Study on Charter School Effects on Student Achievement and on Segregation in Florida Public Schools

Charter schools have now been in operation for two decades in the U.S., and for 15 years in Florida. Florida took the third place in the U.S. in the number of charter schools operated and the student enrollment in 2010. This study examined the assumed effects of charter school policy on the public school system: charter school effect on student achievement in charter schools and in TPSs, and segregation effects and stratification effects on charter schools and traditional public schools (TPSs). I applied three perspectives to investigate charter school effect on student achievement: School effectiveness theory, Market competition theory, and Social inequality theory. The racial/ethnic segregation effect and the socio-economic stratification effect were examined longitudinally and cross-sectionally. Datasets of primary and secondary public schools and county educational and demographic information covering 1998 to 2010 were obtained from multiple sources: the Common Core of Data from NCES, the Florida School Indicator Report, the Florida Department of Education, Florida Statistics Abstract, and Census Bureau. Hierarchical linear modeling was utilized to explore charter school effect in different organizational levels and hierarchical multivariate linear modeling was used to take into account the closely correlated relationships of the demographic compositions in public schools. The analyses of student achievement in charter schools and traditional schools indicated that charter schools and traditional public schools are significantly different from each other, and that the school characteristics were more influential on the school performances than county characteristics or the year effects, especially in the higher grades. Some charter schools achieved better in some subjects and grades in that they started at the lower scores than TPSs but grew faster during the period of 1998-2010. However, the charter school effectiveness turned out to be insignificant or even negative when control variables were introduced such as educational factors and demographic compositions. Market competition theory could not explain the variations in schools' FCAT scores, while social inequality theory explained them better. The findings of this study did not support the School Effectiveness Theory nor the Market Competition Theory in charter school movement. Instead, Social Inequality Theory was proved to be relevant to understand the differences in public school academic achievement. The analyses of segregation and stratification effects showed that charter schools were more racially and socio-economically segregated, and that they exacerbated the segregation and stratification in traditional public schools. The analyses of the Dissimilarity Index (DI) distribution among charter schools and TPSs revealed that the demographic compositions in charter schools deviate more from the county means than do TPSs during the period of 1998 through 2009. Charter schools had much lower proportion of free/reduced price lunch program students than TPSs in every school level, which was negatively related to the percentage of white students but positively to the percentage of black students. The years of charter school policy adoption in a county have similar effects on both groups: The longer it was since a county introduced charter school policy, the fewer black students and the more white students would enroll in charter schools. Overall, charter schools were likely used as pockets for white flight and self-isolation as well and exacerbated socio-economic stratification in public schools. The analyses of charter school DIs supported the warnings of white flight, self-isolation, and socio-economic stratification (Carnoy, 2000; Frankenberg, Lee, & Orfield, 2003; Rivkin, 1994). Findings of this study suggested that the increasing proportion of black students and free/reduced price lunch program recipients have enrolled in TPSs for all school levels along the years during the period of 1998-2009, but that the percentages of white students in TPSs have decreased year by year even though the rates are small. The analyses implied that charter schools were likely to locate around TPSs that had a higher proportion of a certain demographic group: The higher proportion of a certain demographic groups in a certain area would induce charter schools targeting these groups. Hierarchical multivariate linear models (HMLM) were introduced to detect the relative relationships between demographic groups. The multivariate analyses suggested that middle school charters were likely to locate around the TPSs with more white students and fewer Hispanic students, while elementary charter schools opened more around the TPSs with fewer black students. The location and targeting strategies of charter schools affected also the racial/ethnic distributions in high TPSs, even though the relationship got weaker. The proportions of free/reduced lunch program students in TPSs havd a consistently and significantly negative influence on the proportions of white students and a positive influence on the percentages of black and Hispanic students in TPSs. The academic performances of TPSs were highly and negatively related to the proportion of black students, while the relationship becomes much weaker to the percentage of white students and neutral to that of Hispanic students. The cross-sectional multivariate analyses suggested that charter schools created more racially segregated educational institutes in public education in Florida. The racial/ethnic compositions in TPSs were closely interrelated to the issues of the socio-economic stratification and residential division (Carnoy, 2000; Frankenberg, et al., 2003; Rivkin, 1994). The comparisons of the explained variance proportions by HMLM models and those of other models revealed that the percentages of white students were much more sensitive to the socio-economic and residential factors than the proportions of black students were, while the proportions of Hispanic students were much more sensitive to the charter school factors. The findings of this study highlighted the critical role of social context in public educational policies and the importance of policy design. This study rediscovered the old but important principle that charter school policy makers need to take into account the expectable but ignored or unintended consequences of the policy in public education system and the impacts of the policy on the non-choosers in TPSs as well.