Reinventing an Idea: The Political Construction of Charter Schools1

Reinventing an Idea: The Political Construction of Charter Schools1

Bulkley, Katrina E


Policy ‘alternatives” begin as general ideas, and are inherently distinct from an innovation or reform that has been adopted by policy makers in a specific context in order to serve particularpurposes. However, the interplay between the idea itself and that moment in time when it is adopted, including how the context affects the idea, has often not been fully understood. As Hall notes, ‘if we want to accord ideas an explanatory power in analyses of policymaking, we need to know much more about the conditions that lend force to one set of ideas rather than another in a particular historical setting”(Hall 1989, p. 362). In this paper, I draw on literature from political science, sociology, and education to explore the political construction of charter schools during the adoption of charter school laws in three states; this political construction reflects the interplay between the ideas and context by examining the adoption of charter school laws in three states. By examining the influences outside a state and the historical and structural factors, policy discourse, and politics within a state, I explain how these different constructions developed. This comparative historical case study demonstrates the importance for scholars and policy makers of understanding exactly how a disembedded idea becomes integrated with local context in the process of policy formation, and the implications of such variation for later policy analysis and evaluation of policy impact.

A political construction is the particular understanding of a policy idea by policy makers in a specific context at a particular moment in time. This construction of the policy idea incorporates the theory of change or theory of action connecting the adoption of a policy to desired ends and the assumptions found in that theory (Argyris & Schon 1974; Weiss 1995). The term political construction, rather than social construction, is used to reflect that a construction is developed in a political (here, a legislative) process. Literature on policy design also bears a strong resemblance to the concept of political construction. Schneider defines policy design as ‘the content of policy, its text and practices”(Schneider 1997). She calls ‘target populations, goals, assumptions, rationales, implementation structures, rules and tools”the elements of design (Schneider 1997, p. 14).

As in the case of social construction, I am attempting to portray both a particular way of viewing or understanding a concept, the construction of charter schools, and that construction is also a process, rather than simply an end. The idea that a construction involves a dynamic process conveys the ongoing change that happens as any reform concept moves, within a particular context, from vague idea to a specific theory of change as incorporated in some sort of decision (such as the passage of a law) to implementation. While this study focuses on the construction of the charter idea at one specific point in each state, when policymakers passed the initial charter law, construction is an on-going evolution and ideas continue to develop as long as they are being used in policy, with each point in the evolution effecting future interpretations of an idea.

While this study emphasizes the substantial state-to-state variation in the interpretation of the charter school idea, there are some common elements of the initial idea (see Nathan 1996; Kolderie 1990). Charter schools are relatively autonomous public schools that receive a ‘charter’Or contract from an ‘authorizer,” which as a public entity such as a local school board, a public university, or a state board of education. These contracts, which are usually for 3-5 years, provide the operators of a school with more autonomy than a district-run public school, and generally include requirements that schools demonstrate their value (i.e., through improved student test scores) in order to have their contract renewed. Charter schools are essentially a governance reform, and are not directly operated by a school district, as has traditionally been the case with public schools. In most (but not all) cases, these are public Schools of choice”-serving only those students and families that select them -and receive public funding based on the number of students they serve (for a discussion of research on the impacts of charter schools, see Bulkley & Fisler 2003). In some ways, charter schools have similarities with English ‘grant-maintained”schools (Wohlstetter & Anderson 1994; Levacic & Hardman 1999).

This study asks : how was the idea of charter schools ‘bonstnicted”in these three states, and what influenced these constructions? Specifically, the paper provides a comparative case analysis of the interpretation of charter schools during the original adoption of the law in Michigan, Arizona, and Georgia. The paper begins by presenting a framework for the construction of a policy idea in a new context, then provides briefcase descriptions of each state and compares the process across states using the conceptual framework. The final section offers some ‘propositions” about influences on the construction of policy ideas.

Literature Review

An explanation of the construction of a policy idea in specific contexts can draw on a wide array of literature in political science, policy studies, sociology, and the study of organizations. I first look at the ‘multiple streams”literature, which examines linkages between policy problems and alternatives in a political environment. Much of the effort in political science devoted to investigating the adoption of policy innovations has emphasized the importance of ‘inputs,” such as the condition of a state’s economy, party control of the legislature and governor’s office, and actions of neighboring states, to explain when and whether a particular state adopts a particular innovation (Berry 1994; Gray 1973 ; Walker 1969). As Gray notes, this type of heavily quantitative research leaves little room for the process of policymaking (Gray 1994), while the multiple streams literature focuses on the process. The multiple streams literature is supplemented in the second section with literature on the socially-constructed aspects of problems, alternatives, and problem-alternative couplings.

Ideas of’Multiple Streams”

The ‘multiple streams’literature (see Zahariadis 1996) questions the assumption made in some rational models of policy analysis that a ‘solution’Or alternative is only considered or adopted in response to a pre-existing problem (see Patton & Sawicki 1986; Stokey & Zeckhauser 1978). Research in the multiple streams tradition moves toward an explanation of agenda setting and policymaking that emphasizes the active role of individuals. Analyses that rely on the logic of multiple streams examine different parts of the policy and decision-making process individually, and then look at the ways in which these different pieces are brought together.

Kingdon, drawing on Cohen, March and Oison (Cohen, March, & Olsen 1972), identifies three streams central to agenda-setting: problems, policy alternatives or solutions, and politics (Kingdon 1995). Each of these streams, he argues, operate independently, but all three must come together (or become ‘coupled’) during ‘policy windows”(or periods when movement is possible) in order for an issue to rise on the political agenda. Policy entrepreneurs often play an active role in this coupling (see Mintrom 2000). Thus, alternatives are developed separately from problems, and can, in fact, become coupled with different problems.

The independence of these ‘streams” has been questioned by a number of authors (c.f. Mucciaroni 1991). While the streams may be independent before the coupling of a problem and alternative, they are clearly interacting once that link is made. Thus, Kingdon does not directly address the potentially dynamic interaction between a problem or problems and an alternative. In this study, I assume only that, at some level (i.e., national debates), problems and alternatives develop somewhat independently.

Finally, as Mucciaroni points out, ‘Kingdon’s garbage can model views the role of institutions almost exclusively in situational terms. They constitute part of the changing ‘political climate,’ which leaves untheorized the role of their structural characteristics and decision-making processes”(Mucciaroni 1991, p. 467). Institutions, Mucciaroni goes on to argue, can facilitate or constrain problems or alternatives receiving attention from policymakers. Thus, the relative institutional power of actors such as governors and legislators and the norms of policymaking in a context may influence both the problems that are attended to, and the way in which ideas are constructed as solutions to policy problems.

The Social Construction of Problems and Alternatives

While the multiple streams literature provides importance insights, including the idea that problems and alternatives can develop independently and that timing and politics are critical to their linkages, it does not examine how problems or alternatives are actually defined. Social constructivism moves away from ideas of obj ective reality toward explanations that focus on how individuals construct the world around them through social interaction. As Edelman notes, ‘We are acutely aware that observers and what they observe construct one another; that political developments are ambiguous entities that mean what concerned observers construe them to mean” (Edelman 1988, p. 1-2). In other words, rather than focusing on objective realities, these works emphasize how individuals understand the world around them.

