003) and Teske (1991) emphasize that the political tendency of policy actors is a primary determinant affecting the public policy process. Their resea rch shows that the pol itical attitudes Democratic or R epublican determine if policy actors actively or slowly conduct a public service delivery. With this view, future studies need to explore whether the political tendency of school district residents infl uences school districts to more actively implement the state charter school law. to help citizens learn social order s and law s under the same education contents. Pro public school education leaders believed that this goal was achievable under a one size fits all education system. However, its direction and frame, which are presented as the command control and universal public edu cation, are inconsistent with the core Republican values that center on individual creativity, liberty, and personality. Traditionally, Republicans have been suspicious of group minded policies (Urban & Wagoner, 2009). Republicans argue that regular public schools ha ve been dysfunctional in improving individual gifts and producing better student performance, and regular public school weaknesses must be changed with school choice options. This means that Republicans have not liked and trusted the one size fi ts all form shown in public

PAGE 192

176 education. And, they advocate providing education customers with more education services, emphasizing customer rights and requesting other school options. 61 T h us, Republicans are more recognized as charter school friendly entitie s than D emocrats are because their philosophy originally pursues individual creativity and respects personality while D emocrat ic This perspective supports that Republicans want charter schools to e xpand and open across their territory in order for children to receive multiple education services. Based on this logic to the Republican philosophy, it is meaningful for future studies to test this political tendency hypothesis that school districts with more R epublican residents are more likely to implement the state charter school law. At the school district level, we can measure a political tendency of each school district through the school district board members ults in a state or federal legislative election, senate election, or presidential election. The final suggestion is to propose using a qualitative method. Many scholars ( Bernard, 2000; Duffy, 1987; Jick, 1979; Onwuegbuzie & Leech, 2004) point out that a quantitative study has some limits in obtaining in depth answers to research questions. The author has a similar opinion with other scholars in designing and analyzing this empirical study. For instance, the final statistical results demonstrate that pol icy entrepreneurs influence the variation of the local charter school policy implementation. However, t hrough this analyzed result, the author could not understand what it specifically mean s That is, this final finding does not describe who policy entrepr eneurs 61 In the American education policy, Urban and Wagnoner (2009) describe the former president Bill Clinton as a liberal sheep putting on conservative wolf s customs (p. 390). The Obama administration s situation can be also explained as the Clinton administration s situation in leading American education policy. T his tendency shows that the two administrations deviate from the typical Democratic policy direction in pursing the American education policy.

PAGE 193

1 77 are, what kind of roles they play in helping school districts deliver charter school services to their residents, or how they affect charter school services of each school district. Meanwhile the final result of the policy network factor indicates that a strong policy network that each school district and the seven organizations form is an important factor in explaining the variation of the local charter school policy implementation. However, this finding does not show what organization plays a rol e as a main actor, what primary strategic tools an organization uses to connect with each other are, or who an individual actor leading these connections with other organizations in each organization is. Moreover, these days information exchange and gathe ring are accomplished through online methods rather than o f f line ones. It is necessary for a scholar to examine how each organization supports or uses online tools to make its networks more active To seek answers to these complicated questions, future studies need to employ and conduct a qualitative method such as an interview or observation focusing on respondents who work for charter school services in each school district. In applying this qualitative method to future studies, first of all, it is rea sonable to target staff members, who are charged with adopting and carrying out school choice movement tools as each school district interviewees because they know about how school choice movement tools were introduced to their own school district in the past years, who has been deeply involved in carrying out school choice movement tools and which organizations (or stakeholders) strongly and actively support their own school choice movement. Some scholars (Bogdan & Biklen, 2007; Creswell, 2007) emphasize how important informants are in a qualit ative method. They explain that informants such as interviewees working in areas related to a research topic can provide more proper

PAGE 194

178 answers fitting the research topic and explaining main issues in the study. Theref ore, t hese qualitative interviews targeting a staff member working for the school choice movement in each school district will allow this study to obtain deeper knowledge of the aforementioned complex factors that lead a schoo l district to more actively deliver charter schools to their residents. The qualitative research is mainly conduct ed through small N observation s because many times and high costs in conducting a qualitative method are requested ( Bernard, 2000 ; Crouch & Mc Kenzie, 2006; Mahoney, 2000 ; Tjora; 2006 ) This condition proposes that future studies need to first choose a school district that more actively uses school choice movement tools. Therefore, in the Colorado school district case, the three school districts operating many charter schools Jefferson County School District R 1, Douglas County School District RE 1, and Denver County School District 1 can be the most proper school districts in conducting a qualitative method.

PAGE 195

179 REFERENCES Abernathy, S. F. (2005). School choice and the future of American democracy Ann Arbor, MI: University of Michigan Press. Agranoff, R. (2004). Leveraging networks: A guide for public managers working across organizations. In J. M. Kamensky, & T. J. Burlin (Eds.), Collaboration: U sing networks and partnerships (pp. 61 102). Lanham, MD: Rowman and Littlefield Publishers, Inc. Aldrich, H. E. (1979). Organizations and environments Englewood Cliffs, NJ: Prentice Hall. Aldrich, H. E. (2008). Organizations and environments Stanford CA: Stanford Business Books. Aldrich, H E., & Fiol, M C. ( 1994 ) Fools r ush i n? The i nstitutional c ontext of i ndustry c reation. Academy of M anagement Review 19 (4) 645 670. Aldrich, H. E ., & Whetten, D. (1981). Organizations, action sets, and networks: Making the most of simplicity. In P. C. Nystrom & W. H. Starbuck (Eds.), Handbook of organizational design (pp. 385 408). N ew Y ork, NY : Oxford University Press. Alleyne, S. (2014). The advantages of traditional schools Retrieved from http:// classroom.synonym.com/advantages traditional schools 3706.html Alston, L. J., & Gillespie, W. (1989). Resource coordination and transaction costs: A framework for analyzing the firm/market boundary. Journal of Economic Behavior and Organization, 11 191 212. Alvesson, M., & Skoldberg, K. (2009). Reflexive methodology: new vistas for qualitative research (2 nd ed.). Thousand Oaks, CA: S age Publications Inc. Anderson, T. L., & Parker, D. P. (2013). Transaction costs and environmental markets: The role of entrepreneurs. Review of Environmental Economics and Policy, 7 (2), 259 275. Andlovic, M & Lehamann, W. (2014). Interest group influence and interinstitutional power allocation in early second reading agreements: A re examination of aviation emissions trading. Journal of European Public Policy, 21 (6), 802 821. Anrig, G. (2008). An idea whose time has gone: Gonserv atives abandon their support for school vouchers Washington Monthly. Retrieved from http://www. washingtonmonthly.com/ features/2008/0804.anrig.html

PAGE 196

180 Babbie, E. (2010). The practice of social research (12 th ed.). Belmont, CA: Wadworth. Babbie, E. Halley, F. S., Wagner, W. E., & Zaino, J (2013). Adventures in social research: Data analysis using IBM SPSS statistics ( 8 th ed.) Thousand Oaks, CA: S age Publications, Inc. Baez, B., & Abolafia, M. Y. (2002). Bureaucratic entrepreneurship a nd institutional change: A sense making approach. Journal of Public Administration Research and Theory, 12 (4), 525 552. Balla, S J. ( 2001 ) Interstate p rofessional a ssociations and the d iffusion of p olicy i nnovations. American Politics Research 29 (3) 221 245. policing. Economic Inquiry, 25 (1), 103 116. Basilico, E., Broggini, P, Eder, I., & Truax, K. (2002). Reasons, consequences, alternative for below average compensation Retrieved from http:// www.centerfortaxpolicy.org/reports/02 teachers_salaries.pdf Baumgartner, F. R., & Jones, B. D. (1993). Agenda and instability in Americ an politics. Chicago IL : University of Chicago Press. Baumgartner, F. R., & Leech, B. L. (1998). Basic interests Princeton, NJ: Princeton University Press. Belfield, C. R., & Levin, H. M. (2005). Pr i vatizing educational choice: Consequences for parents, schools, and public policy Boulder, CO: Paradigm Publishers. Benigno, P., & Morin, K. (2013). On the road of innovation: C o lorado s charter school law turns 20 Retrieved from http://www.i2i.org/files/file/IP 4 2013.pdf Benton, J. E. (2002). County service delivery: Does government structure matter? Public Administration Review, 62 (4), 471 479. Bernard, H. R. (2000). Social research methods Thousand Oaks, CA: Sage Publication s. Berry, F. S., & Berry, W. D. (1990). State lottery adoptions as policy innovations, American Political Science Review, 84 (2), 395 413. Berry, F. S., & Berry, W. D. (1994). The politics of tax increases in the states. American Journal of Political Scie nce 38 (3), 855 859. Berry, F. S., & Berry, W. D. ( 1999 ). Innovation and diffusion models in policy research. In P.A. Sabatier (Ed.), Theories of the policy process (pp.169 200). Boulder CO : Westview press

PAGE 197

181 Berry, F. S., & Berry, W. D. (2007). Innovation and diffusion models in policy research. In P.A. Sabatier ( 2 nd e d.), Theories of the policy process (pp. 223 260). Boulder, CO: Westview press Berry, F. S., & Flowers, G. (1999). Public entrepreneurs in the policy process: Performance based bu dgeting reform in Florida, Journal of Public Budgeting, Accounting, & Financial Management, 11 (4), 578 617. Bierlein, L. A. (1993). Controversial issues in education al policy Newbury Park CA : Sage Publications, Inc. Birkland, T. A. (2001). An introduct ion to the policy process: Theories, concepts, and models of public policy making Armonk, New York: M. E. Sharpe, Inc. Boatright, R. G. (2011). Interest groups and campaign finance reform in the United States and Canada Ann Arbor, MI: University of Mich igan Press. Bogdan, R. C., & Biklne, S. K. (2007). Qualitative research for edu c ation: An introduction to theories and methods (5 th ed.). Boston, MA: Allyn and Bacon Bord, R. J., O Connor, R. B., & Fisher, A. (2000). In what sense does the public need to understand global climate change? Public Understanding of Science, 9 205 218. Borgatti, S. P., & Foster, P. C. (2003). The network paradigm in organizational research: A review and typology. Journal of Management, 2 9 (6), 991 1013. Borgatti, S. P., Everett, M. G., & Freeman, L. C. (200 2 ). UCINET 6 for windows software for social network analysis Retrieved from http://pages.uoregon. edu/vburris/hc431/ Ucinet_Guide.pdf Boyett, I. (1997). The public sector entrepreneur A definition. International Journal of Entrepreneurial Behavior and Research, 3 (2) 77 92. Bressers, H. T., & O Toole, L. J. (1998). The selection of policy instruments: A network based perspective. Journal of Public policy, 18 (3), 213 239. Breul, J. D., & Kamensky, J. M. (2008). Federal government reform: Lessons from Cli n initiatives. Public Administration Review, 68 (6), 1009 1026. Brewer, G. D., & deLeon, P. (1983). The foundations of policy analysis Homewood, IL: The Dorsey Press.

