Many authors (e.g. Woessmann et al. 2009; Woessmann 2008, Betts and Roemer2007) agree that the educational achievement of students should be independent of their family socioeconomic background. In this concept the educational inequality should be tolerated only if it results from differences in effort, not if it reflects circumstances that are beyond a person’s control – including the socioeconomic background of their parents, peer effects or teacher effects. At the same time, education policy has shifted more towards choice policies in the European systems, by aiming for efficiency it has unintentionally created more inequality. The effect of the choice on equity is hotly debated in empirical and theoretical literature. It is shown in many cases (Australia (Narodowski 2002), Germany (Riedel et al. 2009), various US cities (Lankfort and Wyckoff 2001; Bifulco et al. 2009), England (West 2006; Burgess et al. 2006), and even Finland (Seppänen 2006, 2003)) that choice tends to cumulate better socio-economic background students into certain schools, creating not only positive peer effects, but also a negative externality to the disadvantaged students. Uncontrolled choice has increased ethnic and social segregation that can cause even more social costs than residential segregation (Bifulco et al. 2009). Furthermore, due to limited economic, cultural and social resources school choice tends to constrain disadvantaged families from practicing choice.
The Estonian education system, having its roots in both the elitist pre-World War II and the egalitarian soviet post-World War II, previously relied on a catchment area based model with some exceptional specialised extra curricula schools. After the gradual changes in education policy concerning school choice, an increasing degree of urban families are investing more time and other limited resources into the pre-training processes of their children. Moreover, we see that schools that are highly selective in entrance concentrate high-SES students into an oases in the limited number of downtown schools. Thus, we are studying the consequences of latent school choice development in Estonia. We show that as a result of incremental latent changes, individual educational returns become significantly dependent from the socioeconomic background characteristics. Furthermore, we indicate that these returns are reliant on institutional characteristics, like peer effects, which are enforced by ability selection to the primary level of education. These changes have been the result of recent amendments of policy towards more choice, allowing school autonomy in selecting their intake, without any centrally designed or controlled criteria.
We operationalize student educational returns by using the PISA 2009 disaggregated dataset. 4727 observations are available, however missing data cannot be imputed, and thus the actual number of observations is smaller. Independent variables are also taken from PISA: background characteristics are measured by eleven dimensions i.e. single parenthood, parents’ job status and country of birth, language at home, home possessions. The independent school effect is operationalized by ownership (private or public), admission criteria, extra curriculum activities, existence of standardised testing and language. We also control for the educational resources at home, gender, pre-primary education, reading pleasure, out of school training and relations with teachers.