Reconstructing Local Governance in American Public Education

Document Sample
Reconstructing Local Governance in American Public Education Powered By Docstoc
					Reconstructing Local Governance in American Public Education: Politics, Policy, and Process

Kenneth K. Wong and Warren E. Langevin Vanderbilt University


The political tradition of local school governance predates the existence of the United States. Yet the most prominent methods for selecting local authorities have evolved considerably from their colonial origins in response to the changing political climate of the nation, constituent demands, and legislative innovation. Recognizing the intergovernmental context for policy innovation, this paper examines the diffusion of governance reform in American public education since the release of A Nation at Risk (1983) using discrete event history analysis. Using the complementary log-log model with a cubic spline representation of time, we test the national influence and ideological distance hypotheses for state takeover reform and charter school legislation. Our preliminary results show considerable differences in the respective number of significant relationships and explanatory power for the predicted relationships of both strands of governance reform.

Prepared for delivery at the Annual Meeting of the Southern Political Science Association in New Orleans, January 6 - 8, 2005. This research was supported in part by a grant from the Mid-Atlantic Regional Education Laboratory and the United States Department of Education. The authors appreciate helpful comments from Sean Nicholson-Crotty, William Doyle, Michael McLendon, Thomas Smith, and Matthew Springer. Opinions reflect those of the authors and do not necessarily reflect those of the funding agency. Please send questions and comments to: Kenneth Wong, Vanderbilt University, Peabody 514, 230 Appleton Place, Nashville, TN 37235-5721, or email (


Reconstructing Local Governance in American Public Education The political tradition of local school governance predates the existence of the United States. Yet the most prominent methods for selecting local authorities have evolved considerably from their colonial origins in response to the changing political climate of the nation, constituent demands, and legislative innovation.1 Although many historians regard the structural reforms of the past century as well-intentioned, the passage of state accountability legislation reinvigorated criticism of current governance arrangements for restricting market-driven consumer choice in the provision of public services (Chubb and Moe 1990). More recently, local constituents have strongly opposed the reduction of local discretion under the No Child Left Behind Act (2002) without concurrent acceptance of fiscal responsibility for redistributive spending. Perhaps as a result of the insulation of public education from voters in the past century and the diminishment of civic capacity (Stone et al. 2001), the national accountability movement thus confronts state legislators with a “crisis of legitimacy” – how do state bureaucratic institutions respond to an unprecedented level of public and private demands from governors, business leaders, constituents, and interest groups without the reinforcement of a competitive electoral process?2 And, in the broader consideration of state policymaking, when are elected lawmakers willing to challenge the political and economic constraints of divided localism in order to pursue localized governance reform? By virtue of the tradition of state plenary authority in education, the passage of governance reform represents an intriguing case for the study of policy diffusion. Since Berry and Berry (1990) renewed scholarly interest in the structural determinants of policy adoption across the fifty states, the majority of published studies have examined the response of state legislatures to outward signals from the national government (Mooney and Lee 1995, Soss et al. 2001, Allen et al. 2004). When state policymakers debate the adoption of new public school laws, however, they must consider a different legal context reinforced by constitutional provision and the democratic tradition of local


control. The distinctive character of education policy found in state legislative behavior, according to Gray (1973), suggests an empirically complex pattern of policy innovation that has not received wide scholarly attention. For example, while Mintrom (1997) analyzes open enrollment school choice laws, his substantive focus on policy entrepreneurs does not utilize the theoretical advantages of an intergovernmental context to explain structural tensions evident in current statehouse deliberations (Conley 2003). Accordingly, we believe political scientists should recognize the direct implications for state policymaking when contextual factors in public school governance exemplify “a continuing power shift” from local to state authorities characterized by recurrent intergovernmental conflict (DiLeo 1998, 137). This paper examines the political dynamics of legal and fiscal responsibility shaping the diffusion of governance reform. Through an extension of Wong and Shen’s (2002) substantive focus on charter school laws and state takeover of local school districts, we seek to more fully explain the antecedent conditions for state legislative activism when local bureaucratic constraints create an intergovernmental context for policy innovation. In moving beyond the party competition hypothesis to recent work on the ideological bases of policy learning (Grossback et al. 2004) and national influence (Allen et al. 2004), we engage an emerging debate on the different modes of deliberative conflict over institutional reform proposals (Hill and Klarner 2002). Finally, our study contributes to the general literature by examining the theory of diffusion when the focus of reform is not the exercise of state authority, but the conditioning of local political actors on national terms. Background Localism represents the most recurrent dilemma for state legislators whose constitutional authority may preclude compulsory deference to federal interests but does not necessarily imply constituent support for governance decisions. While the education policy domain is characterized by greater consensus among national and state lawmakers than the increasingly partisan mode of