PAGE 198

182 Brudney, J. L., Hebert, F. T., & Wright, D. S. (1999). Reinventing government in the American states: Measuring and explaining administrative reform. Public Administration, 59 (1), 19 30. Bryk, A. S. & Gomez, L. M. (2008). Reinventing a research and development capacity. In F. M. Hess (Ed.), The future of educational entrepreneurship: possibilities for school reform (pp.181 206). Cambridge, MA: Harvard Education Press. Brynard, P. (2005). Policy implementation: Lessons for service delivery. Joural of Public Administration, 40 (special issue 3), 649 664. Buckley, J., & Schneider, M. (2007). Charter Schools: Hope or Hype? Princeton, NJ: Princeton University Press. Buddin, R., & Zimm er, R. (2005). Student achievement in charter school: A complex picture. Journal of Pol icy Analysis and Management, 24 (2), 351 371. Burt, R. S. (2000). The network structure of social capital. Organizational Behavior, 22, 345 423. Cahn, M. A. (1995). The players: Institutional and noninstitutional actors in the policy process. In S. Z. Theodoulou, & M. A. Cahn (Eds.), Public policy: The essential readings (pp. 201 211). Englewood Cliffs, NJ: Prentice Hall, Inc. Carlson, D., Lavery, L., & Witte, J. F. (2011). The determinants of interdistrict open enrollment flows: Evidence from two states. Educational Evaluation and Policy Analysis, 33 (1), 76 94. Carnochan, S. (2002). Reinventing government: What unbran school districts can learn form charter school. In A. S. Wells (Ed.), Where charter school policy fails: The problems of accountability and equity (pp. 54 76). New York, NY: Teachers College Press. Carpenter, D. M., & Kafer, K. (2009). A typology of Colorado charter schools. Retrieved from http :// www. cde.state.co.us/cdechart/download/typologyreport_012709.pdf Carpenter, D. M., & Kafer, K. (2013). T h e state of charter schools Retrieved from http://www.cde.state.co.us/sites/default/files/Charter%20Report%202013%20040 413c.pdf Center for Education Reform. (2010). Retrieved from http://www.edreform.com/wp content/uploads/2011/09 / CER_Charter_Survey_ 2010.pdf

PAGE 199

183 Center for Education Reform (2012). Charter school laws across the states Retrieved from http://www.edreform.com/wp content/uploads/2012/04/CER_2012_ Charter_Laws.pdf Center for Education Reform (2012 b ). K 12 facts Retrieved from http://www. edreform. com/2012/04/k 12 facts/ Center for Education Reform. (2013 a ). America s attitudes towards education reform: Executive summary Retrieved from http://www.edreform.com/2013/10/americas attitude towards education reform executive summary/ Center for Education Reform. (2013 b ). Charter school law rankings and scoreboard Retrieved from http://www.edreform. com/wp content/uploads/2013/06/CER CharterLaws 2013 _Chart_FINAL.pdf Chubb, J. & Moe, T. (1990). Politics, markets, and the organization of schools Washington. D.C.: Brookin gs. Choy, S. P. (1997). Public and private schools: How do they differ? Retrieved from http://nces.ed.gov/pubs97/97983.pdf Cibulka, J. (2001). The changing role of interest groups in education: Nationalization and the new politics of education productivity. Educational Policy, 15 (1), 12 40. Cigler, A. J., & Joslyn, M (2002). Groups, social capital, and democratic orientation s. In A. J. Cigler, & B. A. Loomis (Eds.), Interest group politics (pp.37 53). Washington DC: CQ press. Cohn, E. (1997). Public and private school choices: Theoretical consideratons and empirical evidence. In E. Cohn (Ed.), Market approaches to education (pp.3 20). New York NY : Elsevier Science Inc. Colorado Department of Education. ( 2007 ) Summary of school district data for school year 2006 2007 Retrieved from http://www.cde.state.co.us/cdereval/download /PDF/ 20062007 SummaryofSchool DistrictData.pdf Colorado Department of Education. (2013). Rural definition Retrieved from http://sites cde.state.co.us/sites/default/files/documents/ruraledcouncil/download/ruraldefiniti onspreadsheet032913.pdf Colorado League of Charter Schools. (201 3 ). Charter school facts. Retrieved from http://www.coloradoleague org/colorado charter schools/charter schools fact sheet.php

PAGE 200

184 Colvin, R. L. (2004). Public school choice: An overview. In Hess, F. M., & Finn, C. E. (Eds.), Leaving no child behind? Options for kids in failing schools (pp.11 36). New York NY : Palgrave Macmillan. Condon, S. (2010). Obama calls for big education reforms more charter schools, longer school year Retrieved from http://www.cbsnews.com/news/obama calls for big education reforms more cha rter schools longer school year Cooper, B. S., & Sureau, J (2008). The collective politics of teacher unionism. In B. S. Cooper, J. G. Cibulka, & L. D. Fusarelli (Eds.), Handbook of Education Politics and Policy (pp. 263 280). New York, NY: Routledge. Corwin, R. G., & Schneider, J. E. (2005). The school choic e hoax: Fixing America s schools Westport, CT: Praeger Publihsers. Coughlin, J. F., & Tompkins, C. J. (2009). Demographics, density, and anticipating the future of the transportation system. Public Works and Policy, 13 (4), 284 287. Creswell, J. W. (2007). Qualitative inquiry and research design: Choosing among five approaches. Thousand Oaks, CA: Sage Publication, Inc. Crouch, M., & McKenzie, H. (2006). The logic of small samples in interview based qualitative research. Social Science Information, 45 (4), 483 499. Daley, D. M., & Garand, J. C. (2005). Horizontal diffusion, vertical diffusion, and internal pressure in state environmental policymaking, 1989 1998. American Politics Research, 33 (5), 615 644. deLeon, L. (1996). Ethics and entrepreneurship. Polic y Studies Journal, 24 (3), 495 510. deLeon, L., & Denhardt, R. B. (2000). The political theory of reinvention. Public Administration Review, 60 (2), 89 97. deLeon, P ( 1999 ). The stages approach to the policy process In P.A. Sabatier (Ed.), Theories of the policy process (pp.19 3 2). Boulder CO : Westview press deLeon, P., & deLeon, L. (2002). What ever happened to policy implementation? An alternative approach. Journal of Public Administration Research and Theory, 12 (4), 467 492. deLeon, P & Varda, D. M. (2009). Toward a theory of collaborative policy networks: Identifying structural tendencies. Policy Studies Journal, 37 (1), 59 74. deLeon, P & Vogenbeck, D. M. (2007). Back to square one: The history and promise of the policy sciences. In J. Rabin, W. B. Hildreth, & G. J. Miller (Eds.), Handbook of public administration (pp. 503 539). New York, NY: Taylor & Francis. deLeon, P., & Weible, C. M. (2010). Policy process research for democracy: A

PAGE 201

185 commentary on Lasswell s vision. Internation al Journal of Policy Studies, 1 (2), 23 34. Denhardt, J. V., & Denhardt, R. B. (2007). The new public service: Serving, not steering Armonk, NY: M.E. Sharpe. Denhardt, R. B. (2008). Theories of public organization (5 th ed.). Belmont, CA: Thomson Wadsworth Denhardt, R. B., & Denhardt, J. V. (2000). The new public service: Serving rather than steering. Public Administration Review, 60 (6), 549 559. DiMaggio, P., & Powell, W. (1983). The iron cage revisited: Institutional iso morphism and collective rationality in o rganizational fields. American Sociological Review, 48 (2), 147 160. Duffy, M. E. (1987). Methodological triangulation: A vehicle for merging quantitative and qualitative research methods. Journal of Nursing Scholarship, 19 (3), 130 133. Dryz ek, J. S. (2002). A post positivist policy analytic travelogue The Good Society, 11 (1), 32 36. Finn, C. E. (2008). Quality control in a dynamic sector. In F. M. Hess (Ed ), The future of educational entrepreneurship: Possibilities for school reform (pp. 1 61 179). Cambridge, MA: Harvard Education P r ess. Finn, C. E. (2013). Margaret Thatcher, education reformer Retrived from http:// www.nationalreview.com/corner/345062/margaret thatcher education reformer chester e finn jr Finn, C. E., M an n o, B. V., & Vanourek, G. (200 0 ). Charter schools in action: Renewing public education Princeton, NJ: Princeton University Press. Finn, C. E., Man n o, B. V., & Vanourek, G. (2001). Chater schools: Taking Stock. I n P. E. Peterson, & D. E. Campbell (Eds.), Charters, vouchers, and p ublic e ducation (pp. 19 42). Washington, D. C.: Brookings Institution Press. Fischer, F. (1998). Beyond empiricism: Policy inquiry in postpositivist perspective. Policy Study Journal, 26 (1), 129 146. Fischer, F. (2003). Reframing public policy: Discursive politics and deliberative practices Oxford, N Y : Oxford University Pres s. Fisher, J. L., & Koch, J. V. (2008). Born, not made : The entrepreneurial personality Westport, CT: Praeger Publishers.