In a constructivist view, clear and objective problems are not simply waiting to be addressed by government action (see Gusfield 1981). Instead, they must be constructed in ways that are consistent with the particular context within which they are discussed. Rochefort and Cobb lay out several different factors that can influence how a condition is defined as a problem, including cultural values, interest groups, scientific information, and professional advice (Rochefort & Cobb 1995). The context in which a condition exists has important effects on how, and if , it is defined as a problem to be addressed by government action. Thus, a particular condition may be defined as a problem in one context, but not in another. Rochefort and Cobb argue that any complex social condition can have a number of different ’causes,”and that part of defining a problem is picking which cause or causes to emphasize, which to downplay or ignore, and what actors have the authority to act. The way in which policymakers define a problem is closely connected to the types of alternatives that they will use to address it.

In the same way that problems are socially constructed, alternatives are themselves constructed through a social process, and do not appear ‘fully formed” on a policy agenda. This raises questions about the assumption implicit in models of innovation diffusion that essentially assume that ‘policies”in different contexts with a particular ‘hame” are fundamentally the same. Recent work by several Scandinavian organizational theorists on the ‘translation’Of new ideas in specific contexts provides important insights into the reasons that policy alternatives are constructed or altered within a context (Czarniawska & Joerges 1996; Sahlin-Andersson 1996; Sevon 1996). First, they argue that ideas must be separated from the actions that come out of those ideas in order to shift from one location to another. In other words, they must be largely ‘uisembedded”from a particular context, and it is in this form, as relatively ‘bontext-free”prototypes, that they can travel between contexts (Sahlin-Andersson 1996). Thus, even when policymakers interested in charters looked to Minnesota (the first state to adopt a charter school law) for inspiration, they may have looked to the concept or idea as well as the action as embodied in the Minnesota law.

Because of the disembedded nature of an idea when it travels, it is likely to be transformed in any environment it enters as it becomes more clearly defined. Consequently, innovation is simultaneously spread by external forces and arises from local circumstances. Thus, ‘The translation model…can help us to reconcile the fact that a text is at the same time object-like and yet it can be read in differing ways…it is the people…who energize an idea any time they translate it for their own or somebody else’s use” (Czarniawska & Joerges 1996, p. 23). In the process of translation, new ideas and the problems with which they become linked are created through a process of social interaction and must ‘fif’with the particular local context.

While the constructivist literature can aid our understanding of the policy making process, there are potential pitfalls in applying it to policy analysis. An emphasis on the construction of problem definitions and alternatives can deemphasize the importance of structural factors, such as the state of the economy, the availability of slack resources in the budget, or the political composition of the legislature. While such structural factors need to be mediated by ideas, they provide the basic framework within which policy decisions are made (Stone 1988).

Conceptual Framework

A number of different analytical levels can aid in understanding the process of construction in a state, and explaining why a particular meaning of charter schools is ultimately used in a charter school law: external forces; structural and historical factors; the current policy discourse; and specific temporal orpolitical issues. While each analytic level might be used independently to explain the construction of charter schools in a particular law, I conceptualize the various components as influencing different points of the process that leads to aparticular construction (see Figure 1 ). These factors serve as sensitizing concepts (Denzin 1989) that ensure the study examines the possibility that they have an impact on the construction of charter schools.

External Forces

It is through the national discourse on education, or particular pieces of it, that state policymakers learn of new ideas. Other reform ideas under discussion nationally at a particular time may also influence the political construction of a new idea and set the stage for the policy discourse within states, helping to shape the types of problems and alternatives that policy makers debate. National discussions can influence state policy discussions in a variety of ways, as state policy makers read pieces by national organizations or influential educational thinkers, attend conferences about reform ideas, or talk informally with colleagues in other states. External forces introduce new ideas into the state policy discourse, where state politics and history shape the specific construction of the idea.

Structural and Historical Factors

State context may constrain the nature of the problem definitions and constructions of an idea to those that are politically viable in a context, including economic, political, and social aspects of the environment in a particular state (Berry 1994; Gray 1973; Walker 1969). Structural factors, such as party control and the power of different branches of government, are closely intertwined with state history. Inclusion of these factors addresses the concerns raised by Mucciaroni (1991) that the multiple streams model does not sufficiently attend to the role of institutions. In addition, the history of education policy in a state may have an impact on the problem-definitions, stories and theories of change that are seen as powerful and thus may be used in the construction of the charter school idea. Two indicators of this history are political culture (including the influence of different actors) (Ela/ar 1970; Marshall, Mitchell, & Wirt 1989) and state versus local control over education (Wirt & Kirst 1997).

Policy Discourse

Hall notes that, ‘Policy making takes place within an institutional framework… but it also occurs within the context of a prevailing set of political ideas”(Hall 1989, p. 383). The policy discourse includes problems and alternatives both within and outside a particular policy domain, such as education. When a problem and alternative are coupled and rise on the policy agenda, a more complete construction must be used to define a problem as caused by certain factors and to explain to people why a particular alternative is an appropriate response for this problem-definition. When an idea receives attention in a state, it enters an environment in which policymakers are discussing various problem-definitions and alternatives both in and outside of education. Policy discourse influences what problem-definitions are available to be linked with a new idea (although the appearance of a new alternative may actually stimulate latent problem-definitions).

Politics and Timing

As the multiple streams literature suggests, the specific timing of a decision, particularly in respect to what other activities are going on at the same time, has important implications for the decision that is (or is not) made. A significant aspect of that timing is who is involved in a particular policy discussion. Clearly, not everyone has access to elite policy discussions. The political context plays a role in determining who has access and who does not to the policy discourse surrounding charter schools. Among possible actors are policy entrepreneurs who can take an active role in connecting an alternative to problems identified in the policy discourse (Mintrom 2000, 1997).

Actors with access vary in how much attention they give to a particular policy issue, and what resources (financial and political, in addition to time) they choose to expend on that issue; those constructions supported by actors with access, attention and resources are expected to be weighed more heavily than constructions supported by less influential actors. As March notes, ‘Understanding decisions in one arena… requires an understanding of how participation in those decisions fits into the lives of participants”(March 1994, p. 199). Some actors may have had the opportunity to help shape the charter idea, but did not take it because they placed a higher priority on other issues. Finally, it is important to consider the goals of actors involved in charter school discussions that are unrelated to the policy itself.

Following the multiple streams framework, it is in a state’s policy discourse that the problems and alternatives streams develop and flow, intersecting and separating as time passes (Kingdon 1995). It is in the politics and timing arena that the politics stream flows, with its own separate set of processes. Due to the different roles each factor plays in the process, and the extent to which they are intertwined (i.e. history and politics cannot be completely disentangled), identifying the ‘most significant” is not realistic. The coming together of these three streams, when an alternative becomes coupled with a problem or set of problems at a politically opportune moment, allows for a particular political construction (or set of constructions) to emerge and be used in a law.