congressional policymaking (Aldrich and Rohde 2001), local officials hold significant power that constrains the principal-agent hierarchies forged by vertical divisions in the federal system (Chubb 1985; Peterson, Rabe, and Wong 1986). Although state constitutions generally do not elaborate upon the informal character of state-local relations, the politics of local control reinforce the assertion that “local governments possess considerable power” not only in deciding the election of public officials but also in restricting how state officials use their powers of office (Briffault 1990, 112). At the most basic level, the unique electoral basis for local institutional legitimacy allows constituents to simultaneously uphold the popular mandate of local school board members in making responsive decisions and constrain the ability of state legislators to enforce sustainable intergovernmental relationships. It is consequently not surprising that given the dramatic extent to which state legislatures revised the operational terms of the intergovernmental partnership in adopting governance reform legislation consistent with national policy goals, they also transformed the calculus of local resistance to policy innovation. With San Antonio Independent School District v. Rodriguez (1973) as the legal foundation for maintaining a tradition of state plenary authority in education, publicly-elected state lawmakers paradoxically enacted broad governance reform in the years following A Nation at Risk (1983) to empower state and local educational officials in their respective spheres of influence. For those local school districts with recurrent patterns of fiscal, managerial, and academic failure, twenty states transferred governance authority from local school boards to contracted management teams with the legal approval of the chief state school officer and state board of education. On the other hand, thirty-nine states authorized the creation of charter schools with limited bureaucratic constraints and regulatory oversight to introduce market-driven competition as an alternative to traditional public schooling and reshape the political incentives of school board members (Hill, Pierce, and Guthrie 1997). Although they are not mutually exclusive in their potential application


by state policymakers, each strategy clearly proposes a different core approach to governance reform within the context of paradigmatic policy change (Wallis and Dollery 1999). Yet what is not clear are the different modes of political conflict shaping the governance strategies conceived by policy actors in response to the constraints of divided localism. In this paper, we explore the intergovernmental context shaping the diffusion process to get a more complete understanding of the electoral tensions for local institutional legitimacy interwoven in governance reform. The Diffusion of Governance Reform in American Public Education The introduction of discrete event history analysis remains one of the most significant developments in the study of state policy innovation. To provide an empirical test of their unified model of state policymaking, Berry and Berry (1990) suggest a binary approach based on legislative policy outputs, coded in dichotomous terms, over a specified time interval. The basic method of discrete event history models in policy diffusion compares the relative contributions of internal and external determinants in the legislative process through the hazard rate, such that the probability of state policy adoption is conditional on prior event non-occurrence as well as the covariates: h(t) = Pr( T = ti | T ti , x), (1)

where the random variable T is a discrete variable indicating the time of event occurrence, ti denotes the individual observations at specific time intervals, and x represents the covariate relationships of interest. In simpler terms, the hazard rate provides a measure of relative likelihood for state policy innovation given that state legislators have not previously adopted the specific policy of interest. With the preceding determination of the hazard rate for policy innovation, it is important to consider the functional form of the discrete-time model. Many diffusion studies rely on maximum likelihood estimation through probit or logistic regression on the categorical dependent variable (Berry and Berry 1990; Mooney and Lee 1995; Wong and Shen 2002). Although the choice of functional form should reside on whether the policy outcome reflects an underlying normal or


binomial distribution, most scholars will reach similar conclusions from both functions in discrete event history analysis over short time periods. Yet when the basic assumptions of binary data analysis are not reflected in the overall risk set, according to Buckley and Westerland (2004), the more attractive option is the complementary log-log function of the form: log [– log (1 – λ i)] = β0 + β1x1i + β2x2i + … + βkxki , (2)