PAGE 202

186 Foster, R. H., McBeth, M. K., & Clemons, R. S. (2010). Public policy pedagoy: Mixing methodologies using cases. Journal of Public Affairs Education, 16 (4), 517 540. Fraze, S. D., Hardin, K. K., Brashears, M. T., Haygood, J. L., & Smith, J. H. (2003). The effects of delivery mode upon survey response parte and perceived attitudes of Texas agri science teachers. Journal of Agricultural Education, 44 (2), 27 37. Frederickson, H. G. (1996). Comparing the reinventing government movement with the new public admin istration. Public Administration Review, 56 (3), 263 270. Freeman, L. C. (2004). The development of social network analysis: A study in the sociology of science North Charleston, SC: BookSurge, LLC. Friedman, M. (1962). Cap i talism and freedom Chicago, IL: University of Chicago Press. Friedman, M. (2002). Cap i talism and freedom (40 th Anniversay ed.). Chicago, IL: University of Chicago Press. Frumkin, P., Manno, B. V., & Edgington, N. (2011). The strategic management of charter schools: Frameworks and tools for educational entrepreneurs Cambridge, MA: Harvard Education Press. Fuhrman, S. H. (2001). Introduction. In S. H. Furhman (Ed.), From the capital to the classroom: standards based reform in the states: The one hundredth yearbook o f the national society for the study of education, Part 2 (pp.1 12). Chicago IL : University of Chicago Press. Fusarelli, L. D. (2003). The political dynamics of school choice: Negotiating contested terrain New York, NY : Palgrave Macmillan. Gauri, V. ( 1998). School choice in Chile: Two decades of educational reform Pittsburgh, PA: University of Pittsburgh Press. George, D., & Mallery, P. (2013). IBM SPSS statistics 21 step by step: A simple guide and reference ( 13 th ed.). Boston, MA : Pearson. Gerge n, C., & Vanourek, G. (2008). Talent development: looking outside the education sector. In F. M. Hess (Ed ), The future of educational entrepreneurship: possibilities for school reform (pp. 23 44). Cambridge, MA: Harvard Education Press. Gerson, M. (2013). A victory for school choice. Real Clear Politics Retrieved from http://www.realclearpolitics.com/articles/2013/04/02/a_victory_for_school_ choic e_117752.html

PAGE 203

187 Gillard, D. (2011). Education in England: A brief history Retrieved from http://www educationengland.org.uk/history/chapter08.html Gonring, P. Teske, P., & Jupp, B. (2007). Pay for performance teacher comp ensation: Cambridge, MA: Harvard Education Press. Good, T. L., & Braden J. S. (2000). Charter schools another reform failure or a worthwhile investment? Phi Delta Kappan, 81 (10), 745 755. Gore, Albert (1993). From red tape to results: Creating a government that works better & costs less Washington, D.C.: Government Printing Office. Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78 (6), 1360 1380. Granovetter, M. (1982). The strength of weak ties: A network theory revisited. In P. Marsden & N. Lin. (Ed s .), Social structure and network analysis. Beverly Hills, CA: Sage. Gray, V. (1973). Innovation in the States: A d iffusion s tudy American Political Science Review, 67 (4), 11 74 1185. Griffin, J. (2013). Denver Post. Retrieved from http://www.denverpost.com/c i_23261474/colorados charter schools their history and their future# Grossberg, B. (2014). Five major differences between public and private schools Retrieved from http://privateschool.about.com/od/choosingaschool/a/Five Major Differences Between Public And Private Schools.htm Gruber, J. (2011). Public finance and public policy (3 rd ed.). New York, NY: Worth Publishers. Gujarati, D. N. (2003). Basic econometrics New York, NY: McGraw Hill Higher Education. Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis (5 th ed.). Upper Saddle River, NJ: Prentice Hall. Hannan, M. T., & Carroll, G. R. ( 1992 ) Dynamics of organizationa l populations: Density, legitimation, and competition New York NY : Oxford University Press. Hawkins, C. (2011). Electoral support for community growth management policy Social Science Quarterly, 92 (1), 268 284.

PAGE 204

188 Heclo, H. (1974). Modern social politics in Britain and Sweden New Haven CT : Yale University Press. Heclo, H. (1995). Issue networks and the executive establishment. In S. Z. Theodoulou & M.A. Cahn (Eds.), Public policy: T he essential readings (pp. 46 58). Englewood Cliffs, NJ : Prent ice Hall Henig, J. R., & Sugarman, S. D. (1999). The nature and extent of school choice. In J. R. Henig, & S. D. Sugarman (Eds.), School choice and school controversy: Politics, policy, and law (pp.13 35). Washington, D.C.: Brookings Institution Press. Hess, F. M. (2008). The future of educational entrepreneurship: P ossibilities for school reform Cambridge, MA: Harvard Education Press. Hess, F. M. (2010). Does school choice Work ? National Affairs, 5 (Fall), 35 53. Hill, P. J. (1999). Public choice: A review. Faith and Economics, 34 1 10. Hill, P. T. (2010 ) Learning as we go: Why school choice is worth the wait Stanford, CA: Hoover Instituion Press. Hill, P. T., & Jochim, A. E. (2009). Political perspectives on school choice. In M. Berends M. G. Springer, D. Ballou, & H. J. Walberg (Eds ) Handbook of research on school choice (pp. 3 17 ). New York, NY: Routledge. Hill, P. T., & Lake, R. (2010). The charter school catch 22. Journal of School Choice, 4 232 235. Hirsch, E. (2002). The Colorado charter schools act. In S. Vergari (Ed.). The charter school landscape (pp. 93 112). Pittsburgh, PA: University of Pittsburgh Press. Holcomb, P. A., & Nightingale, D. S. (2003). Conceptual undrpinnings of implementation an alysis. In M. C. Lennon, & T. Corbett (Eds.), Policy into action: Implementation research and welfare reform (pp. 39 55 ). Washington, D.C.: Urban Institute Press. Holzinger, K., & Knill, C. (2005). Causes and conditions of cross national policy convergenc e. Journal of European Public Policy, 12 (5), 775 796. Howlett, M. (2002). Do networks matter? Linking policy network structure to policy outcomes: evidence from four Canadian policy sectors 1990 2000. Canadian Journal of Political Science, 35 (2), 235 267. Howell, P. L., & Miller, B. B. (1997). Sources of funding for schools. The Future of Children 7 (3), 39 50.

PAGE 205

189 Howell, W. G., & Peterson, P. E. (2006). The education gap: Vouchers and urban schools Washington, D.C.: Brookings Institiution Press. Huffington P ost (2012). Best education in the world: Finland, South Korea top country rankings, U.S. rated average Retrieved from http://www.huffingtonpost.com /2012/11/27/best education in the wor_n_2199795.html Huson, J. A. (2014). The a dvantages of traditional schools Retrieved from http:// education.seattlepi.com/advantages traditional schools 2140.html Jefferson, S. P. (2004). Charter schools: A descriptive study of empowerment within the operation of charter schools. Boca Raton, FL: Disseration.com. Jeong, M. (2006). Local choices for development impact fees. Urban Affairs Review, 41 (3), 338 357. Jick, T. D. (1979). Mixing qualitative and quantitative methods: Triangulation in action. Administrative Science Quarte rly, 24 (4). 602 611. Julnes, P., & Holzer, M. (2001). Promoting the utilization of performance measures in public organizations: An empirical study of factors affecting adoption and implementation. P ublic Administration Review, 61 (6), 693 708. Kamarck, E C. (2004). Applying 21 st century government to the challenge of homeland security. In J. M. Kamensky & T. J. Burlin (Eds.), Collaboration: Using networks and partnerships (pp. 103 146). Lanham, MD: Rowman and Littlefield Publishers, Inc. Kamensky, J. management reform. Public Administration Review, 56 (3), 247 255. Kearney, R. C., & Hays, S. W. (1998). Reinventing government, the new public management and civil service systems in inte rnational perspective: The danger of throwing the baby out with the bathwater. Review of Public Personnel Administration, 18 (4), 38 54. Kellough, J. E, & Selden, S. C. (2003). The reinvention of public personnel administration: An analysis of the diffusi on of personnel management reforms in the states. Public Administration Review, 63 (2), 165 176. Kemerer F R. (2009). A legal perspective on school choice. In M. Berends M. G. Springer, D. Ballou, & H. J. Walberg (Eds ) Handbook of research on school choice (pp. 19 34). New Haven CT : Routledge. Kennedy, P. (2008). A guide to econometrics (6 th ed.). Malden, MA: Wiley Blackwell.

PAGE 206

190 Kennedy, R. (2014). Comparison of private and public schools: A look at the difference and similarities Retrieved from http://privateschool.about.com/od/ choosingaschool/qt/comparison.htm Kettl, D. F. (2002). The transformation of governance: Public administration for twenty first century America. Baltimore, MD: Johns Hopki ns University Press. Khakee, A. (2003). The emerging gap between evaluation research and practice. Evaluation, 9 (3), 340 352. Kingdon, J. W (2003). Agendas, alternatives, and public policies. New York, NY: Addison Wesley Educational Publishers Inc. Kirst, M. W. (2007). Politics of charter schools: Competing national advocacy coalitions meet local politics. Peabody Journal of Education, 82 (2 3), 184 203. Klingner, D. E. (2006). Diffusion and adoption of innovations: A development perspective. In Unit ed Nations (Eds.), Innovations in governance and public administration: Replicating what works (pp. 61 66). New York, NY: United Nations Publication. Kolderie, T. (2005). Retrieved from http://www.educationevolving.org/pdf/Ray Budde Origins Of Chartering.pdf Kumar, R. (1996). Research methodology: A step by step guide for beginners Thousand Oaks, CA: SAGE Publicati on Ltd. Landry, R., Lamari, M., & Amara, N. (2003). The extent and determinants of the utilization of university research in government agencies. Public Administration Review, 63 (2), 192 205. Lane, J. E. (2000). The public sector: Concepts, models and approaches (3 rd ed.). Thousand Oaks, CA : Sage Publications Lawrence, C. (2014). The advantages of traditional schools Retrieved from http:// everydaylife.globalpost.com/advantages traditional schools 16699.html Lee, J. (2014). T h e mechanism of stat e charter school policy adoption: The case of American states. International Journal of Public Administration, 37 (1), 20 34. Lee, J., & Choi, Y. (2011). What determinants influence the implementation of growth management policy in Colorado? International Review of Public Administration, 16 (1), 135 156. Lee, J., & Jeong, J. (2012). Explanatory factors for the uneven implementation of charter school policy. Korean Journal of Policy Studies 27 (1), 153 172.

PAGE 207

191 Lee, J & Kim, S. ( 2010 ) Determinants of c harter s chool p olicy a doption in U.S. s tates: A c ase s tudy of Colorado s chool d istricts. International Review of Public Administration 15 (2) 29 44. Lee, C., Chang, K., & Berry, F. S. (2011). Testing the development and diffusion of e government and e democracy: A global perspective. PublicAdministration Review 71 (3), 444 454. Leech, N. L., Barrett, K. C & Morgan, G. A. (20 11 ). IB M SPSS f o r intermediate statistics : U se and interpretation (4 th ed.). New York, NY: Routledge. Leisey, D. E., & Lavaroni, C. (2000). The educational entrepreneur: making a difference San Rafael, CA: Edupreneur Press. Lemieux, P. (2004). The public choice revolution: Public choice has demystified and unedified the state. Regulation, 27 (3), 22 29. Lenkowsky, L. & Perry, J. L. (2000). Reinventing government: The case of national service. Public Administration Review, 60 (4), 298 307. Levin, H. R. (2009). An economic perspective on school choice. In M. Berends M. G. Springer, D. Ballou, & H. J. Walberg (Eds ) Handbook of research on school choice (pp. 19 34). New York, NY : Routledge. Levy, T. (2010). Charter schools legislation and the element of race. Western Journal of Black Studies, 34 (1), 43 52. Lowi, T. J. (1964). American business, public policy, case studies, and political theory. World Politics, 16 (4), 677 715. Lubell, M., Feiock, R., & Ramirez, E. (2005). Political institutions and conservation by local government. Urban Affaris Review, 40 (6), 706 739. Lubell, M., Scholz, M., & Mete, M (2002). Watershed partnerships and the emergence of collective action institutions. American Journal of Political Science 46 (1), 148 163. Lynn, L. E. (1999). A place at the table: Policy analysis, its postpositive critics, and the future of practice. Journal of Policy Analysis and Management, 18 (3), 411 424. Mahoney, J (2000). Strategies of causal inference in small N analysis. Socio logical M ethods and Research, 28 (4), 387 424. Makse, T, & Volden, C. ( 2011 ) The role of policy attributes in the diffusion of innovations. Journal of Politics 73 (1) 108 124.

