10 likes | 131 Views
The Policy Diffusion of Universal Preschool: An Event History Analysis Approach. F. Chris Curran. Peabody College, Vanderbilt University. Theoretical Framework and Background. Sensitivity Analysis. Conclusions. Introduction. Limitations and Future Work. Descriptives.
E N D
The Policy Diffusion of Universal Preschool: An Event History Analysis Approach F. Chris Curran Peabody College, Vanderbilt University Theoretical Framework and Background Sensitivity Analysis Conclusions Introduction Limitations and Future Work Descriptives EHA Descriptives and Survivor Curve Event History Analysis Model Data Results Motivation: There is an increased emphasis on early intervention and expansion of preschool; however, there is a lack of research on the policy environments that facilitate such expansion. Research Questions: To what extent does the interstate factor of adoption of policies of universal preschool by nearby states predict the adoption of state policies of universal preschool? To what extent do intrastate factors such as political party control, education expenditures, family income, size of the student population, and recent early childhood policy changes predict adoption of policies of universal preschool? • Interstate Predictors: • No relationship between variables measuring nearby state adoptions of the policy and a given state’s likelihood to adopt. • Diffusion of the policy was not supported. • Intrastate Predictors: • Democratic party control of legislature predicts higher likelihood to adopt universal preschool. • Earlier adoption of a policy of universal kindergarten predicts earlier adoption of universal preschool in one model. • Policy Innovation and Diffusion: • Innovation = a policy that is new to a state regardless of how long it may have existed in another state1 • Diffusion = the way in which an innovation is transferred through various avenues in society1,2 • Background: • Nine states have the policy, while 3 have fully implemented it. • Preschool has a positive impact on children’s achievement.3 • Education governors are predictive of adoption of P-16 councils.4 universal preschool = h0(t)exp( B1(regional proximity) + B2(universal kindergarten adoption decade) + B3(compulsory kindergarten) + B4(% Republican legislature) + B5(Republican governor) + B6(average family income) + B7(total ed expenditures lagged) + B8(% change in ed expenditures lagged) + B9(# of K students lagged) + B10(% change in # of K students lagged)) • Logistic Regression: • To test for sensitivity to the methodology, namely EHA, logistic regressions were performed. Statistical significance was lost in most cases, but the direction of coefficients persisted. • Covariate Analysis: • To test for robustness to the set of controls included, a series of models were run in which each covariate was removed. The findings remained statistically significant in 5/9 to 7/9 of the models depending on the independent variable of interest. Figure 1: Adopters of Policies of Universal Preschool • Data was compiled from the following sources: • US Census • Klarner’sBook of the States • NCES’ Common Core of Data • PreK Now initiative sponsored by the Pew Center on the States • NIEER Preschool Yearbook 2011 • Universal Kindergarten from Cascio (2010) • Limitations: • A very low number of adoptions of the policy restricted the number of covariates that could be included due to limits on degrees of freedom. • Future Work: • Apply this methodology to targeted preschool programs and to policies of universal kindergarten. Figure 2: Survivor Curve for Adoption of Policy of Universal Preschool Berry and Berry, 1999 Walker, 1969 Abbott-Shim, Lambert, & McCarty, 2009; Casto & Mastropieri, 1986; McKey et al., 1985; Puma, Bell, & Cook, 2005 Mokher, 2010