where λ i represents the log-odds ratio of the probability of state policy adoption to the probability of prior event non-occurrence. The major distinction of the complementary log-log function is an asymmetrical response curve that gradually departs from the initial event of policy innovation, but more rapidly approaches complete diffusion than the probit and logit specifications (Agresti 1990).3 Dependent Variable: States Adoption of Governance Reform The passage of governance reform began with the state takeover of Newark Public Schools in New Jersey and Pike County School District in Kentucky. To fully explain the competitive legal context for different reform strategies, we create two independent models of the diffusion process for state takeover reform and charter school legislation. Our dataset begins in the first year after the introduction of each policy innovation. The dependent variable is coded zero for each year in which the state legislature does not adopt the particular governance reform strategy and one in the year of adoption. For states not adopting by the end of 2002, the full series of state-year observations is included in the respective dataset. Independent Variables: Vertical Diffusion Forces The prominent role of federal policymakers in prioritizing governance reform since the release of A Nation at Risk (1983) suggests that vertical forces may strongly influence the diffusion process. While the national government may signal preferences to state legislatures using several different mechanisms, the availability of federal funding is among the most influential determinants of policy innovation (Allen et al. 2004; Welch and Thompson 1980). Since 1965, the United States


Congress has utilized the fiscal incentive strategy in education by allocating categorical grants to state education agencies with a clear expectation that states will adopt national reform goals (Wong 1999). In 1994, one major goal when national lawmakers reauthorized federal spending was dual endorsement of charter school laws and state takeover as a reform strategy. Accordingly, we model the broad conditions of national policy leadership in governance reform using the binary variable approach, with all years prior to the passage of the Improving America’s Schools Act (1994) coded as zero and subsequent years coded as one (Allen et al. 2004).4 The national influence hypothesis suggests that states will be more likely to adopt both reform strategies after federal lawmakers legitimate the implementation of governance reform in public education. Independent Variables: Internal Determinants The annual budgetary process is many state legislatures regularly elicits the most divisive confrontations in American government due to the considerable level of state expenditures for social welfare programs and electoral constraints on state revenue collections. Without the expansive borrowing capacity of the federal government, state lawmakers cannot invest their political capital in expensive reform proposals during budgetary years characterized by fiscal insolvency. At the same time, local interests are well-positioned to maintain current governance arrangements when an overall budgetary crisis shifts public attention toward statewide fiscal stability rather than an interconnected network of local governmental activities during the limited duration of state legislative sessions. Recent studies of policy diffusion recognize this limitation on state legislative behavior by incorporating a measure of fiscal health, defined as the ratio of total state revenue minus total state spending to total state spending, as a time-varying covariate in their final model and hypothesizing an overall direction of association based on potential costs of implementation (Berry and Berry 1990; Grossback et al. 2004; Allen et al. 2004). To understand the intergovernmental dimensions of policy innovation, however, we recognize the need to specify our


hypothesized relationships for two different strands of governance reform. In regard to state takeover, we hypothesize that poor fiscal health will restrict policy innovation due to the major discretionary expenses represented in state appointment of contracted management teams and greater resource allocation to the targeted district (Wong et al. 2004). In contrast, we hypothesize the opposite relationship for states’ adoption of charter school laws because policy entrepreneurs have consistently framed the creation of privately-operated charter schools as a market-oriented, efficient policy option in statehouse deliberations (Hassel 1999; Mintrom 2003). Fiscal considerations may be more evident in the relationship between local governments and the states in terms of actual governance authority. Although Wirt (1985) found that the local share of total state spending does not dictate formal levels of local control, current scholars assert the extent of decentralization within the public finance system accords local officials with varying levels of political power and responsibility in the policy process (Wong 1999). In conceptual terms, the local revenue share demonstrates the degree to which local taxpayers support local political institutions. Given the unique intergovernmental partnership for financing and delivering public elementary and secondary education, we believe the considerable influence held by local governments and their constituents in revenue generation has been the principal source of tension in state legislative efforts to restore performance accountability through the establishment of strict conditions for local school board autonomy. For example, state legislators may equate local fiscal capacity as prospective funding for their market-driven reform initiatives when local governmental interests fail to perceive institutional reform as a direct threat to their policymaking influence. As a result, we hypothesize states with higher levels of local fiscal responsibility will be less likely to implement takeover reform due to the electoral tensions of divided localism. Conversely, we hypothesize states with higher levels of local fiscal responsibility will be more likely to adopt