PAGE 208

192 Malen, B. (2001). Generating interest in interest groups. Educational Policy, 15 (1), 168 186. Maranto, R., Kayes, M. S., & Maranto, A. G. (2006). Charter schools and school reform: In M. S. Kayes, & R. Maranto (Eds.), A guide to charter schools: Research and practical advice for educators (pp. 1 11). Lanham, MD: Rowman and Littlefield Education. Mawhinney, H. B., & Lugg, C. A. (2001). Interest groups in United States education, Educational Policy, 15 (1), 3 11. Mazzoni, T. L. (1995). State policymaking and school reform: Influences and influentials. In J. D. Scribner, & D. H. Clayton (Eds.), The study of educational politics (pp. 53 73). Washington, D.C.: The Falmer Press. McBeath, J., Reyes, M. E., & Ehrlander, M. (2008). Education reform in the American states Charlotte, NC: Information Age Publishing, Inc. McClendon, M. J. (2004). Statistical analysis in the social sciences Belmont, CA: Wadsworth/Thomson Learning. 1 ). Managerial strategies and behavior in networks: A model with evidence from U.S. public education. Journal of Public Administration and Theory, 11 (3), 271 293 Milward, H. B., & Provan, K. G. (2000). Governing the hollow state. Journal of Public Administration Research and Theory, 10 (2), 359 379. Mintrom, M. (1997). Policy entrepreneurs and the diffusion of innovation. American J ournal of Political Science, 41 (3 ), 738 770. Mintrom, M. (2000 ). Policy entrepreneurs and school choice. Washington, D.C.: Georgetown University Press. Mintrom, M (2001). Educational governance and democratic practice. Educational Policy, 15 (5), 615 643. Mintrom, M., & Vergari, S. (19 96). Advocacy coalitions, policy entrepreneurs, and policy change. Policy Studies Journal, 24 (3), 420 434. Mintrom, M., & Vergari, S. (1997). Education reform and accountability issues in an intergovernmental context. Publius, 27 (2), 143 166. Mintrom, M. & Vergari, S. (1998). Policy networks and innovation diffusion: The case of state education reforms. Journal of P olitics, 60 (1), 126 148.

PAGE 209

193 Miron, G., & Nelson, C. (2002). What s public about charter schools? Lessons learned about choice and accountability Thousand Oaks, CA: Corwin Press. Molnar, A. (1996). Charter s chools: The s miling f ace of d isinvestment Educational Leadership 54 (2), 9 15 Moon, M. J. (1999). The pursuit of Managerial entrepreneurship: Does organization matter? P ubli c Administration Review, 59 (1), 31 43. Morgan, G. A., Leech, N. L. Gloeckner, G. W., & Barrett, K. C. (2013). IB M SPSS for introductory statistics: Use and interpretation (5 th ed.). New York, NY: Routledge. Myers, N. R., & Cibulka, J. G. (2008) Religious faith and policy in public education: A political and historical analysis of the Christian right in American schooling. In B. S. Cooper, J. G. Cibulka, & L. D. Fusarelli (Eds.), Handbook of Education Politics and Policy (pp. 232 245). New York, N Y: Routledge. National Charter School Resource Center. (2010). November 2010: Meeting the challenge of rural charter school development Retrieved from http://www. charterschoolcenter.org/newsletter/november 2010 meeting challenge rural charter school development Nester, W. R. (2013). The age of Jackson and the art of Ameircan power, 1815 1848 Washington, D.C.: Potomac Books. North, D. C. (1990). Institutions, in stitutional change, and economic performance Cambridge, NY: Cambridge University Press. Nownes, A. (2001). Pressure and power: organized interests in American Politics Boston, MA: Houghton Mifflin. umb for variance inflation factors. Quality and Quantity, 41 (5), 673 690. findings: The role of mixed methods research. The Qualitative Report, 9 (4), 770 792. Osborne D. (1993). Reinventing government. Public Productivity & Management Review 16 (4), 349 356. Osborne, D., & Gaebler, T. (1992). Reinventing government: how the entrepreneurial spirit is transforming the public sector Reading, MA: Addison Wesley.

PAGE 210

194 Osborn e, D., & Plastrik, P. (1997). Banishing bureaucracy: The five strategies for reinventing government Reading, MA: Addison Wesley. Osguthorpe, R. (2002). Utah reading policy: Problems, processes, and participants (CIERA Report). Ann Arbor, MI: University of Michigan, Center for the Improvement of Early Reading Achievement. Ostrom, E ( 2007 ) Institutional r ational c hoice: An a ssessment of the i nstitutional a nalysis and d evelopment f ramework. In P A. Sabatier ( 2 nd e d. ) Theories of the p olicy p rocess (pp 21 64) Boulder, CO: Westview press Ostrom, E (20 11 ). Background on the institutional analysis and development framework. Policy Studies Journal, 39 (1), 7 27. R. & Berner, M. (2003). Research methods for public administrators New York, NY:Longman. Practical research methods for nonprofit and public administrator New York, NY: Longman. Peters, B. G., & Savoie, D. J. (1994). Reinvent ing Osborne and Gaebler: Lessons from the Gore commission. Canadian Public Administration, 37 (2), 302 322. Peterson, P. E., & Campbell, D. E. (2001). Introduction: A new direction in public education? I n P. E. Peterson, & D. E. Campbell (Eds.), Charters, vouchers, and p ublic e ducation (pp. 1 16). Washington, D. C.: Brookings Institution Press. Pfeffer, J., & Salancik, G. R. (1974). Organizational decision making as a political process: The case of a university budget. Administrative Science Q u arterly 19 (2), 135 151. Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations New York, NY: Harper and Row. Polsby, N. (1984). Political innovation in America New Haven, CT: Yale University Press. Portz, J. (1996). Proble m definition and policy agendas: Shaping the education agenda in Boston. Policy Studies Journal, 24 (3), 371 386. Powell, W. W. & DiMaggio, P. J. (1991). The new institutionalism in organizational analysis Chicago, IL: University of Chicago Press. Provan, K. G., & Milward, H. B. (2001). Do networks really work? A framework for evaluating public sector organizational networks. Public Administration Review, 61 (4), 414 423.

PAGE 211

195 Putnam, R. (1993). Making democracy work: Civic traditions in modern Italy Pr inceton, NJ : Princeton University. Putnam, R (1995). B ow l Journal of Democracy, 6 (1) 65 78. Putnam, R. (1995). Tuning in, Tuning out: The strange disappearance of social capital in America. PS: Political S cience and Politics, 28 (4) 664 683. Putnam, R. (2000). Bowling alone: The collapse and revival of American community New York, NY : Simon and Schuster. Rabb, J., & Winstead, D. (2003). Perspectives of the ultimate consumers: Policymakers and program ma nagers. In M. C. Lennon, & T. Corbett (Eds.), Policy into action: Implementation research and welfare reform (pp. 21 37). Washington, D.C.: Urban Institute Press. Remler, D. K., & Van Ryzin, G. G. (201 1 ). Research m ethods in p r actice: Strategies for description and causation Thousand Oaks, CA: S age Publications, Inc. Renzulli, L. A. (2005). Organizational environments and the emergence of charter schools in the United States. Sociology of Education, 78 (January), 1 26. Renzulli, L. A. & Roscigno, V. J. (2005). Charter school policy, implementation, and diffusion across the United States. Sociology of Education, 78 (October), 344 366. Renzulli, L. A. & Roscigno, V. J (2007). C h arter schools and the public good. Contexts, 6 (1), 31 36. Rivera, J., & deLeon, P. (2004). Is greener whiter yet? The sustainable slopes program and the voluntary environmental performance of Western ski areas, Policy Studies Journal, 3 2 (3), 417 437. Rivera, J., deLeon, P., & Koerber, C. (2006). Is greener whiter yet? The sustainable slopes program after five years, Policy Studies Journal, 34 (2), 195 221. Rivera, M. D., Streib, G., Willoughby, K. G. (2000). Reinventing government in council manager cities: Examining the role of city managers. Public Performance and Management Review, 24 (2), 121 132. Roberts, N. C., & King, P. J. (1991). Policy entrepreneurs: Their activity structure and function in the policy process. Journal of Public Administration Research and Theory, 1 (2), 147 175.

PAGE 212

196 Robert s, N. C., & King, P. J. (1996). Transforming public policy: Dynamics of policy entrepreneurship and innovation San Francisco, CA: Jossey Bass Publishers. Rogers, E. M. (1962). Diffusion of innovation New York, NY : Free Press. Rogers, E. M (2003). Dif fusion of innovation (5 th ed.). New York, NY : Free Press. Roth, W. D., & Mehta, J. D. (2002). The Rahomon effect: Combining positivist and interpretivist approaches in the analysis of contested events, Sociological Methods and Research, 31 (2), 131 173. R uhil, A. V. S., Schneider, M. Teske, P., & Ji, B. (1999). Institutions and reform: Reinventing local government. Urban Affairs Review, 34 (3), 433 455. Rusaw, A. C. (1997). Reinventing local government: A case study of organizational change therough commun ity learning. Public Administration Quarterly, 20 (4), 419 432. Russell, G. D., & Waste, R. J. (1998). The limits of reinventing government. American Review of Public Administration, 28 (4), 325 346. Sabatier, P. A. (2007). The need for better theories. In P.A. Sabatier ( 2 nd e d.), Theories of the policy process (pp. 3 17 ). Boulder, CO: Westview press Saint Martin, D. (2001). How the reinventing government movement in public administration was exported from the U.S. to other countries. International Journ al of Public Administration, 24 (6), 573 604. Schachter, H. L. (1995). Reinveinting government or reinventing ourselves: Two models for improving government performance. Public Administration Review, 55 (6), 530 537. Schneider, M., Teske, P., & Mintrom, M. (1995). Public entrepreneurs: Agents for change in American government Princeton, NJ: Princeton University Press. Schneider, M Teske, P., & Marschall, M. (2000). Choosing schools: Consumer choice and the quality of American schools. Princeton, NJ: Pri nceton University Press. Scott, J. (2000). Social network analysis: A handbook (2 nd ed.). Thousand Oaks, CA: Sage Publications. Smith, J ., Wohlstetter, P., Farrell, C. C., & Nayfack, M. B. (2011). Beyond ideological warfare: The maturation of research on charter schools. Journal of School Choice, 5 444 507.