Data and Methodology

In order to examine the process of constructing the charter school idea, I conducted three case studies in Arizona, Michigan, and Georgia of the adoption of each state’s original charter school law. In all three states, these original laws have been amended in both small and significant ways. For example, in Michigan, the original law was altered a year after passage due to legal challenges. In Georgia, revisions over time have included the introduction of student choice, not found in the original bill. The use of case studies is designed to provide detailed contextual information on the political construction of charter schools underlying the charter law in each state, and the reasons for the use of a particular construction (or, in some cases, multiple constructions). case study research allows for the rich contextual analysis required to understand the intentions and expectations of policymakers, and the environment within which they were operating (Yin 1994). In addition, methodologists generally consider evidence from multiple case studies to be more compelling, and the results more theoretically generalizable (Yin 1994).

The sample for this study was selected based on theoretical replication (Yin 1994). That is, instead of predicting the same outcomes as a result of the same inputs, I expected different outcomes in these states, but because of similar processes. These cases were selected for several reasons: they adopted charter school laws in the same time period (1993-1994), they differ in terms of political composition at the time of adoption (Georgia had a Democratic legislature and Governor; Michigan had a Republican Senate and Governor and an evenly split House; and Arizona had a Republican legislature and Governor), and national studies have placed their charter laws in different categories (Arizona and Michigan’s laws are described as expansive, while Georgia’s is considered quite restrictive) (Buechler 1996 (July)).

Between 1996 and 1998,1 conducted semi-structured interviews with 11-12 people in each state, including state legislators and their aides, interest group and think tank representatives, the governor’s staff members, and other observers of educational policymaking. In addition, I collected and analyzed relevant documents, including the charter school law itself, over 500 newspaper accounts (from both local newspapers and national sources such as Education Week), interest group publications and notes from legislative hearings. When available, secondary sources, such as academic articles, were also used. I coded the interview transcripts and documents by theme and analyzed them to determine what constructions were actually used to justify charter schools in the states studied and the process that led to that construction (Miles & Huberman 1994).

Interpreting the Charter School Idea in Three States

The charter school idea was first discussed in the late 1980s, and became a viable policy alternative when Minnesota legislators adopted a law in 1991. For the original advocates of charter schools, especially those in Minnesota, the idea was seen as bringing together some of the perceived strengths of decentralization (especially to the school building level), deregulation, a growing interest in accountability based on student performance, and school choice (Nathan 1996; Kolderie 1990). The charter school idea has, as will be shown below, proven to be a particularly malleable one, with both different constructions and different specific legal provisions; Wells and her colleagues ( 1999) referred to it as the ’empty vessel” of school reform. The Center for Education Reform, among other organizations, has pointed out this malleability by regularly ‘rating”the quality of charter school laws based on their own particular policy preferences (see

The following case descriptions offer a brief glimpse into the process of constructing the charter school idea in each state, and a description of the theory of change (as reflected in the law) that was used in each case.2 Some of the key legal differences included those allowing for more or less autonomy; for example, more autonomy was associated with deregulation, the ability to have government entities other than school districts grant charter contracts, the ability for private schools to convert to charter school status, and the ability for charter schools to hire noncertified teachers (see Table 1 ). The following case descriptions provide more detail on the process of construction in the three states through a comparative analysis.


Michigan’s charter law passed in 1993, amid contentious political battles around educational financing and reform. At the time, there were a multitude of ‘broblems”and alternatives linked with education under consideration. One of the highest items on Governor Engler’s agenda was inter-district school choice, which would allow students to choose to attend public schools in other school districts if space was available. School choice met with great resistance from education groups, especially the powerful Michigan Education Association. Charter schools proved to be a compromise position. Governor John Engler, a powerful force behind the push for charters, attached a meaning to the concept that emphasized market forces and minimal government. However, the concept of charter schools reflected by the final bill shows a reliance on both markets and government to connect charter schools to improved educational quality.

The basic elements of the charter law allowed for increased autonomy through multiple sponsoring agencies, including local school boards, intermediate school districts, community college and public universities, but not through explicit deregulation (for a discussion of forms of autonomy for charter schools, see Wohlstetter, Wenning, & Briggs 1995). The combination of autonomy from districts and unions and the need to attract ‘customers” would lead, supporters thought, to higher quality education in these new schools. The problems that charter schools were seen as addressing included excessive control of public education by the ‘establishment” and a lack of competition for public schools and the most prominent goals of the charter law were to improve student outcomes within charter schools and the existing system and to increase parental satisfaction and control.


At the time the charter school law passed in Arizona, in june of 1994, the legislature was embroiled in a battle over school vouchers, and charters were also a compromise position in this state. In general, leading policy makers were concerned with issues of control and efficiency. Specifically, they were concerned about what was seen as the monopolistic and bureaucratic control of schools, insufficient parental control in public schools, and a system of public education they believed was inefficient.

Originally, the charter school idea was raised by an ad-hoc group called the ‘Education Reform Group,”which was committed to reform through consensusbuilding, and involved people from across the political spectrum, including Republican state legislators, business groups and education organizations. They supported a construction of charter schools that emphasized a limited educational marketplace and more school-level control. However, in the final debate, the ‘ERG” construction of charter schools was not adopted, and the conservative Republican version was. This ‘expansive”charter bill was designed to stimulate competition through the creation of a large number of charter schools. Among these ‘expansive” features were that charter schools: could be authorized by one of two state boards or a local school district; did not have to hire certified teachers; were exempt from parts of the state education code; and were fiscally independent entities.

The construction underlying the adopted Arizona law was built on school autonomy through multiple authorizers, student choice and substantial deregulation alongside competition as a means to improve school and system quality. The combination of parental control and school autonomy, policymakers believed, would result in a diverse range of charter schools that could offer a variety of programs because they did not have to be “one-size-fits-all,”a higher quality of public schools (charter and district-run), and a more efficient use of public funds.


Georgia’s charter school law (adopted 1993) passed during a period of relative political calm, and Governor Zell Miller was looking for an opportunity to adopt ‘experimental”programs and to demonstrate his on-going commitment to education. In this context, the Georgia charter school law was an incremental step towards less state and more teacher control of public schools. After a relatively uneventful debate, charter school legislation was adopted that was considered a minor, no-cost bill that was not expected to result in major changes in public education; rather, as one person who worked with the governor said, it was to be an “arrow in the quiver” of education reform. Specifically, the bill required charter school applicants to go to their local school board with a clear mission and set of goals, and request a charter that would allow them exemptions from much of the state education code and some additional control at the school level. Unlike in most other states that have passed charter legislation, charter schools in the 1993 Georgia law were not schools of choice.

Rather than being linked to problems of a ‘monopolistic bureaucracy,”charter schools were linked to concerns about excessive state control, inadequate teacher control and low school quality. The primary goals of the legislation were to improve quality and innovation and increase school-level, especially teacher, influence and move away from state control. The law offers substantial deregulation from state rules, but little increased autonomy from districts.

An interesting twist in the Georgia context was that, two years prior to adoption, two state legislators who had a very different interpretation of the idea introduced the charter concept. The version of charter schools they advocated incorporated considerable parental control through choice and would have allowed private schools to convert to charter status -both of these components were strongly opposed by other Democrats and representatives of public education groups. Despite their efforts, including bringing national charter advocates to Georgia, their idea of charter schools did not receive much support and essentially faded away.