charter school laws, which can be attributed to a joint recognition of the prospective electoral benefits for education reform and the availability of intergovernmental revenue streams. Following the conventional approach of several diffusion studies, we incorporate timevariant measures of constituency size and citizen wealth as control variables in our final model of policy innovation. Based on the growing prominence of education reform as a campaign issue in general statewide elections, we argue that states with higher student enrollment totals will have greater incentives for governance reform because of the organizational basis for political change in larger constituencies and the underlying citizen demand for responsive democratic governance (March and Olsen 1995; Fusarelli 2002). Of course, a rival explanation of the economic constraints to institutional reform suggests that states with higher per capita personal income levels, measured in real dollars, will be more likely to adopt takeover and charter school laws based on the wealth of state residents and their related propensity to favor performance accountability standards.5 Given the expansive presence of ideological arguments in state legislative deliberations over governance reform, the role of ideology should be modeled in the diffusion process. At the federal level, one might reasonably argue that relevant measures of party control and divided government can explain the ideological dimensions of legislative success in the present conditional party government mode of congressional debate (Aldrich and Rohde 2001). Still, as Erikson et al. (1989) found in their national survey of state legislation, party control does not consistently explain state policy outcomes. In comparing the results of Wong and Shen (2002) with previous diffusion studies, we are reasonably certain that traditional measures of inter-party competition will not reveal the key ideological motives of elected politicians over time. Accordingly, we use the government ideology measure developed by Berry et al. (1998) to control for different levels of state policy liberalism over time, where higher values are associated with greater policy liberalism. Overall, we believe the politics of elite consensus in public school governance necessitate “pragmatic movement


toward the middle” that will not yield statistically significant results in our final model (Fusarelli 2002, 154). Although the extent to which state lawmakers differ in their response to fiscal responsibility has been partly attributed to citizen ideology (Wood and Theobald 2003), we believe the intergovernmental context for governance reform motivates the formation of broad regulatory coalitions without strong ideological foundations (Lowi 1972). Nonetheless, we hypothesize states with higher government ideology measures will be less likely to pass both strands of governance reform in recognition of current officeholders’ preferences for state bureaucratic involvement through traditional policy levers rather than market-driven alternatives (Wong 1999). Independent Variables: External Determinants One of the most important considerations in the diffusion process is the mechanism by which state legislators receive information about new policy options. Although there are multiple pathways of information in modern state government, including both regional and national organizations such as the National Conference of State Legislatures, a straightforward approach introduced in the recent work of Grossback et al. (2004) and Hill and Klarner (2002) focuses on the role of elite ideological preferences in policy adoption. Recognizing the diverse policy preferences represented in government ideology measures, Grossback et al. (2004) emphasize how state legislatures interpret their overall ideological distance from states having already passed the relevant policy innovation. Whereas the time-varying covariate of government ideology provides a direct reflection of elite preferences, this relative measure of ideological context offers a contrasting perspective on policy learning. Moreover, Grossback et al. (2004) contend that states will value their ideological similarity with the most recent adoptive state to the same extent as the average cumulative ideological preferences of prior adopters.6 Thus, the ideological distance hypothesis for governance reform suggests that an increase in the ideological distance between a potential adoptive


state and previous adopters will decrease the likelihood for successful diffusion of takeover reform and charter school laws. The final variables in our unified model of state policymaking control for alternative spatial and temporal explanations of policy innovation. First, we follow Berry and Berry (1990) and others in modeling spatial diffusion as the percentage of neighboring states having adopted the specific policy of interest in the previous year. The regional diffusion hypothesis suggests that states will be more likely to adopt either reform strategy when a higher percentage of neighboring states has enacted the legislation in a previous session. While the disadvantage of this conventional approach reduces our risk set by two states, Alaska and Hawaii, we believe the neighboring state formulation is more theoretically justified than competing measures of regional diffusion (Allen et al. 2004).7 The unfortunate lack of scholarly consensus on policy diffusion effects (Mooney 2001), however, is not evident in current discussions on serial correlation. To control the unobserved relationship among state-year observations in a given state, we use the methodological advantages of robust variance estimation in clustering observations on each state, which results in corrected standard errors and improved model fit (Buckley and Westerland 2004; Williams 2000). The reliable measurement of temporal dependence is one of the most notable developments in the study of policy diffusion. A basic assumption for modeling duration dependence in discretetime formulations is the difficult expectation of a flat hazard rate for event occurrence over time (Box-Steffensmeier and Jones 2004). By accounting for temporal dependence among observations and across the cumulative risk set, many political scientists have refined the predictive capabilities of discrete-time event history models. There are presently several attractive options for appropriate model specification. For example, Mintrom (1997) incorporates temporal dummy variables based on counting process theory, while King and Zeng (2001) favor logistic regression with randomized sampling of non-event observations. Based on preliminary analyses, we instead prefer the cubic