PAGE 213

197 Smith, K. B., & Meier, K. J. (1995). Public choice in education: Markets and the demand for quality education. P olitical Research Quarterly, 4 8 (3), 461 478. Smith, K., & Petersen, J. (2008). Social purspose capital markets in K 12. In F. M. Hess (Ed ), The future of educational entrepreneurship: possibilities for school reform (pp. 89 117). Cambridge, MA: Harvard Education Press. Spring, J. (2 005). Political agendas for education: From the religious right to the G reen P arty (3 rd ed.). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Spring, J. (2010). Political agendas for education: From change we can believe in to putting America first (4 th ed .). New York, NY: Routledge. Stedman, L. C. (1994). The Sadia report and U.S. achievement: An assessment. Journal of Educational Research, 87 (3), 133 146. Stewart, J., Hedge, D. M., & Lester, J. P. (2008). Public policy: An evolutionary approach (3 rd ed .). Boston, MA: Thomson Wadsworth. Stillman, R. C. (1991). Preface to public administration: A search for themes and direction New York Stillman, R. C. (1996). The American bureaucracy: The core of modern government Chicago, IL: Nelson Hall Inc. Stine, R. A. (1995). Graphical interpretation of variance inflation factors. American Statistician, 49 ( 1 ), 53 56 Stoddard, C., & Corcoran, S. P. (2007). The political economy of school choice: Support for charter schools across states and school districts. Journal of Urban Economics, 62 (1), 27 54. Teske, P. (199 0 ). After Divestiture: The p olitical e conomy of s tate t elecommunications r egulation Albany, N Y: State University of New York Press Teske, P. (1991). Interests and institutions on state regulation. American J ournal of Political Science, 35 (1), 139 154. Teske, P. (2004). Regulation in the States Washington, D.C.: Brookings Institution Press. Te sk e P, Fitzpatrick, J & Kaplan, G. ( 2006 ) The i nformation g ap? Review of Policy Research 23 (5) 969 9 81.

PAGE 214

198 Teske, P., & Williamson, A. (2006). Entrepreneurs at work. In F. M. Hess (Ed.), Educational entrepreneurship (pp. 45 62). Cambridge, MA: Harvard Education Press. Tjora, A. H. (2006). Writing small discoveries: An exploration of fresh observers observations. Qualitative Research, 6 (4), 429 451. dimensions of managerial networking. American Revi ew of Public Administration, 43 (3), 251 272. Tucker, A. M. & Lauber, W. F. (1995). School choice programs: What s happening in the states? Washington, D.C.: The Heritage Foundation. Tyran, J., & Sausgruber, R. (2005). The diffusion of policy innovations: An experimental investigation. Journal of Evolutionary Economics, 15 (4), 423 442. United States Census Bureau (2012). State & County QuickFacts Retrived from http:// quickfacts.cen sus.gov/qfd/states/08000.html Urban, W. J., & Wagoner, J. L. (2009). American educaton: A history New York, NY: Routledge. Vergari, S. (1999). Charter schools: A primer on the issues. Education and Urban Society, 31 (4), 389 405. Vergari, S. (2000). Mo rality politics and educational policy: The abstinence only sex education grant. Educational Policy, 14 (2), 290 231. Vergari, S. (2002). T he charter school landscape Pittsburgh, PA: University of Pittsburgh Press. Vergari, S. (2007). The politics of charter schools. Educational Policy, 21 (1), 15 39. Vito, G. F., & Kunselman, J. (2000). Reinventing government: The views of police middle managers. Police Quarterly, 3 (3), 315 330. Vollmer, R. J. (2010). Schools cannot do it alone: Building public suppo public schools Fairfield, IA: Enlightenment Press. Wagner, W. E. (2013). Using IBM SPSS statistics for research methods and social science statistics (4 th ed.). Thousand Oaks, CA: Sage Publications, Inc. Walberg, H. J. (2007). School Ch oice: The findings Washington, D.C.: Cato Institute.

PAGE 215

199 Walker, J. L. (1969). The diffusion of innovations among the American states. American Political Science Review, 63 (3), 880 899. Walker, J. L. (1981). The diffusion of knowledge, policy communities and agenda setting: The relationship of knowledge and power. In J. E. Tropman, M. J. Dluhy, & R. M. Lind (Eds.), New strategic perspectives on social policy (pp. 75 96). New York, NY : Pergamon. Walt, G., Shiffman, J., Schneider, H., Murray, S. F., Brugha, R., & Gilson, L. (2008). challenges. Health Policy and Planning, 23 (5), 308 317. American Behavioral Scientist 30 (6 ) 612 631. Weible, C. M., Heikkila, T., deLeon, P., & Sabatier, P. A. (2012). Understanding and influencing the policy process. Policy Science, 45 1 21. Wheelan, C. (2011). Introduction to public policy New York, NY: W. W. Norton and Company. Wirt, F., & Kirst, M. (1997). The political dynamics of American education. Berkeley, CA: McCutchan. Witt e J. F. (2000). voucher program Princeton, NJ: Princeton University Press. Witt e J. F., Schlomer, P. A., & Shober, A. F. (2007). Going charter? A study of school district competition in Wisconsin. Peabody Journal of Education, 82 (2 3), 410 439. Wong, K. K., & Klopott, S. (2009). Politics and governance in charter schools. In M. Berends M. G. Springer, D. Ballou, & H. J. Walberg (Eds ) Handbook of research on school choice (pp. 1 15 1 3 5 ). New York, NY: Routledge. Wong, K. K., & Langevin, W. E. (2005). The diffusion of governance reform in American public education. Nashville, TN: National Center on School Choice. Wong, K. K., & Langevin, W. E. (2007). Policy expansion of school choice in the American States. P eabody Journal of Education, 82 (2 3), 440 472. Wong, K. K., & Shen, F. X. (2002). Politics of state led reform in education: Market competition and electoral dynamics. Educational Policy, 16 (1), 161 192.

PAGE 216

200 Zajac, G. (1997). Reinventing government and reaffirming ethics: Implications for organizational development in the public service. Public Administration Quarterly, 20 (4), 385 404. Ziebarth, T. (2005). Peaks and valleys: Colorado s charter school landscape Retrieved from http:// www.ppionline.org/documents/Colorado_Charter_1220.pdf

PAGE 217

201 APPENDIX A PPROVAL L ETTER

PAGE 218

202 APPENDIX B S URVEY I TO SUPERINTENDENTS CONSENT You are invited to take part in Colorado educational reform research. This study is being conducted by John (Jeongho) Lee, a doctoral student in the School of Public Affairs at the University of Colorado Denver. The primary go al of this study is to examine why and under what circumstances school districts in Colorado implement innovative education policies State level research on this topic has been studied since the mid 1990s, but, this topic has not been studied much at the school district level. Moreover, Colorado does not specifically have any studies on this topic. The researcher expects this study to add to our knowledge about Colorado education reforms. To accomplish this goal, the researcher needs to conduct a survey. P lease carefully consider your responses to the survey so that this study can contribute to an improved Colorado educational environment for our students. These survey questionnaires aim to collect information that cannot be obtained from existing data. Yo ur responses would be greatly appreciated in answering the questions on dissertation. This survey should only take about 15 to 20 minutes. If you have any questions about this st udy, you can contact the researcher through either e mail ( Jeong.Lee@ucdenver.edu ) or phone (850 339 5372). You can also e mail the Paul.Teske@ucdenver.edu If you have any questions about your rights as a participant in a research study, you can call the Colorado Multiple Institutional Review Board (COMIRB) at 303 724 1055. Your name will rema in anonymous on data and your responses will not been revealed to other researchers. Thanks for your response and time, Statement of Consent : I have read the above information. I consent to participate in this study. Your Signature ________________________ Date ________ ______ _____________ Your Name (printed) ________ ______ ___________________ _____________________

PAGE 219

203 Survey Questions Q. 1 Please write down the name of your school district. __________________ ____ __. Q. 2. Pertaining to the teacher union: What percentages of K 12 teachers in your school districts are members of the Colorad o Ed ucation Association (C EA) or the American Federation of Teachers (AFT)? Please circle below. 1) 0 10% 2) 11 20% 3) 21 30% 4) 31 40% 5) 41% 50% 6) 51 60% 7) 61 70% 8) 71 80% 9) 81 90 % 10) 91 100% Q.3. Pertaining to alternative innovations: Q. 3 1) How many K 12 magnet schools are there in your school district? Please write down the number of them. ____ _______. Q. 3 2) Do you agree that the open enrollment program has been implemented well in your school district? Please circle below. (1) strongly disagree (2) disagree (3) neither disagree nor agree (4) agree (5) strongly agree Q. 3 3) Do you agree that the home schooling program has been implemented well in your school district? Please circle below. (1) strongly disagree (2) disagree (3) neither disagree nor agree (4) agree (5) strongly agree Q. 4 Pertaining to the goal of your own school district : Does your school district have a strong goal to use school choice programs to reform your K 12 school system? (1) strongly disagree (2) disagree (3) neither disagree nor agree (4) agree (5) strongly agree Please go to the next page

PAGE 220

204 Q. 5 Pertaining to policy entrepreneurs: Q. 5 1) Teske & Williamson (2006) define educational entrepreneurs as individuals seeking to instigate change in the public education system that will disrupt, transform, or radically alter the way education is provided. D o you agree that any educational entrepreneurs strongly influence the consideration or implementation of school choice programs as public educational innovations and reforms in your school district? Please circle below. (1) strongly disagree (2) disagree (3) neither disagree nor agree (4) agree (5) strongly agree Q. 5 2) How many educational entrepreneurs do you think are for the consideration or implementation of school choice programs as public educational innovations and reforms in your school district? Please write down the number of them __________ Q. 6 Pertaining to organized interest groups: Q. 6 1) Scholars indicate that organized interest groups influence the school choice movement as educational innovations and reforms. In the case of your school district, do you agree with this? Please circle below (1) strongly disagree (2) disagree (3) neither disagree nor agree (4) agree (5) strongly agree Q. 6 2) How many organized interest groups do you think positively affect (that is, support school choice programs) you reforms? Please write down the number of them __ ________. Q. 6 3) How many organized interest groups do you think negatively affect (that is, reforms? Please write down the number of them __ ________. Please go to the next page