Critical Differences in Constructions

As the case descriptions suggest, these three state laws each reflect a different construction of the charter school idea, and the legal provisions of the laws reflect these differences. In Arizona and Michigan, the national discussions about school choice among more free-market oriented groups had a forceful impact. While the logic of decentralization can be found in the theories of change in both Arizona and Michigan, and the logic of deregulation as well in Arizona, it was choice that drove charter advocates. Georgia policymakers did not rely on the logic of choice at all, but instead shaped the construction of charter schools around a combination of decentralization/site-based management and deregulation.

Each interview used in this study was coded based on whether or not a particular “rationale” for charter schools was mentioned. The findings from this particular analysis reflect the variations in constructions of the charter idea. For example, the idea of charters as a way to introduce competition into public education was mentioned in every interview in Michigan and Arizona, but only 20% of interviews in Georgia. On the other hand, deregulation was cited as a purpose of adopting a law in only 25% of the Michigan interviews, but in 56% of Arizona interviews and 80% of Georgia interviews. Finally, decentralizing more control to schools was only cited in 13% of Michigan interviews, but in 80% of Georgia interviews and 89% of Arizona interviews.

Building Constructions

This section examines the effects across states of the four factors in the conceptual framework, along with some other issues that arose during the analysis.

External Forces

In all three states, policymakers gathered information about charter schools and other reforms from a variety of external sources. While the most common source of data was the actual laws adopted by other states, there were also partisan national organizations that provided information. These included national conservative groups such as the Heritage Institute and the Center for Education Reform, which had direct links to conservative state think tanks that were very active in the debates about charter schools (the Mackinac Institute in Michigan and the Goldwater Institute in Arizona), and Democratic organizations such as the Democratic Leadership Council, (the DLC had connections to Governor ZeIl Miller in Georgia).

External forces often interacted with internal factors when policymakers were choosing which organizations to rely on for information on charter schools. At least to some extent, people found what they were looking for. For example, while a broad range of organizations were discussing charter schools in the early 1990s, conservative free-market oriented Arizona policymakers primarily received information, as filtered through the Goldwater Institute, which favored a competition-based understanding of charter schools.

While policymakers maintained that they had made the charter idea their own, they still liked to connect their law to the national charter movement, and the state variations reflect the lack of unity among charter advocates at the national level. Arizona’s law, with its high level of autonomy for schools through choice and deregulation, is closest to conservative national actors such as the Center for Education Reform. Michigan’s inclusion of charter schools in a system that emphasizes accountability brings its law somewhat closer to advocates in Minnesota (see Kolderie 1990), while Georgia’s law reflects the use of a school-site control form of the charter idea far more similar to that of some of the first to use the term ‘charter schools”(Budde 1989; Shanker 1988). Overall, individuals, groups, and the media were responsible for transporting the charter school idea into state policy discourse (Mintrom 1996).

National discussions about education reform more generally also provided policy arguments for state-level constructions of charter schools, but which aspects of this national discourse policymakers attended to most varied based on specific state contexts. Across all three states were some common themes, such as the focus on achievement and educational quality, which were consistent with ideas throughout the country. There was also shared interest in shifting power towards local actors such as parents and teachers, providing more local autonomy, and increasing accountability. However, the national conservative focus on competition took root in Arizona and Michigan, but not in Georgia.

Structural and Historical Factors

Political Culture and History

The role of state structural and historical factors, and especially the state’s political culture, set the stage for and constrained the construction of charter schools. Marshall and her colleagues argue that, within a state’s political culture, there is an ‘assumptive world”involving ‘the understandings of the rules among those who participate in state education policy-making” (Marshall, Mitchell, & Wirt 1989, p. 10). This assumptive world includes issues such as which actors have the right to participate in policy discussions, what are legitimate policy goals, and what problem-definitions and alternatives are not available for consideration.

Both Arizona and Georgia have traditionalistic political cultures, where a primary purpose of government is the maintenance of state elites. However, the elites in these two states are rather different. Arizona’s elites are primarily business leaders, and education groups and leaders are on the outside of these privileged circles. Elites in Georgia, on the other hand, include leaders in the Democratic Party, and education leaders such as local school officials.

In Arizona, there was no need to maintain the existing education governance system, as the people empowered by that system were not the elites in the early 1990s. Arizona elites were, in fact, often in conflict with education leaders in the state, and were perfectly willing to lessen the power of the education establishment. Thus, a law that shifted power to parents was compatible with the preferences of the state’s elite. The assumptive world of Arizona policymakers valued efficiency and choice more so than in many other states, and this was reflected in the theory of change underlying the state’s charter law. Consistent with the emphasis on efficiency and choice, and the power of business interests, was a heavy reliance on the logic of the free-market. Embedded in the state’s charter law is an assumption that competition, which was viewed as driving the successful business environment, would lead to similar success in public education.

In Georgia, the construction of charter schools that kept these schools within the confines of the existing education governance structure was consistent with a traditionalistic political culture (Fleischmann & Pierannunzi 1997; Gray 1996). The assumptive world for Georgia policymakers included a focus on maintaining the status quo, and the role of elites in that status quo, and elite policymakers were uninterested in fundamental changes to a system that favored them. The Georgia charter school law was an incremental alteration of the status quo that kept power relationships largely unchanged. Rather, it was a pendulum swing in the direction of decentralization in response to the increased centralization that resulted from the implementation of the Quality Basic Education Act in the 1980s and early 1990s (Wohlstetter 1994).

Assumptive worlds include those issues that are figuratively ‘off the table”in policy discussions. For Georgia policymakers, one such issue was school choice. During the desegregation period, School choice” that provided predominantly white parents with public money to send their children to private academies was one strategy for maintaining a segregated educational environment (Lugg 1996). Memories of this negative use of choice remain in Georgia, and, according to respondents, made modern choice reforms unpalatable for many in the state at the time the charter law was adopted. Consequently, Georgia policymakers constructed the charter idea around issues of teacher control and deregulation, avoiding politically-charged school choice.

In Michigan, charter schools that were largely uncontrolled by either state actors or regulations would have conflicted with a moralistic political culture that emphasized the role of government in serving the common good (Browne & VerBurg 1995). If charter schools operated primarily outside the control of government, the state would have conceded its ability to ensure that these schools served public needs. While Governor Engler originally wanted charter schools to receive substantial waivers from state regulations, the process of debating the law resulted in the state maintaining control in a way consistent with its political culture.

Structural Factors

The most significant structural factor in the political construction of charter schools in these three states was the party control of the legislative houses and the governor’s office. The push towards an understanding of charter schools that emphasized competition was ideologically linked to Republican reform ideas in Arizona and Michigan, while a construction more closely tied to traditional education reform ideas such as site-based management was in keeping with the moderate Democratic control of Georgia politics. However, party control cannot tell the whole story about the construction of charter schools. For example, it is difficult to imagine, even if Republicans had been more powerful in Georgia politics, that, given the state’s history with school choice, a strong choice/competition understanding of charters would have found its way into legislation at that time.