spline approach to conserve degrees of freedom (Beck et al. 1998).8 In our final model, we follow Buckley and Westerland (2004) in estimating a three-knot cubic spline of the probability of adopting each governance reform strategy on the linear time counter variable and include the linear predictions as a covariate to model duration dependency. Results and Discussion The preliminary results of our discrete-time event history analysis of state takeover reform and charter school legislation are displayed in Table 1. Both models offer strong predictive capacity for state policy innovation based on three conventional measures of overall model specification. The first measure is the model chi-squared test. Due to the methodological limitations of the likelihood-ratio test within cluster-correlated data (Williams 2000), we provide the adjusted Wald statistic for both models, which suggests overall model significance through rejection of the null hypothesis.9 In addition, we also evaluate model performance by comparing each model’s sensitivity and specificity in correctly identifying event and non-event occurrence. As expressed in the goodness of fit measure, we observe similar levels of model performance for both models consistent with previous diffusion studies. A third indication of appropriate model specification is provided by the link test, which helps to evaluate specification by modeling an alternative pattern of covariate relationships. Following our general expectations, we observe non-significant results in comparing alternative specifications of both full models and can thus expect correct specification. Based on the observed pattern of covariate effects, our preliminary results yield substantive differences from previous studies. Most importantly, we do not find significant support for the national influence hypothesis, which suggests that federal endorsement of state legislation expedites the diffusion process. This preliminary result is particularly surprising for the passage of charter school legislation because the Improving America’s Schools Act (1994) and the more recent No Child Left Behind Act (2002) provide tremendous fiscal support for charter school entrepreneurs


Table 1. Determinants of State Adoption of Educational Governance Reform
Independent Variables Improving America’s Schools Act (1994) Fiscal Health Local Revenue Share Student Population Per Capita Personal Income Government Ideology Ideological Distance Neighbors’ Adoption Duration Dependency Constant Adjusted Wald Chi-Square ( df ) State Takeover Reform -.125 (.825) -5.328** (3.183) -.023** (.012) .008*** (.002) -.004 (.022) .011 (.012) .012 (.018) 1.950* (1.32) 55.040** (28.207) -5.284 (3.503) 83.56 (9) Charter School Law .372 (.620) .340 (1.974) .003 (.017) .004*** (.001) .008 (.009) -.010 (.008) -.016 (.014) .312 (.753) 6.805* (4.244) -4.257*** (1.153) 22.02 (9)

Overall Model Significance 0.000 0.009 2 0.965 0.882 Count R Number of Cases 513 306 *p<.10, **p<.05, ***p<.01 Notes: Coefficients from complementary log-log model with robust standard errors clustered on state; all significance tests except constant are one-tailed; All analyses performed with STATA 8.0.

across the country. On the other hand, the failure for the government ideology or ideological distance variables to reach standard levels of statistical significance matches our initial predictions. While the ideological dimensions of state legislative behavior should not be discounted from the study of policy diffusion, we believe that the political dynamics of governance reform are better explained by relevant internal determinants. A more detailed analysis of the observed pattern of statistically-significant relationships offers different conclusions on competing reform strategies. The main difference between the takeover reform and charter school legislation models is the respective number of significant relationships and the explanatory power for the predicted relationships. In preliminary results for the takeover reform model, the notable contributions for the time-varying measures of fiscal health,