PAGE 221

205 Q. 7 Pertaining to the frequency of sharing information: Q. 7 1) Below is a list of 1 2 state level public organizations that are involved broadly in For each organizat ion, please indicate the frequency of contacts ( e.g. meetings, phone calls, or e mails ) your own school district had with the organization in 2011 relating to school choice programs as educational innovations and reforms. ( 0 ) = No contact ( 1 ) = 1 contact (2) = 2 4 contacts (3) = 5 7 contacts ( 4 ) = 8 1 0 contacts (5) = more than 1 0 contacts PUBLIC ORGANIZATION FREQUENCY 1 BEST Board (Public School Capital Construction Assistance) 0 1 2 3 4 5 2 Boards of Cooperative Educational Services 0 1 2 3 4 5 3 Colorado Association of School Executives 0 1 2 3 4 5 4 Colorado Education Association 0 1 2 3 4 5 5 Colorado Charter School Institute 0 1 2 3 4 5 6 Colorado Department of Education 0 1 2 3 4 5 7 Colorado Department of Higher Education 0 1 2 3 4 5 8 Colorado Children's Campaign 0 1 2 3 4 5 9 Colorado League of Charter Schools 0 1 2 3 4 5 10 Colorado Legacy Foundation 0 1 2 3 4 5 11 Colorado State Board of Education 0 1 2 3 4 5 12 Education Leadership Council 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 222

206 Q. 7 2) Pertaining to networks between your own school district and other Colorado school districts : For each school district, please indicate the frequency of contacts ( e.g. meetings, phone calls, or e mails ) your own school district had with the school district in 2011 in relating to school choice programs as educational innovations and reforms (Please skip your own school district). ( 0 ) = No contact ( 1 ) = 1 contact (2) = 2 4 contacts (3) = 5 7 contacts ( 4 ) = 8 1 0 contacts (5) = more than 1 0 contacts Please go to the next page

PAGE 223

207 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 1 Academy 20 (El Paso) 0 1 2 3 4 5 2 Adams 12 Five Star Schools (Adams) 0 1 2 3 4 5 3 Adams County 14 (Adams) 0 1 2 3 4 5 4 Adams Arapahoe 28J (Arapahoe) 0 1 2 3 4 5 5 Agate 300 (Elbert) 0 1 2 3 4 5 6 Aguilar Reorganized 6 (Las Animas) 0 1 2 3 4 5 7 Akron R 1 (Washington) 0 1 2 3 4 5 8 Alamosa RE 11J (Alamosa) 0 1 2 3 4 5 9 Archuleta County 50 JT (Archuleta) 0 1 2 3 4 5 10 Arickaree R 2 (Washington) 0 1 2 3 4 5 11 Arriba Flagler C 20 (Kit Carson) 0 1 2 3 4 5 12 Aspen 1 (Pitkin) 0 1 2 3 4 5 13 Ault Highland RE 9 (Weld) 0 1 2 3 4 5 14 Bayfield 10 JT R (La Plata) 0 1 2 3 4 5 15 Bennett 29J (Adams) 0 1 2 3 4 5 16 Bethune R 5 (Kit Carson) 0 1 2 3 4 5 17 Big Sandy 100J (Elbert) 0 1 2 3 4 5 18 Boulder Valley RE 2 (Boulder) 0 1 2 3 4 5 19 Branson Reorganized 82 (Las Animas) 0 1 2 3 4 5 20 Briggsdale RE 10 (Weld) 0 1 2 3 4 5 21 Brighton 27J (Adams) 0 1 2 3 4 5 22 Brush RE 2(J) (Morgan) 0 1 2 3 4 5 23 Buena Vista R 31 (Chaffee) 0 1 2 3 4 5 24 Buffalo RE 4 (Logan) 0 1 2 3 4 5 25 Burlington RE 6J (Kit Carson) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 224

208 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 26 Byers 32J (Arapahoe) 0 1 2 3 4 5 27 Calhan RJ 1 (El Paso) 0 1 2 3 4 5 28 Campo RE 6 (Baca) 0 1 2 3 4 5 29 Canon City RE 1 (Fremont) 0 1 2 3 4 5 30 Centennial R 1 (Costilla ) 0 1 2 3 4 5 31 Center 26 JT (Saguache ) 0 1 2 3 4 5 32 Cheraw 31 (Otero) 0 1 2 3 4 5 33 Cherry Creek 5 (Arapahoe) 0 1 2 3 4 5 34 Cheyenne County RE 5 (Cheyenne) 0 1 2 3 4 5 35 Cheyenne Mountain 12 (El Paso) 0 1 2 3 4 5 36 Clear Creek RE 1 (Clear Creek) 0 1 2 3 4 5 37 Colorado Springs 11 (El Paso) 0 1 2 3 4 5 38 Cotopaxi RE 3 (Fremont) 0 1 2 3 4 5 39 Creede School District (Mineral) 0 1 2 3 4 5 40 Cripple Creek Victor RE 1 (Teller) 0 1 2 3 4 5 41 Crowley County RE 1 J (Crowley ) 0 1 2 3 4 5 42 Custer County School District C 1 (Custer) 0 1 2 3 4 5 43 De Beque 49JT (Mesa ) 0 1 2 3 4 5 44 Deer Trail 26J (Arapahoe) 0 1 2 3 4 5 45 Del Norte C 7 (Rio Grande) 0 1 2 3 4 5 46 Delta County 50 (J) (Delta) 0 1 2 3 4 5 47 Denver County 1 (Denver) 0 1 2 3 4 5 48 Dolores County RE NO.2 (Dolores) 0 1 2 3 4 5 49 Dolores RE 4A (Montezuma) 0 1 2 3 4 5 50 Douglas County RE 1 (Douglas) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 225

209 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 51 Durango 9 R (La Plata) 0 1 2 3 4 5 52 Eads RE 1 (Kiowa) 0 1 2 3 4 5 53 Eagle County RE 50 (Eagle) 0 1 2 3 4 5 54 East Grand 2 (Grand) 0 1 2 3 4 5 55 East Otero R 1 (Otero) 0 1 2 3 4 5 56 Eaton RE 2 (Weld) 0 1 2 3 4 5 57 Edison 54 JT (El Paso) 0 1 2 3 4 5 58 Elbert 200 (Elbert) 0 1 2 3 4 5 59 Elizabeth C 1 (Elbert) 0 1 2 3 4 5 60 Ellicott 22 (El Paso) 0 1 2 3 4 5 61 Englewood 1 (Arapahoe) 0 1 2 3 4 5 62 Falcon 49 (El Paso) 0 1 2 3 4 5 63 Florence RE 2 (Fremont) 0 1 2 3 4 5 64 Fort Morgan RE 3 (Morgan) 0 1 2 3 4 5 65 Fountain 8 (El Paso) 0 1 2 3 4 5 66 Fowler R 4J (Otero) 0 1 2 3 4 5 67 Frenchman RE 3 (Logan) 0 1 2 3 4 5 68 Garfield 16 (Garfield) 0 1 2 3 4 5 69 Garfield RE 2 (Garfield ) 0 1 2 3 4 5 70 Genoa Hugo C113 (Lincoln) 0 1 2 3 4 5 71 Gilpin County RE 1 (Gilpin) 0 1 2 3 4 5 72 Granada RE 1 (Prowers) 0 1 2 3 4 5 73 Greeley 6 (Weld) 0 1 2 3 4 5 74 Gunnison Watershed RE1J (Gunnison) 0 1 2 3 4 5 75 Hanover 28 (El Paso) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 226

210 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contact s. SCHOOL DISTRICT (County) FREQUENCY 76 Harrison 2 (El Paso) 0 1 2 3 4 5 77 Haxtun RE 2J (Phillips ) 0 1 2 3 4 5 78 Hayden RE 1 (Routt) 0 1 2 3 4 5 79 Hinsdale County RE 1 (Hinsdale) 0 1 2 3 4 5 80 Hi Plains R 23 (Kit Carson) 0 1 2 3 4 5 81 Hoehne Reorganized 3 (Las Animas) 0 1 2 3 4 5 82 Holly RE 3 (Prowers) 0 1 2 3 4 5 83 Holyoke RE 1J (Phillips) 0 1 2 3 4 5 84 Huerfano RE 1 (Huerfano) 0 1 2 3 4 5 85 Idalia RJ 3 (Yuma) 0 1 2 3 4 5 86 Ignacio 11 JT (La Plata) 0 1 2 3 4 5 87 Jefferson County R 1 (Jefferson) 0 1 2 3 4 5 88 Johnstown Milliken RE 5J (Weld) 0 1 2 3 4 5 89 Julesburg RE 1 (Sedgwick) 0 1 2 3 4 5 90 Karval RE 23 (Lincoln) 0 1 2 3 4 5 91 Keenesburg RE 3 (J) (Weld) 0 1 2 3 4 5 92 Kim Reorganized 88 (Las Animas) 0 1 2 3 4 5 93 Kiowa C 2 (Elbert) 0 1 2 3 4 5 94 Kit Carson R 1 (Cheyenne) 0 1 2 3 4 5 95 La Veta RE 2 (Huerfano) 0 1 2 3 4 5 96 Lake County R 1 (Lake) 0 1 2 3 4 5 97 Lamar RE 2 (Prowers) 0 1 2 3 4 5 98 Las Animas RE 1 (Bent) 0 1 2 3 4 5 99 Lewis Palmer 38 (El Paso) 0 1 2 3 4 5 100 Liberty J 4 (Yuma) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 227

211 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 101 Limon RE 4J (Lincoln) 0 1 2 3 4 5 102 Littleton 6 (Arapahoe) 0 1 2 3 4 5 103 Lone Star 101 (Washington) 0 1 2 3 4 5 104 Mancos RE 6 (Montezuma) 0 1 2 3 4 5 105 Manitou Springs 14 (El Paso) 0 1 2 3 4 5 106 Manzanola 3J (Otero) 0 1 2 3 4 5 107 Mapleton 1 (Adams) 0 1 2 3 4 5 108 MC Clave RE 2 (Bent) 0 1 2 3 4 5 109 Meeker RE1 (Rio Blanco) 0 1 2 3 4 5 110 Mesa County Valley 51 (Mesa) 0 1 2 3 4 5 111 Miami/Yoder 60 JT (El Paso) 0 1 2 3 4 5 112 Moffat 2 (Saguache) 0 1 2 3 4 5 113 Moffat County RE:No 1 (Moffat) 0 1 2 3 4 5 114 Monte Vista C 8 (Rio Grande) 0 1 2 3 4 5 115 Montezuma Cortez RE 1 (Montezuma) 0 1 2 3 4 5 116 Montrose County RE 1J (Montrose) 0 1 2 3 4 5 117 Mountain Valley RE 1 (Saguache) 0 1 2 3 4 5 118 North Conejos RE 1J (Conejos) 0 1 2 3 4 5 119 North Park R 1 (Jackson) 0 1 2 3 4 5 120 Norwood R 2J (San Miguel) 0 1 2 3 4 5 121 Otis R 3 (Washington) 0 1 2 3 4 5 122 Ouray R 1 (Ouray) 0 1 2 3 4 5 123 Park (Estes Park) R 3 (Larimer) 0 1 2 3 4 5 124 Park County RE 2 (Park) 0 1 2 3 4 5 125 Pawnee RE 12 (Weld) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 228

