Beyond School Effects: Private Schooling, Segregation and Standardization of School Systems in Latin America

Introduction

A considerable body of studies have shown that specific institutional arrangements in educational systems help in understanding cross-national differences in educational outcomes (Woessman 2003, Brunello and Checchi 2007, Bol and van de Werfhorst 2011, Bol et al 2014, Mijs 2016). The design of educational institutions may face policy trade-offs in the tasks of school systems that are served by them. Deregulation as privatization and school autonomy may enhance efficient sorting of students and then maximize learnings but at the cost of exacerbating social inequalities. A centralized education system may guarantee equality of educational opportunities, but it is not clear if it increases or hinder the overall performance level (Bol and van de Werfhorst 2011, Pedró et al 2015). This study is aimed to fill this gap, first, by departing from the widely supported assumption that the organization of educational systems affect, partly, the educational outcomes of students. Second, I focus on developing countries –Latin American countries- as we know less about the impact of institutions in educational outcomes in the region. Third, I attempt to address the following research question: what is the effect of the level of privatization increase achievement inequalities, given the level of standardization of the school systems in Latin American countries?

Data and Methods

I analyse data from the Tercer Estudio Regional Comparativo y Explicativo (TERCE), implemented in in year 2013 by UNESCO office in Santiago, Chile. TERCE is the most recent large-scale assessment that exclusively cover students and schools in Latin American countries. Fifteen countries participated: Argentina, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru and Uruguay. Following the comparative research body, I use two-level hierarchical model to account for the multilevel structure of the data, students nested in schools, with random school effects and country fixed effects, to identify variability in the educational institutions of interest, given the small number of countries. The dependent variable is performance in mathematics in 6th grade as it is the subject that is most clearly learned at schools (Coleman 1975, Bol et al 2014), and seems to be more sensitive to socioeconomic background than other subjects. The main predictor is socioeconomic status of student, and I control for several indicators related to learning home environment, and sociodemographic variables. At school level, I control for several organizational characteristics and social composition.

Findings
The main findings show that country-specific configurations of school systems are associated
to difference in mathematics achievement. Differences between schools in performance are partly explained by differences at country level. In this regard, I have chosen two important dimensions of school systems, for the Latin American region: the level of standardization and privatization. These results confirm some recent findings that achievement inequality is larger in school systems with a great level of differentiation between schools, in which the stratification triggered by the private sector is one important indicator (Chmielewski and Reardon, 2016).
In case of the level of standardization results show that is associated to a lesser degree of achievement inequality. However, the main finding points to persistent inequalities as much as private sector in school systems is bigger. Further, the models predict that these inequalities are not decreasing as the standardization level increases. In this respect, I also find support for a diminishing effect of policies that points to equalization of opportunities (Woessman 2003, Bol and van de Werfhorst 2011). The results suggest higher inequalities as the stratification induced by private school sector increases. These effects are still significant after adding school level controls, which suggest that over and above school processes, uneven between school sorting induces by private sector.