local revenue share, total student enrollment, and duration dependency suggest that political and economic constraints deserve greater consideration in the education policy domain (Wong et al. 2004). Based on our initial hypotheses, we believe the final results validate our theoretical basis for the existence of unique principal-agent relationships in the intergovernmental context of American public education. In the charter school law model, we can only assume that state legislatures with larger constituencies are more likely to prioritize structural changes through governance reform than state legislatures with smaller constituencies and less competitive electoral pressures.10 Conclusion Since the release of A Nation at Risk (1983), the process of governance reform has arguably generated the strongest resistance from local officials of any reform strategy for reinvigorating the venerable concept of performance accountability in American public schools. For example, most scholars who have endeavored to study the process of state takeover have focused on the removal of governance authority, the formation of new political alliances, the emergence of new policy priorities for urban mayors and state governors, and the implementation challenges for reversing decades of managerial failure (Burns 2003; Wong and Shen 2003; Henig and Rich 2004). Yet the study of governance reform has not matched the pace of policy and practice. In this preliminary study, we attempted to extend the diffusion literature by focusing on the political and economic dimensions of the governance reform process. Based on our preliminary analysis, we find good reason for scholars to more closely examine the political dimensions of state takeover and charter school reform. In the consideration of state takeover, the statistical significance of fiscal health, local revenue share, total student enrollment, and duration dependency measures provide an innovative foundation for future political-economic work that more closely examines conditional factors in the restoration of local


control. For charter school reform, it is likely that future scholars may need to find more innovative methodological approaches to improve the predictive capacity of diffusion models. Overall, we believe that scholars must continue to examine two questions that this paper has introduced: First, how do state bureaucratic institutions respond to an unprecedented level of public and private demands from governors, business leaders, constituents, and interest groups without the reinforcement of a competitive electoral process? And, in the broader consideration of state policymaking, when are elected lawmakers willing to challenge the political and economic constraints of divided localism in order to pursue localized governance reform?


Appendix A: Variable Measurement and Data Sources
Variable Dependent Variables State Takeover of Local School District Coded 1 for a state's intervention in district governance affairs in a given year and 0 otherwise Coded 1 for a state's adoption of charter schools law in a given year and 0 otherwise Ziebarth (2004) and original data collected by the authors United States Department of Education (2003) and original data collected by the authors Wong (1999) Measurement Source

Charter Schools Law

Independent Variables: Vertical Diffusion Forces Improving America’s School Act (1994) Coded 0 for all years prior to 1995 and 1 thereafter Independent Variables: Control Variables Fiscal Health Ratio of total state revenue minus total state spending to total state spending Local Revenue Share Percentage of state's elementary and secondary education spending contributed by local governments Student Population Total number of state’s public elementary and secondary students in a given year Per Capita Personal Income Measure of state’s per capita personal income in real dollars based on Consumer Price Index Government Ideology Elected officeholders' liberal-toconservative ideology score, with higher scores indicating greater liberalism Ideological Distance Relative measure of state policy learning process based on prior adopters’ government ideology Neighbors’ Adoption Percentage of neighboring states having previously adopted the policy of interest Duration Dependency Measure of time dependency using linear predictions of a cubic spline function on event occurrence

United States Bureau of the Census (various years) National Center for Education Statistics Common Core of Data (various years) National Center for Education Statistics Common Core of Data (various years) Bureau of Economic Analysis (various years) Berry et. al (1998) and related updates on Inter-University Consortium for Political and Social Research Grossback, Nicholson-Crotty, and Peterson (2004) Berry and Berry (1990)

Beck, Katz, and Tucker (1998)


Appendix B: Descriptive Statistics for State Takeover of Local School Districts
Variable Fiscal Health Local Revenue Share Student Population Per Capita Personal Income Government Ideology Ideological Distance Neighbors’ Adoption Mean 0.099 44.732 756702 14501.86 45.685 28.480 20.217 Standard Deviation 0.111 13.902 711243 1889.191 24.781 19.488 23.546 Minimum -0.298 13.778 88128 10014.54 0 0.089 0 Maximum 1.248 89.698 4950474 19470.36 96.371 81.3 100

Appendix C: Descriptive Statistics for Charter School Laws
Variable Fiscal Health Local Revenue Share Student Population Per Capita Personal Income Government Ideology Ideological Distance Neighbors’ Adoption Mean 0.091 44.922 751660 14484.58 48.715 22.077 28.407 Standard Deviation 0.089 13.642 748672 2015.896 25.918 13.442 29.173 Minimum -0.216 13.778 97137 10377.05 0 0.011 0 Maximum 0.326 89.668 5254844 20665.39 97.917 58.389 100