212 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 126 Peyton 23 JT (El Paso) 0 1 2 3 4 5 127 Plainview RE 2 (Kiowa) 0 1 2 3 4 5 128 Plateau RE 5 (Logan) 0 1 2 3 4 5 129 Plateau Valley 50 (Mesa) 0 1 2 3 4 5 130 Platte Canyon 1 (Park) 0 1 2 3 4 5 131 Platte Valley RE 3 (Sedgwick) 0 1 2 3 4 5 132 Platte Valley RE 7 (Weld) 0 1 2 3 4 5 133 Poudre R 1 (Larimer) 0 1 2 3 4 5 134 Prairie RE 11 (Weld) 0 1 2 3 4 5 135 Primero Reorganized 2 (Las Animas) 0 1 2 3 4 5 136 Pritchett RE 3 (Baca) 0 1 2 3 4 5 137 Pueblo City 60 (Pueblo) 0 1 2 3 4 5 138 Pueblo County Rural 70 (Pueblo ) 0 1 2 3 4 5 139 Rangely RE 4 (Rio Blanco) 0 1 2 3 4 5 140 Ridgway R 2 (Ouray) 0 1 2 3 4 5 141 Roaring Fork RE 1 (Garfield) 0 1 2 3 4 5 142 Rocky Ford R 2 (Otero) 0 1 2 3 4 5 143 Salida R 32 (Chaffee) 0 1 2 3 4 5 144 Sanford 6J (Conejos) 0 1 2 3 4 5 145 Sangre De Cristo RE 22J (Alamosa) 0 1 2 3 4 5 146 Sargent RE 33J (Rio Grande) 0 1 2 3 4 5 147 Sheridan 2 (Arapahoe) 0 1 2 3 4 5 148 Sierra Grande R 30 (Costilla) 0 1 2 3 4 5 149 Silverton 1 (San Juan) 0 1 2 3 4 5 150 South Conejos RE 10 (Conejos) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 229

213 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 151 South Routt RE 3 (Routt) 0 1 2 3 4 5 152 Springfield RE 4 (Baca) 0 1 2 3 4 5 153 St. Vrain Valley RE 1J (Boulder) 0 1 2 3 4 5 154 Steamboat Springs RE 2 (Routt) 0 1 2 3 4 5 155 Strasburg 31 J (Adams) 0 1 2 3 4 5 156 Stratton R 4 (Kit Carson) 0 1 2 3 4 5 157 Summit RE 1 (Summit) 0 1 2 3 4 5 158 Swink 33 (Otero) 0 1 2 3 4 5 159 Telluride R 1 (San Miguel) 0 1 2 3 4 5 160 Thompson R 2J (Larimer) 0 1 2 3 4 5 161 Trinidad 1 (Las Animas) 0 1 2 3 4 5 162 Valley RE 1 (Logan) 0 1 2 3 4 5 163 Vilas RE 5 (Baca) 0 1 2 3 4 5 164 Walsh RE 1 (Baca) 0 1 2 3 4 5 165 Weld County RE 1 (Weld) 0 1 2 3 4 5 166 Weld County S/D RE 8 (Weld) 0 1 2 3 4 5 167 Weldon Valley RE 20(J) (Morgan) 0 1 2 3 4 5 168 West End RE 2 (Montrose) 0 1 2 3 4 5 169 West Grand 1 JT (Grand) 0 1 2 3 4 5 170 Westminster 50 (Adams) 0 1 2 3 4 5 171 Widefield 3 (El Paso) 0 1 2 3 4 5 172 Wiggins RE 50(J) (Morgan) 0 1 2 3 4 5 173 Wiley RE 13 JT (Prowers) 0 1 2 3 4 5 174 Windsor RE 4 (Weld) 0 1 2 3 4 5 175 Woodland Park RE 2 (Teller) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 230

214 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 176 Woodlin R 104 (Washington) 0 1 2 3 4 5 177 Wray RD 2 (Yuma) 0 1 2 3 4 5 178 Yuma 1 (Yuma) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Thank you so much for your responses

PAGE 231

215 APPENDIX C S URVEY II TO ORGANIZATIONS CONSENT You are invited to take part in Colorado educational reform research. This study is being conducted by John (Jeongho) Lee, a doctoral student in the School of Public Affairs at the University of Colorado Denver. The primary goal of this study is to examine why and under what circumstances school districts in Colorado implement innovative education policies State level research on this topic has been studied since the mid 1990s, but, this topic has not been studied much at the school district level. Moreove r, Colorado does not specifically have any studies on this topic. The researcher expects this study to add to our knowledge about Colorado education reforms. To accomplish this goal, the researcher needs to conduct a survey. Please carefully consider your responses to the survey so that this study can contribute to an improved Colorado educational environment for our students. These survey questionnaires aim to collect information that cannot be obtained from existing data. Your responses would be greatly appreciated in answering the questions on dissertation. This survey should only take about 15 to 20 minutes. If you have any questions about this study, you can contact the resea rcher through either e mail ( Jeong.Lee@ucdenver.edu ) or phone (850 339 5372). You can also e mail the Paul.Teske@ucdenver. edu If you have any questions about your rights as a participant in a research study, you can call the Colorado Multiple Institutional Review Board (COMIRB) at 303 724 1055. Your name will remain anonymous on data and your responses will not been revealed to other researchers. Thanks for your response and time, Statement of Consent : I have read the above information. I consent to participate in this study. Your Signature ________________________ Date ________________ ______ _____ Your Name (printed) ____________________________ ____________________ ______

PAGE 232

216 Survey Questions Q.1. Please write down the name of your organization. _____________________. Q.2. Pertaining to the frequency of contacts between your own organization and the other 1 1 public organizations at the state level, Below is a list of 1 2 state level public organizations that are involved broadly in Co indicate the frequency of contacts ( e.g. meetings, phone calls, or e mails ) your own organization had with the organization in 2011 relating to educational innovations and ref orms (Please skip your own organization). ( 0 ) = No contact ( 1 ) = 1 contact (2) = 2 4 contacts (3) = 5 7 contacts ( 4 ) = 8 1 0 contacts (5) = more than 1 0 contacts Please go to the next page

PAGE 233

217 Note: (0)= No contact; (1)=1 contact; (2)=2 4 contacts; (3)=5 7 contacts; (4)=8 10 contacts; (5)=more than 10 contacts. PUBLIC ORGANIZATION FREQUENCY 1 BEST Board (Public School Capital Construction Assistance) 0 1 2 3 4 5 2 Boards of Cooperative Educational Services 0 1 2 3 4 5 3 Colorado Association of School Executives 0 1 2 3 4 5 4 Colorado Education Association 0 1 2 3 4 5 5 Colorado Charter School Institute 0 1 2 3 4 5 6 Colorado Department of Education 0 1 2 3 4 5 7 Colorado Department of Higher Education 0 1 2 3 4 5 8 Colorado Children's Campaign 0 1 2 3 4 5 9 Colorado League of Charter Schools 0 1 2 3 4 5 10 Colorado Legacy Foundation 0 1 2 3 4 5 11 Colorado State Board of Education 0 1 2 3 4 5 12 Education Leadership Council 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Q. 3. Pertaining to policy networks between your own organization and other Colorado school districts, For each school district, please indicate the frequency of contacts ( e.g. meetings, phone calls, or e mails ) your own organization had with the school district in 2011 in relating to educational innovations and reforms. ( 0 ) = No contact ( 1 ) = 1 contact (2) = 2 4 contacts (3) = 5 7 contacts ( 4 ) = 8 1 0 contacts (5) = more than 1 0 contacts Please go to the next page

PAGE 234

218 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 1 Academy 20 (El Paso) 0 1 2 3 4 5 2 Adams 12 Five Star Schools (Adams) 0 1 2 3 4 5 3 Adams County 14 (Adams) 0 1 2 3 4 5 4 Adams Arapahoe 28J (Arapahoe) 0 1 2 3 4 5 5 Agate 300 (Elbert) 0 1 2 3 4 5 6 Aguilar Reorganized 6 (Las Animas) 0 1 2 3 4 5 7 Akron R 1 (Washington) 0 1 2 3 4 5 8 Alamosa RE 11J (Alamosa) 0 1 2 3 4 5 9 Archuleta County 50 JT (Archuleta) 0 1 2 3 4 5 10 Arickaree R 2 (Washington) 0 1 2 3 4 5 11 Arriba Flagler C 20 (Kit Carson) 0 1 2 3 4 5 12 Aspen 1 (Pitkin) 0 1 2 3 4 5 13 Ault Highland RE 9 (Weld) 0 1 2 3 4 5 14 Bayfield 10 JT R (La Plata) 0 1 2 3 4 5 15 Bennett 29J (Adams) 0 1 2 3 4 5 16 Bethune R 5 (Kit Carson) 0 1 2 3 4 5 17 Big Sandy 100J (Elbert) 0 1 2 3 4 5 18 Boulder Valley RE 2 (Boulder) 0 1 2 3 4 5 19 Branson Reorganized 82 (Las Animas) 0 1 2 3 4 5 20 Briggsdale RE 10 (Weld) 0 1 2 3 4 5 21 Brighton 27J (Adams) 0 1 2 3 4 5 22 Brush RE 2(J) (Morgan) 0 1 2 3 4 5 23 Buena Vista R 31 (Chaffee) 0 1 2 3 4 5 24 Buffalo RE 4 (Logan) 0 1 2 3 4 5 25 Burlington RE 6J (Kit Carson) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 235

219 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHO OL DISTRICT (County) FREQUENCY 26 Byers 32J (Arapahoe) 0 1 2 3 4 5 27 Calhan RJ 1 (El Paso) 0 1 2 3 4 5 28 Campo RE 6 (Baca) 0 1 2 3 4 5 29 Canon City RE 1 (Fremont) 0 1 2 3 4 5 30 Centennial R 1 (Costilla ) 0 1 2 3 4 5 31 Center 26 JT (Saguache ) 0 1 2 3 4 5 32 Cheraw 31 (Otero) 0 1 2 3 4 5 33 Cherry Creek 5 (Arapahoe) 0 1 2 3 4 5 34 Cheyenne County RE 5 (Cheyenne) 0 1 2 3 4 5 35 Cheyenne Mountain 12 (El Paso) 0 1 2 3 4 5 36 Clear Creek RE 1 (Clear Creek) 0 1 2 3 4 5 37 Colorado Springs 11 (El Paso) 0 1 2 3 4 5 38 Cotopaxi RE 3 (Fremont) 0 1 2 3 4 5 39 Creede School District (Mineral) 0 1 2 3 4 5 40 Cripple Creek Victor RE 1 (Teller) 0 1 2 3 4 5 41 Crowley County RE 1 J (Crowley ) 0 1 2 3 4 5 42 Custer County School District C 1 (Custer) 0 1 2 3 4 5 43 De Beque 49JT (Mesa ) 0 1 2 3 4 5 44 Deer Trail 26J (Arapahoe) 0 1 2 3 4 5 45 Del Norte C 7 (Rio Grande) 0 1 2 3 4 5 46 Delta County 50 (J) (Delta) 0 1 2 3 4 5 47 Denver County 1 (Denver) 0 1 2 3 4 5 48 Dolores County RE NO.2 (Dolores) 0 1 2 3 4 5 49 Dolores RE 4A (Montezuma) 0 1 2 3 4 5 50 Douglas County RE 1 (Douglas) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 236