The economies of all three states, along with the country, were emerging from a recession in the early 1990s, and there was little support for large new education programs (with the exception of those programs in Georgia that were directly funded by a new lottery). This may well have made charter schools, especially charter programs that were relatively revenue-neutral, more appealing to policymakers than costlier reform alternatives.

A heavy reliance on structural factors in explaining the adoption of charter schools demonstrates a basic problem with studies of policy formation and adoption (Walker 1969 ; Gray 1973). As Sabatier points out, when discussing such models, ‘the aggregated nature of the factors -and the neglect of individual-level processes -results in a model that inadequately accounts for the importance of agency in policy development”(Sabatier 1991, p. 150). The construction of charter schools in these three states was generally consistent with political culture, history, and structural factors. However, the role of individuals, timing and various other factors were critical in determining what construction, among all possible constructions that were consistent with history and structure, would actually be used in a state’s law.

Policy Discourse

The problems and alternatives prominent in the policy discourse in each state at the time of adoption were, as predicted by Kingdon, often developed somewhat separately from each other, and problem-alternative pairs were frequently created after the specific problem-definition and alternative were discussed independently. There were some common elements in the discourse in all three states, tied to national discussions about the perceived need for more local and parental control and higher quality education. Rather than a single, objective account of the causes of school quality, policymakers developed a variety of explanations based on their own understandings of education (consistent with literature on the social construction of policy problems).

Clear differences in both the style and content of education policy discussions across the cases were tied to the historical and structural factors. Somewhere in the problem and alternatives streams in all three states, which included essentially every problem-definition and alternative ever considered in the state, were the full range of issues discussed nationally. This range included alternatives such as teacher professionalism, site-based management, removing requirements for teacher certification and vouchers. However, it was only in Michigan, where there was a more contentious political environment, that a broad range of issues was actively in discussion.

Georgia and Arizona each had a ‘prevailing set of political ideas”that was more coherent than that in Michigan, reflecting the dominance of one political party in these states. Georgia’s definition of problems in education centered on general concerns about educational quality, low education funding, and the perceived excessive regulation by the state tied to the implementation of the 1985 Quality Basic Education Act (Wohlstetter 1994). Alternatives mapped onto these problems included deregulation and support for school-level innovation. There was little discussion in Georgia during this period of major changes in the structure of public education.

Arizona also had aprevailing set of ideas, almost all of which were tied to attempts to alter the basic structures of education. The discourse in Arizona centered on the perception of a problematic system of public education that relied on excessive bureaucratic controls, with few incentives for schools to improve the quality of their product or respond to the demands of their consumers. While other problems were also discussed, such as low funding, the discourse was dominated by the problem-definition of a monopolistic system that needed competition to spur its improvement.

Kingdon argues that problems and alternatives develop separately because they are often discussed by different sets of people (Kingdon 1995). In these three cases, however, problems and alternatives were often discussed by the same people. Despite the crossover of participants in the problem and alternatives streams, in all three states charter schools were able to attach to different problem-definitions as they were debated. The discourse in each state operated in conjunction with politics and timing in the construction of charter schools.

Politics and Timing

The previous sections may seem to suggest that the construction of charter schools in each state was predetermined by its close connections to the structure, history and discourse of the time. However, while constrained by these factors, the meaning attached to the charter idea was also influenced by more happenstance occurrences in the politics stream during the debate and passage of each law. Several important issues connect politics and timing to the construction of charter schools in each state, including: other policy issues, within and outside of education, being debated at the time; who had access to the education reform and charter discussions; and, among those who had access, who chose to actually be involved in the debate.

In Michigan, the education financing controversy dominated the legislative session in the fall of 1993, and overshadowed all other policy discussions. As a result, while everyone was aware of charter schools, groups such as the teacher unions were unable to expend the energy they might have otherwise to either push an alternative understanding of the charter idea or defeat the passage of any charter bill. Another important piece of the political environment was the debate over school choice. Governor Engler and his allies “traded away”choice with educator groups and moderate Republicans in favor of a choice-oriented charter bill. Absent the threat of public school choice, which educator groups opposed much more fervently than charters, education organizations would likely have fought against any charter bill more intensely.

Arizona policymakers were also considering a variety of reforms alongside charter schools, including school vouchers. While the general interest in decentralization and more parental control certainly influenced the construction of charters, the presence of a viable voucher bill may have played the biggest role in creating the highly autonomy-granting, competition-oriented understanding of charters found in the 1994 legislation. The voucher possibility both emboldened conservative legislators and the governor in their quest for an understanding of charters similar to vouchers, and it limited the ability of education groups and others who feared vouchers to fight the construction of charters favored by conservatives.

The discourse about charters in Arizona really had two components, one pushed by the more moderate and consensus-driven ERG and the other by voucheroriented Republicans. A broad range of groups had access to the ERG component of the discussion. However, educator groups and Democrats were largely shut out of the closed-door Republican conversations in which the understanding of charters actually used in the bill was determined. Outside of the education arena, Governor Symington’s push for tax cuts supported an approach to education reform that did not involve any additional spending, and made any construction of charters that was cost-neutral or a cost-savings seem far more appealing to many Republicans than reform plans with a high price tag.

The politics and timing of the Georgia bill reflected the much calmer political environment and the relatively unambitious charter bill did not cause considerable anxiety among most policymakers. The desire of Governor Miller to adopt some small, almost token education reform was reflected in the passage of charters, while his desire not to upset anyone led to a construction of charters that was not far from the status quo. Republicans were allowed some access to the debate and were granted concessions largely in order to avoid heated debate over the idea.

The issue of access of education groups to the debate influenced the construction of charter schools in each state. When they were at the core of the discussions, as in Georgia and the ERG debates in Arizona, an understanding of charter schools that emphasized teacher control and de-emphasized competition was developed. In the Michigan and the conservative Republican discussion in Arizona, however, educators were only able to work on the margins of the debate and influence specific provisions of the charter bill, rather than helping to shape the overall meaning attached to charters. Historically, unions have had the most impact on policy in Michigan, the only of these states that has ‘closed-shop” schools, where the Michigan Education Association (MEA) was for many years one of the strongest teachers unions in the country. However, during the early 1990s, the MEA experienced both internal and external difficulties that diminished their political influence, and their impact on the construction of charter schools during these years of Republican power was less than it might have been at other times. The primary role that the MEA played was in convincing legislators, especially Democrats, to oppose specific provisions of the charter bill.

Finally, conservative think tanks were present in all three states (the Mackinac Institute in Michigan, the Goldwater Institute in Arizona, and the Georgia Public Policy Foundation). Both Mackinac and Goldwater were active participants in supporting a construction of charter schools that was competition-oriented, as well as other types of reforms designed to create more educational choice such as vouchers and open enrollment. The power of these groups was primarily through their close ties with like-minded politicians. Mackinac and Goldwater brought an understanding of charter schools promoted by national conservative organizations such as the Heritage Foundation, and the similarities of the construction of charter schools in Michigan and Arizona can be traced, at least in part, to these national groups. While present in the state, the Georgia Public Policy Foundation was inactive in the discussions about education reform and charter schools in the early 1990s. Even if they had been active, the most prominent reform they promoted at the time was higher academic standards, not student choice.

Other Factors in the Process

Policy Entrepreneurs. Individuals in each state could be identified as successful entrepreneurs, and played a more important role than anticipated in the conceptual framework. Each individual supported a particular construction of charter schools, was able to influence the actions of important policymakers, and obtained at least some of what he sought in the content of the bill. Absent these individual advocates, it is unclear whether charter legislation would have been passed in these states. When necessary, they made compromises to make the idea more politically viable. This perspective parallels Mintrom and Vergari ‘s emphasis on the role of policy entrepreneurs in the creation of charter school laws (Mintrom & Vergari 1998; Mintrom 2000).

Not only were there successful entrepreneurs, there were also unsuccessful or only partially successful policy entrepreneurs advocating charter schools in each state. In each case, there was at least one individual who promoted a construction of charter schools that was not used in the final adoption. For example, in Michigan, Republican Representative Bill Bryant introduced the charter school idea to the state, but that of Engler and his allies overwhelmed his more modest vision. Bryant’s conception of charters didn’t meet with much success because he was not sufficiently influential in his party, and he was interested in a more moderate approach to education reform than many other Republicans were at the time. Similarly, the two Democratic state legislators who originally introduced the idea in Georgia were unsuccessful entrepreneurs, as their more radical version of charters met with resistance in their own party, and the lack of support from the governor condemned their bill.

What the successful and failed entrepreneurs demonstrate is the need for entrepreneurs to: ( 1 ) support policy ideas, and particular versions of those ideas, that fit with the context of the state environment, including the contextual factors discussed earlier; and (2) to have the support of key political actors.

Policy Adoption Windows. Kingdon suggests that policy windows, which allow a particular problem-alternative pair to rise on the agenda at a given time, can be opened by events in any of the three streams of problems, alternatives, or politics (Kingdon 1995). Similarly, I suggest that windows must be opened for the adoption of a policy with a particular construction attached to it.

In each state, the policy adoption window was opened somewhat differently. In Georgia, the primary impetus for passing this bill seems to have come from the politics stream, as Governor Miller was interested in demonstrating his on-going commitment to education reform while not angering his core constituencies. However, the opportunity for the form of charter bill adopted was also increased through the problem stream, where people were concerned about the amount of state regulation and were seeking deregulatory alternatives.

Policy adoption windows in Michigan and Arizona opened through action in all three streams. In Michigan, the politics surrounding the refinancing of the education system gave Engler an opportunity to trade financing preferences for policy preferences, which led to the adoption of a choice-oriented charter bill. This political context was combined with an ideological focus on choice and competition. The existence of a strong ideology shared by powerful actors influenced both the definition of problems and the creation of alternatives. Ideology dominated the Arizona context, and the policy adoption window that opened was related to a general interest in education reform and a strong interest by conservative policymakers in passing reform that increased competition and parental choice.

Softening Up. New alternatives are seldom introduced to an environment and immediately acted upon. Instead, argues Kingdon, there is a ‘softening up”process, in which actors in a particular context are exposed to a new alternative over time. he says that, ‘Without this preliminary work, a proposal sprung even at a propitious time is likely to fall on deaf ears”(Kingdon 1995, p. 128). The findings in this paper support the existence of a softening up process. In all three states, the charter school idea was raised and discussed for roughly 2-3 years prior to the passage of the charter school law. Even when the idea was constructed differently during this period, the mere fact that people were somewhat familiar with the term may have made them more comfortable with adopting charter legislation.

Problems Chasing Alternatives, or Alternatives Chasing Problems? One of the key insights offered by Kingdon is that alternatives do not always arise because of specific problems. Instead, problems can chase alternatives or alternatives can chase problems (see also Cohen, March, & Olsen 1972). Both scenarios, and combinations of each, can be found in these three cases. In Georgia, charters chased problems. First, charters were linked to a problem-definition involving the basic structure of public education. The second incarnation emphasized the quality to be found within the existing structure and the need to empower that existing capacity. When regulation and state control became labeled the primary problems, charters became a new way of deregulating education.

In Arizona, there was a clear, ideologically-based definition of the problem among conservative legislators that emphasized the monopolistic bureaucracy and the need for more competition. Charters were tied to this problem-definition, or ‘chased”by it. As a result, charters were constructed with a theory of change that relied heavily on competition. One interesting factor in Arizona is that the problem chasing charters as an alternative was itself likely created by another alternative. The idea of vouchers and greater choice in education was an appealing alternative that led conservative policy makers to define educational problems as resulting from a lack of competition.

The situation in Michigan is not as clear-cut as the other two states. There were important actors who defined the problems in education similarly to conservative Republicans in Arizona, and who saw in charters an opportunity to increase choice and advance their reform agenda. However, some policymakers were intrigued by the idea of charters, and saw it as a way to solve a number of educational problems including both a lack of choice and a lack of school-site autonomy. These three cases support Kingdon’s basic premise that alternatives do not necessarily arise in response to particularproblems. In addition, these cases suggest the idea that in cases when there is a strong ideology, with clearly defined problems, these problems may be more likely to chase alternatives than when there is not such a dominant ideology.

Theoretical Implications

This section builds on the comparative analysis to offer some theoretical ‘propositions” about the process of constructing a policy idea within a given context, and the construction of the idea that emerges from that process.


Several aspects of the construction of charter schools shed light on the general construction of policy ideas in particular contexts.

Proposition 1: External debates set the stage for the construction of a policy idea, while internal forces bring the streams together.

In each of the three states studied, national discussions about both charter schools, and education reform more broadly, helped to define the terms of the state debates. Specifically, the content of the policy soup of alternatives, and the types of problems seen as in need of being solved, could be connected to broader conversations taking place outside the state context. As well, policy entrepreneurs at the state level played a role in connecting the internal state discussions to the national charter school debate. However, the actual process of constructing the charter idea and then putting it into action took place largely internally. History, political culture and structural factors constrained the construction of charter schools, while politics and timing were critical in bringing the streams together to form a problem-alternative link or set of links that could garner sufficient support to be passed as legislation.

Proposition 2: The local context influences the breadth of problems and alternatives considered.

The range of actively discussed problems and alternatives in their respective streams varies across contexts. Kingdon argues that certain problems and alternatives will rise to the top of the two streams, including those that meet a “value acceptability”threshold among central actors (Kingdon 1995). When a political environment is relatively unified, those issues that rise to the top of each stream will be more limited. For example, in Georgia, where moderate Democrats had been in power for decades, certain problems and alternatives related to school choice were simply absent from the policy discourse. However, the more contested the political environment, the broader the range of problems and alternatives discussed. Thus, in Michigan, with its high level of party competition and political power across an expanse of ideological perspectives, more explanations for perceived problems in education and alternative alternatives were debated.

Proposition 3: just as problem-definitions must be created, alternative-definitions must also be contextually created.

While Proposition 2 addresses the range of problems and alternatives considered, this proposition addresses the need for the specific construction of a policy idea to be developed with the state context. Rochefort and Cobb describe how social conditions are transformed into problems amenable to policy action through an explanation of the cause of the problem (Rochefort & Cobb 1995). Absent such a problem-definition, there is unlikely to be policy action on an existing condition. Similarly, policy alternatives or solutions are not discussed in a political context without a theory that connects them to either some defined problem or some desirable outcome. While the alternative may exist prior to an explanation connecting it to a problem or outcome, the alternative-definition is necessary for action to be taken. This is similar to saying that alternatives travel “disembedded” from the way they have been implemented, and must be translated and made more specific in a new context in order to be adopted (Sahlin-Andersson 1996).

In some cases, there may be multiple alternative-definitions that can be reconciled in the specifics of a law, or there can be competing alternative-definitions, not all of which are reflected in legislation. Thus, the appeal of charter schools in the three states studied was linked to the way the idea was defined, and the way the idea was defined was related to the context within which it was defined.

Proposition 4: Fertile ground is a necessary, but not sufficient, condition for action on the construction of a policy idea.

One of the clearest findings from this study is that a new policy alternative, with its alternative-definition, must not contradict the context in which the idea is being constructed. This is different from saying that a new alternative must fit perfectly with the context, as new alternatives can help to shape a new understanding of policy problems in ways that make the alternative seem the logical answer. For example, the push towards choice and vouchers in Arizona and Michigan helped to redefine the perceived problems with public education in those states.

However, an alternative-definition that is contrary to a context will not meet with much success. For example, when policymakers in Georgia advocated a charter construction that was inconsistent with the history and politics of the state, their efforts met with little success. The existence of fertile ground for a particular understanding of a policy idea is not sufficient for the idea to be seriously considered and acted upon. Individuals such as policy entrepreneurs and elected officials must have the position, influence and willingness to push the idea.

Proposition 5: Policy entrepreneurs can move an alternative in the direction of having action taken on it, but only if their alternative-definition fits with the context.

Similar to the presence of fertile ground for an alternative-definition, policy entrepreneurs are a helpful but not sufficient condition for the use of a policy idea. Mintrom and Vergari suggest that policy entrepreneurs can essentially drive the process of moving an idea towards adoption as a policy (Mintrom 1996,1997). This research supplements their analyses by demonstrating the many constraints imposed on entrepreneurs by the environment within which they operate.

Theory of Change

This study illuminates the process of construction that can result in a single policy idea being used for various policy goals, and with different underlying theories of change. However, it also provides information about ideas themselves, and the ways in which they are understood in a political context such as a state, as demonstrated by the following two propositions.

Proposition 6: An idea can be understood in multiple ways in the same context.

Wells and her colleagues first discussed the multiple meanings that are attached to charter schools, and the fact that these multiple meanings can co-exist in a single state (Wells et al. 1999). My case studies support these earlier findings, and demonstrate the importance of political construction in the development of multiple meanings. There were two scenarios found in these cases where multiple meanings were present. In the first, found in both Georgia and Michigan, groups with different understandings of the charter idea were able to compromise on the specifics of legislation, while still holding onto their alternative interpretations of the theory of change underlying the legislation. For this situation to arise, it appears that supporters of different interpretations must be able to agree on at least some basic details of the content of a policy. It is difficult to imagine the union supporters of the Georgia bill being able to find sufficient common ground with the competitionoriented advocates of the Michigan or Arizona bills to share support for a single bill.

A different situation occurred in Arizona, where two alternative understandings of charter schools competed, with different bills representing each interpretation vying for legislative support. In the end, supporters of one meaning of charter schools, which involved heavy reliance on parental choice and deregulation and was closely tied to vouchers, defeated supporters of the alternative version.

Not only can an idea have multiple meanings simultaneously, it can also have different meanings at different points in time. In Georgia, the charter idea was first introduced by two legislators who attempted to connect it to other reforms such as home schooling. However, when their efforts failed, the idea temporarily disappeared. When it reappeared, it was framed quite differently, and was attached to efforts to shift control to schools rather than to parents. The reinvention of the charter idea in Georgia, however, may only have been possible because the first incarnation met with so little success. Had the original understanding of charters in the state taken hold more firmly, it may not have been possible for the second meaning to completely supplant the first. Proposition 6 highlights the extent to which it is difficult to generalize about the process and outcomes of policy making, when so many factors influence how an idea is politically constructed.

Other Issues

One final theoretical implication of this study involves the ability to generalize about the process and outcomes of policymaking. Within each of these cases, the process that led to the adoption of a charter school law that was attached to a particular theory of change makes sense, and fits well with the history and politics of the state. However, the course of events across the three states varied considerably, and attempting to describe a specific process that led from the first introduction of the charter school idea to the adoption of a particular construction of charters would be quite difficult. In addition, generalizing from the history and structure of a state to the way an idea is constructed also encounters difficulties, as politics and timing play an important role in the construction of new ideas.


In common day language, there is often debate over whether a particular policy idea is ‘good’Or bad.”However, just as the devil is often in the details of legislative design, so the actual meaning of an idea is embedded in the specifics of the context within which it is being debated. Thus, understanding policy ideas necessitates understanding how they are shaped and altered by a particular time and place. Thus, the process of political construction involves bringing a policy idea together with problem-definitions in a context that can ultimately shape the construction itself. In the case of charter schools, a single idea was constructed so differently across these states, especially if one compares Arizona or Michigan and Georgia, that understanding political support for the idea, and interpreting its ‘success” as a policy requires a rich understanding ofthat construction. Similarly, policy ideas that cross international borders often have significantly different meanings and goals, based on the process of local construction.


1 Authors ‘ Note : Several people provided thoughtful critiques of this work at various stages in its production, including Michael Kirst, Priscilla Wohlstetter, David Tyack, James March, and William Firestone, as well as anonymous reviewers for Educational Foundations. Funding for this research was provided by the U.S. Department of Education’s National Institute on Educational Governance, Finance, Policymaking and Management (Grant#OERI-R308 A60003) through the Consortium for Policy Research in Education (CPRE), and through a research training grant from the Spencer Foundation. The opinions expressed in this research are those of the authors and do not necessarily reflect the views of the Spencer Foundation, the Consortium for Policy Research in Education, the National Institute on Educational Governance, Finance, Policymaking and Management, the Office of Educational Research and Improvement, the United States Department of Education or the institutional partners of CPRE.

2 For a more complete discussion of each case, see Bulkley (1999). For a description of the theory of change in each state, see Bulkley (In press).


Argyris, Chris, & Donald A. Schon. 1974. Theory in practice: Increasing professional effectiveness. San Francisco: Jossey-Bass.

Berry, Frances Stokes. 1994. Sizing up state policy innovationresearch. Policy Studies Journal 22 (Autumn): 442-456.

Browne, William P., & Kenneth VerBurg. 995.Michiganpolitics and government: Facing change in a complex state. Lincoln, NE: University of Nebraska Press.

Budde, Ray. 1989. Education by charter. Phi Delta Kappan March: 518-520.

Buechler, Mark. 1996 (July). Charter schools: Legislation and results after four years. Bloomington, IN: Indiana Education Policy Center.

Bulkley, Katrina E. 1999. Telling stories: The political construction of charter schools. Dissertation, School of Education, Stanford University, Stanford, CA.

Bulkley, KatrinaE. In press. Understanding the charter school concept in legislation: The cases of Arizona, Michigan and Georgia. International Journal of Qualitative Studies in Education.

Bulkley, Katrina E., & Jennifer Fisler. 2003. A decade of charter schools: From theory to practice. Educational Policy 17 (3).

Cohen, Michael D., James G. March, & Johann P. Olsen. 1972. A garbage can model of organizational choice. Administrative Science Quarterly 17(1): 1-25.

Czarniawska, Barbara, & Bernward Joerges. 1996. Travel of ideas. In Translating organizational change, edited by B. Czarniawska and G. Sevon. Berlin: Walter De Gruyer.

Denzin, Norman. 1989. The research act. Third ed. Englewood Cliffs, NJ: Prentice-Hall.

Edelman, Murray. 1988. Constructing the political spectacle. Chicago: University of Chicago Press.

Elazar, Daniel J. 1970. Cities of the prairie: The metropolitan frontier and American politics. New York: Basic Books.

Fleischmann, Arnold, & Carol Pierannunzi. 1997l .Politics in Georgia. Athens, GA: University of Georgia Press.

Gray, Virginia. 1973. Innovation in the states: A diffusion study. American Political Science Review67 (December): 1174-1185.

Gray, Virginia. 1994. Competition, emulation and policy innovation. In New perspectives on americanpolitics, edited by L. C. Dodd & C. Jillson. Washington, DC: Congressional Quarterly Press.

Gray, Virginia. 1996. The Socioeconomic and Political Context of States. In Politics in the American States: A Comparative Approach, Sixth Edition, edited by V. Gray & H. Jacob. Washington, DC: Congressional Quarterly, Inc.

Gusfield, Joseph. 1981. The culture of public problems: Drinking-driving and the symbolic order. Chicago: University of Chicago Press.

Hall, Peter. 1989. Conclusion: The politics of Keynesian ideas. In The political power of economic ideas, edited by P. Hall. Princeton, NJ: Princeton University Press.

Kingdon, John. 1995. Agendas, alternatives, and public policies. 2nd ed. New York: HarperCollins College Publishers.

Kolderie, Ted. 1990. Beyondchoice to newpublic schools: Withdrawing the exclusivefranchise in public education. Washington, DC: Progressive Policy Institute.

Levacic, Rosalind, & Jason Hardman. 1999. The performance of grant maintained schools in England: an experiment in autonomy. Journal ofEducation Policy 15 (2):157-167.

Lugg, Catherine A. 1996. For God and country: Conservatism and American school policy. New York: Peter Lang.

March, James G. 1994. A primer on decision making. New York: Free Press.

Marshall, Catherine, Douglas Mitchell, & Frederick Wirt. 1989. Culture and education policy in the American states. New York: Falmer Press.

Miles, Matthew, & A. Michael Huberman. 1994. Qualitative data analysis: An expanded sourcebook. Thousand Oaks, CA: Sage Publications.

Mintrom, Michael. 1996. Policy entrepreneurs, innovation diffusion, and policy research: The case of state charter school laws: Unpublished paper.

Mintrom, Michael. 1997. Policy entrepreneurs and the diffusion of innovation. American Journal of ‘Political Science 41 (3):738-770.

Mintrom, Michael. 2000. Policy entrepreneurs and school choice. Washington, DC: Georgetown University Press.

Mintrom, Michael, & Sandra Vergari. 1998. Policynetworksandinnovation diffusion: The case of state education reforms. Journal of Politics 60 (1): 126-147.

Mucciaroni, Gary. 1991. A critique of the garbage can model of policy-making. Polity 24 (3):459-482.

Nathan, Joe. 1996. Charter schools: Creating hope and opportunity for American education. San Francisco: Jossey-Bass.

Patton, Carl, & David Sawicki. 1986. Basic Methods of Policy Analysis and Planning. Englewood Cliffs, NJ: Prentice-Hall.

Rochefort, David A., & Robert W. Cobb. 1995. Problem Definition: An Emerging Perspective. In The Politics of Problem Definition, edited by D. A. Rochefort&R. W. Cobb. Lawrence, KS: University of Kansas Press.

Sabatier, Paul. 1991. Toward better theories of the policy process. PS (June): 147-156.

Sahlin-Andersson, Kerstin. 1996. Imitating by editing success: The construction oforganizational fields. In Translating organizational change, edited by B. Czarniawska & G. Sevon. Berlin: Walter De Gruyer.

Schneider, Anne. 1997. Institutions and policies for democracy: Improving policy design, response. Paper read at American Political Science Association Annual Conference, at Washington, DC.

Sevon, Guje. 1996. Organizational imitation in identity transformation. In Translating organizational change, edited by B. Czarniawska & G. Sevon. Berlin: Walter DeGruyer.

Shanker, Albert. 1988. Restructuring our schools. Peabody Journal of Education 65 (3):88100.

Stokey, Edith, & Richard Zeckhauser. 1978. A primer for policy analysis. New York: W.W. Norton and Company.

Stone, Deborah. 1988. Policy paradox and political reason. Glenview: Scott, Foresman.

Walker, Jack L. 1969. The diffusion of innovations among the American states. American Political Science Review 63 ($eptember):880-899.

Weiss, Carol. 1995. Nothing as practical as a good theory: Exploring theory-based evaluation for comprehensive community initiatives for children and families. In New approaches to evaluating community initiatives, edited by Connell, Kubisch, Schorr & Weiss. Washington, DC: The Aspen Institute.

Wells, Amy Stuart, Cynthia Grutzik, Cybill Carnochan, J. Slayton, & A. Vasudeva. 1999. Underlying policy assumptions of charter school reform: The multiple meanings of a movement. Teachers College Record 1OO (3):513-535.

Wirt, Frederick, & Michael Kirst. 1997. The political dynamics of American education. Berkeley, CA: McCutchan Publishing.

Wohlstetter, Priscilla. 1994. Appendix C: Georgia: Reform at the crossroads. In Ten years of state education reform, 1983-1993: Overviewwithfour case studies,editedbyO.Mnssell & S. Furhman. New Brunswick, NJ: Consortium for Policy Research in Education.

Wohlstetter, Priscilla, & Lesley Anderson. 1994. What can U.S. charter schools learn from England’s grant-maintained schools? Phi Delta Kappan 75 (6):486-491.

Wohlstetter, Priscilla, R. Wenning, & K.L. Briggs. 1995. Charter schools in the United States: The question of autonomy. Educational Policy 9 (4):331 -358.

Yin, Robert. 1994. case study research: Design and methods, second edition. Beverly Hills, CA: Sage.

Zahariadis, Nikolaos. 1996. Ambiguity, time and multiple streams. Paper read at American Political Science Association Conference, at San Francisco, CA.

Katrina E. BuIkIe/ is a professor with the Graduate School of Education at Kutgers University, New Brunswick, New Jersey.

Copyright Caddo Gap Press Winter 2004

Provided by ProQuest Information and Learning Company. All rights Reserved