1. The Progressive era reform movement sought to insulate educational authorities from politicians. At the local level, reformers attempted to restrict the role of mayors and local regime actors in using public schools for political patronage. In contrast, many states gradually moved toward gubernatorial appointment. The term “crisis of legitimacy” should be correctly attributed to Boyd (2003), whose article surveys the current landscape of accountability efforts to reshape the traditional structure of American educational governance. According to Box-Steffensmeier and Jones (2004), the complementary log-log function may be preferable for event history analysis because it is the discrete-time analog of the Cox proportional hazards model. While we recognize the methodological limitations of this coding decision, we believe using the same technique as Allen et al. (2004) is the most judicious approach for testing their national influence hypothesis. Even though one might reasonably argue this method does not reveal the actual extent of federal influence, it seems appropriate for the mechanism of federal involvement in education policy. For the sake of clarity, we present basic descriptive statistics for both variables in the appendices. In computing the model coefficients and corrected standard errors, however, we rescaled total student enrollment and per capita personal income for a more straightforward tabular presentation. The formula for the ideological distance variable is the absolute value of a weighted average for the most recent adoptive state and all previous adopters minus the government ideology of a potential adopter, such that Ideological Distance = ABS([(MostRecentAdopterIdeo. + AllOtherAdopterIdeo.)/ 2] – Potential Adopter) We believe the current study may be less susceptible to this disadvantage because the adoption context for Hawaii does not have comparable pressures for governance reform with only one local school district. Also, the exclusion of any party control variables makes it possible to keep the state of Nebraska in the risk set. In preliminary analyses, we found the cubic spline approach yielded superior fit to linear, quadratic, and lowess representations. For a detailed explanation of time measures, see Box-Steffensmeier and Jones (2004). We use the adjusted Wald statistic because the total number of clusters is less than one-hundred.









10. An alternative explanation is large population states are more inclined to follow federal initiatives due to greater reliance on federal redistributive dollars. Because we did not initially specify this relationship in our full model, however, we leave this research idea to future work on the deliberative context for policy adoption

Agresti, Alan. 1990. Categorical Data Analysis. New York: Wiley. Aldrich, John H., and David W. Rohde. 2001. “The Logic of Conditional Party Government: Revisiting the Electoral Connection.” In Congress Reconsidered, ed. Lawrence C. Dodd and Bruce I. Oppenheimer. 7th ed. Washington, DC: Congressional Quarterly. Allen, Mahalley D., Carrie Pettus, and Donald P. Haider-Markel. 2004. “Making the National Local: Specifying the Conditions for National Government Influence on State Policymaking.” State Politics and Policy Quarterly 4: 318-344. Beck, Nathaniel, Jonathan N. Katz, and Richard Tucker. 1998. “Taking Time Seriously: Time-Series-Cross-Section Analysis with a Binary Dependent Variable.” American Journal of Political Science 42: 1260-1288. Berry, Frances Stokes, and William D. Berry. 1990. “State Lottery Adoptions as Policy Innovations: An Event History Analysis.” American Political Science Review 84: 395-415.