220 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 51 Durango 9 R (La Plata) 0 1 2 3 4 5 52 Eads RE 1 (Kiowa) 0 1 2 3 4 5 53 Eagle County RE 50 (Eagle) 0 1 2 3 4 5 54 East Grand 2 (Grand) 0 1 2 3 4 5 55 East Otero R 1 (Otero) 0 1 2 3 4 5 56 Eaton RE 2 (Weld) 0 1 2 3 4 5 57 Edison 54 JT (El Paso) 0 1 2 3 4 5 58 Elbert 200 (Elbert) 0 1 2 3 4 5 59 Elizabeth C 1 (Elbert) 0 1 2 3 4 5 60 Ellicott 22 (El Paso) 0 1 2 3 4 5 61 Englewood 1 (Arapahoe) 0 1 2 3 4 5 62 Falcon 49 (El Paso) 0 1 2 3 4 5 63 Florence RE 2 (Fremont) 0 1 2 3 4 5 64 Fort Morgan RE 3 (Morgan) 0 1 2 3 4 5 65 Fountain 8 (El Paso) 0 1 2 3 4 5 66 Fowler R 4J (Otero) 0 1 2 3 4 5 67 Frenchman RE 3 (Logan) 0 1 2 3 4 5 68 Garfield 16 (Garfield) 0 1 2 3 4 5 69 Garfield RE 2 (Garfield ) 0 1 2 3 4 5 70 Genoa Hugo C113 (Lincoln) 0 1 2 3 4 5 71 Gilpin County RE 1 (Gilpin) 0 1 2 3 4 5 72 Granada RE 1 (Prowers) 0 1 2 3 4 5 73 Greeley 6 (Weld) 0 1 2 3 4 5 74 Gunnison Watershed RE1J (Gunnison) 0 1 2 3 4 5 75 Hanover 28 (El Paso) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 237

221 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 76 Harrison 2 (El Paso) 0 1 2 3 4 5 77 Haxtun RE 2J (Phillips ) 0 1 2 3 4 5 78 Hayden RE 1 (Routt) 0 1 2 3 4 5 79 Hinsdale County RE 1 (Hinsdale) 0 1 2 3 4 5 80 Hi Plains R 23 (Kit Carson) 0 1 2 3 4 5 81 Hoehne Reorganized 3 (Las Animas) 0 1 2 3 4 5 82 Holly RE 3 (Prowers) 0 1 2 3 4 5 83 Holyoke RE 1J (Phillips) 0 1 2 3 4 5 84 Huerfano RE 1 (Huerfano) 0 1 2 3 4 5 85 Idalia RJ 3 (Yuma) 0 1 2 3 4 5 86 Ignacio 11 JT (La Plata) 0 1 2 3 4 5 87 Jefferson County R 1 (Jefferson) 0 1 2 3 4 5 88 Johnstown Milliken RE 5J (Weld) 0 1 2 3 4 5 89 Julesburg RE 1 (Sedgwick) 0 1 2 3 4 5 90 Karval RE 23 (Lincoln) 0 1 2 3 4 5 91 Keenesburg RE 3 (J) (Weld) 0 1 2 3 4 5 92 Kim Reorganized 88 (Las Animas) 0 1 2 3 4 5 93 Kiowa C 2 (Elbert) 0 1 2 3 4 5 94 Kit Carson R 1 (Cheyenne) 0 1 2 3 4 5 95 La Veta RE 2 (Huerfano) 0 1 2 3 4 5 96 Lake County R 1 (Lake) 0 1 2 3 4 5 97 Lamar RE 2 (Prowers) 0 1 2 3 4 5 98 Las Animas RE 1 (Bent) 0 1 2 3 4 5 99 Lewis Palmer 38 (El Paso) 0 1 2 3 4 5 100 Liberty J 4 (Yuma) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 238

222 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 101 Limon RE 4J (Lincoln) 0 1 2 3 4 5 102 Littleton 6 (Arapahoe) 0 1 2 3 4 5 103 Lone Star 101 (Washington) 0 1 2 3 4 5 104 Mancos RE 6 (Montezuma) 0 1 2 3 4 5 105 Manitou Springs 14 (El Paso) 0 1 2 3 4 5 106 Manzanola 3J (Otero) 0 1 2 3 4 5 107 Mapleton 1 (Adams) 0 1 2 3 4 5 108 MC Clave RE 2 (Bent) 0 1 2 3 4 5 109 Meeker RE1 (Rio Blanco) 0 1 2 3 4 5 110 Mesa County Valley 51 (Mesa) 0 1 2 3 4 5 111 Miami/Yoder 60 JT (El Paso) 0 1 2 3 4 5 112 Moffat 2 (Saguache) 0 1 2 3 4 5 113 Moffat County RE:No 1 (Moffat) 0 1 2 3 4 5 114 Monte Vista C 8 (Rio Grande) 0 1 2 3 4 5 115 Montezuma Cortez RE 1 (Montezuma) 0 1 2 3 4 5 116 Montrose County RE 1J (Montrose) 0 1 2 3 4 5 117 Mountain Valley RE 1 (Saguache) 0 1 2 3 4 5 118 North Conejos RE 1J (Conejos) 0 1 2 3 4 5 119 North Park R 1 (Jackson) 0 1 2 3 4 5 120 Norwood R 2J (San Miguel) 0 1 2 3 4 5 121 Otis R 3 (Washington) 0 1 2 3 4 5 122 Ouray R 1 (Ouray) 0 1 2 3 4 5 123 Park (Estes Park) R 3 (Larimer) 0 1 2 3 4 5 124 Park County RE 2 (Park) 0 1 2 3 4 5 125 Pawnee RE 12 (Weld) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 239

223 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 126 Peyton 23 JT (El Paso) 0 1 2 3 4 5 127 Plainview RE 2 (Kiowa) 0 1 2 3 4 5 128 Plateau RE 5 (Logan) 0 1 2 3 4 5 129 Plateau Valley 50 (Mesa) 0 1 2 3 4 5 130 Platte Canyon 1 (Park) 0 1 2 3 4 5 131 Platte Valley RE 3 (Sedgwick) 0 1 2 3 4 5 132 Platte Valley RE 7 (Weld) 0 1 2 3 4 5 133 Poudre R 1 (Larimer) 0 1 2 3 4 5 134 Prairie RE 11 (Weld) 0 1 2 3 4 5 135 Primero Reorganized 2 (Las Animas) 0 1 2 3 4 5 136 Pritchett RE 3 (Baca) 0 1 2 3 4 5 137 Pueblo City 60 (Pueblo) 0 1 2 3 4 5 138 Pueblo County Rural 70 (Pueblo ) 0 1 2 3 4 5 139 Rangely RE 4 (Rio Blanco) 0 1 2 3 4 5 140 Ridgway R 2 (Ouray) 0 1 2 3 4 5 141 Roaring Fork RE 1 (Garfield) 0 1 2 3 4 5 142 Rocky Ford R 2 (Otero) 0 1 2 3 4 5 143 Salida R 32 (Chaffee) 0 1 2 3 4 5 144 Sanford 6J (Conejos) 0 1 2 3 4 5 145 Sangre De Cristo RE 22J (Alamosa) 0 1 2 3 4 5 146 Sargent RE 33J (Rio Grande) 0 1 2 3 4 5 147 Sheridan 2 (Arapahoe) 0 1 2 3 4 5 148 Sierra Grande R 30 (Costilla) 0 1 2 3 4 5 149 Silverton 1 (San Juan) 0 1 2 3 4 5 150 South Conejos RE 10 (Conejos) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 240

224 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 151 South Routt RE 3 (Routt) 0 1 2 3 4 5 152 Springfield RE 4 (Baca) 0 1 2 3 4 5 153 St. Vrain Valley RE 1J (Boulder) 0 1 2 3 4 5 154 Steamboat Springs RE 2 (Routt) 0 1 2 3 4 5 155 Strasburg 31 J (Adams) 0 1 2 3 4 5 156 Stratton R 4 (Kit Carson) 0 1 2 3 4 5 157 Summit RE 1 (Summit) 0 1 2 3 4 5 158 Swink 33 (Otero) 0 1 2 3 4 5 159 Telluride R 1 (San Miguel) 0 1 2 3 4 5 160 Thompson R 2J (Larimer) 0 1 2 3 4 5 161 Trinidad 1 (Las Animas) 0 1 2 3 4 5 162 Valley RE 1 (Logan) 0 1 2 3 4 5 163 Vilas RE 5 (Baca) 0 1 2 3 4 5 164 Walsh RE 1 (Baca) 0 1 2 3 4 5 165 Weld County RE 1 (Weld) 0 1 2 3 4 5 166 Weld County S/D RE 8 (Weld) 0 1 2 3 4 5 167 Weldon Valley RE 20(J) (Morgan) 0 1 2 3 4 5 168 West End RE 2 (Montrose) 0 1 2 3 4 5 169 West Grand 1 JT (Grand) 0 1 2 3 4 5 170 Westminster 50 (Adams) 0 1 2 3 4 5 171 Widefield 3 (El Paso) 0 1 2 3 4 5 172 Wiggins RE 50(J) (Morgan) 0 1 2 3 4 5 173 Wiley RE 13 JT (Prowers) 0 1 2 3 4 5 174 Windsor RE 4 (Weld) 0 1 2 3 4 5 175 Woodland Park RE 2 (Teller) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Please go to the next page

PAGE 241

225 Note: (0)=No contact; (1)=1 contact; (2)= 2 4 contacts; (3)= 5 7 contacts; (4)=8 1 0 contacts; (5)=more than 1 0 contacts. SCHOOL DISTRICT (County) FREQUENCY 176 Woodlin R 104 (Washington) 0 1 2 3 4 5 177 Wray RD 2 (Yuma) 0 1 2 3 4 5 178 Yuma 1 (Yuma) 0 1 2 3 4 5 Note: (0)= No contact ; (1)= 1 contact ; (2)= 2 4 contacts ; (3)= 5 7 contacts ; (4)= 8 10 contacts ; (5)= more than 1 0 contacts Thank you so much for your responses