Berry, William D., Evan J. Ringquist, Richard C. Fording, and Russell L. Hanson. 1998. “Measuring Citizen and Government Ideology in the American States, 1960-93.” American Journal of Political Science 42: 327-348. Box-Steffensmeier, Janet M., and Bradford S. Jones. 2004. Event History Modeling: A Guide for Social Scientists. Cambridge: Cambridge University Press. Boyd, William L. 2003. “Public Education’s Crisis of Performance and Legitimacy: Introduction and Overview of the Yearbook.” In American Educational Governance on Trial: Changes and Challenges, ed. William L. Boyd and Debra Miretzky. Chicago: University of Chicago Press. Buckley, Jack, and Chad Westerland. 2004. “Duration Dependence, Functional Form, and Corrected Standard Errors: Improving EHA Models of State Policy Diffusion.” State Politics and Policy Quarterly 4: 94-113. Burns, Peter. 2003. “Regime Theory, State Government, and a Takeover of Urban Education,” Journal of Urban Affairs 25(3): 285-303. Briffault, Richard. 1990. “Our Localism: Part I – The Structure of Local Government Law,” Columbia Law Review 90: 1-115. Chubb, John E. 1985. “The Political Economy of Federalism,” American Political Science Review 79(4): 994-1015. Chubb, John E., and Terry M. Moe. 1990. Politics, Markets, and America’s Schools. Washington, DC: Brookings. Conley, David T. 2003. Who Governs Our Schools? New York: Teachers College Press. DiLeo, Daniel. 1998. “The State-Local Partnership in Education.” In Governing Partners: State-Local Relations in the United States, ed. Russell L. Hanson. Boulder: Westview. Fusarelli, Lance D. 2002. “The Political Economy of Gubernatorial Elections: Implications for Education Policy.” Educational Policy 16: 139-160. Gray, Virginia. 1973. “Innovation in the States: A Diffusion Study.” American Political Science Review 67, 1174-1185. Grossback, Lawrence J., Sean Nicholson-Crotty, and David A.M. Peterson. 2004. “Ideology and Learning in Policy Diffusion.” American Politics Research 32: 521-545. Hassel, Bryan C. 1999. The Charter School Challenge. Washington, DC: Brookings. Henig, Jeffrey R., and Wilbur C. Rich, eds. 2004. Mayors in the Middle. Princeton: Princeton University Press. Hill, Kim Q., and Carl Klarner. 2002. “The Many Faces of Elite Power in the System of 1896.” Journal of Politics 64: 1115-1136. King, Gary, and Langche Zeng. 2001. “Logistic Regression in Rare Events Data.” Political Analysis 9: 137-163. Lowi, Theodore J. 1972. “Four Systems of Policy, Politics, and Choice.” Public Administration Review 32: 298-310. March, James G., and Johan P. Olsen. 1995. Democratic Governance. New York: Free Press. Mintrom, Michael. 1997. “Policy Entrepreneurs and the Diffusion of Innovation.” American Journal of Political Science 41: 738-770. Mintrom, Michael. 2003. Policy Entrepreneurs and School Choice. Washington, DC: Georgetown University Press. Mooney, Christopher Z. 2001. “Modeling Regional Effects on State Policy Diffusion.” Political Research Quarterly 54, 103-124.


Mooney, Christopher Z., and Mei-Hsien Lee. 1995. “Legislative Morality in the American States: The Case of Pre-Roe Abortion Regulation Reform.” American Journal of Political Science 39: 599-627. National Commission on Excellence in Education. 1983. A Nation at Risk: The Imperative for Educational Reform. Washington, DC: United States Department of Education. Peterson, Paul E. 1995. The Price of Federalism. Washington, DC: Brookings. Peterson, Paul E., Barry G. Rabe, and Kenneth K. Wong. 1986. When Federalism Works. Washington, DC: Brookings. Soss, Joe, Sanford F. Schram, Thomas P. Vartanian, and Erin O’Brien. 2001. “Setting the Terms of Relief: Explaining State Policy Choices in the Devolution Revolution.” American Journal of Political Science 45: 378-395. Stone, Clarence N., Jeffrey R. Henig, Bryan D. Jones, and Carol Peirannunzi. 2001. Building Civic Capacity: The Politics of Reforming Urban Schools. Lawrence: University Press of Kansas. Welch, Susan, and Kay Thompson. 1980. “The Impact of Federal Incentives on State Policy Innovation.” American Journal of Political Science 24: 715-729. Williams, R.L. 2000. “A Note on Robust Variance Estimation for Cluster-Correlated Data.” Biometrics 56: 645-646. Wirt, Frederick. 1985. “The Dependent City? External Influences upon Local Control.” Journal of Politics 47: 83-112. Wong, Kenneth K. 1999. Funding Public Schools: Politics and Policies. Lawrence: University Press of Kansas. Wong, Kenneth K., Warren E. Langevin, and Francis X. Shen. 2004. “When School Districts Regain Control: The Political Economy of State Takeover of Local Schools and Its Withdrawal.” Paper presented at the Annual Meeting of the American Political Science Association, Chicago, Ill. Wong, Kenneth K., and Francis X. Shen. 2002. “Politics of State-Led Reform in Education: Market Competition and Electoral Dynamics.” Educational Policy 16: 161-192. Wong, Kenneth K., and Francis X. Shen. 2003. “Big City Mayors and School Governance Reform: The Case of School District Takeover,” Peabody Journal of Education 78: 5-32.


Shared By: