Essays on educational decisions, academic performance, and inequality of educational opportunity

This thesis investigates the determinants of inequality of educational opportunity (IEO) in the comprehensive educational systems of England and Sweden. Reducing IEO remains a pressing problem for industrialized countries with democratic regimes. Despite major governmental efforts for widening participation in post-compulsory education, it is still possible to observe substantial socioeconomic gaps in educational achievements and attainments. Sociologists analysing IEO tend analytically to separate ‘performance’ from ‘choice’ effects. Performance effects are all those, independently of being determined by genetic or socio-cultural traits, that are expressed in socio-economic differences in academic performance (such as grades, or school marks). Such indicators of performance have a strong effect on educational transitions. Choice effects are in turn all those that are expressed in actual educational decisions for students at similar levels of academic performance. Over the last three decades, research in this area has shown that performance effects are relatively larger than choice effects and that this difference has widened among younger cohorts of students. Importantly, quantifying the relative importance of these two mechanisms can inform and help policy makers to decide ‘when’ and ‘what’ type of interventions should be implemented to reduce inequality. However, the current methodology utilized to quantify the relative importance of performance and choice effects presents both technical and conceptual limitations. First, the standard decomposition methods are likely to yield biased estimates due to confounding and scaling bias. Second, the model of performance and choice effects stresses micro-behavioural mechanisms at the individual level, disregarding the role of school organizational factors. These limitations compromise the possibility of drawing valid inferences and, thus, of informing policy makers. With these concerns in mind, over the course of three empirical chapters, I firstly explored the weaknesses in the standard decomposition approaches used in this field. Secondly, I estimated the relative importance of performance and choice effects using novel data from Children on Immigrants Longitudinal Survey in Four European Countries (CILS4EU) and causal inference approaches. Finally, I analysed the role of ability grouping, i.e., how the organisation of learning environment, affects students’ performance and choices. The findings presented in this thesis make a number of important advances that arguably have implications for policy. First, I provide a more comprehensive framework and guidelines to quantify the relative importance and performance effects. Second, I show that total IEO has continued to decline in England and Sweden for students born in 1995/6. However, once more robust decomposition methods and sensitivity analysis were employed, I found that choice effects are in fact substantially more important than what has been claimed in past research. Third, employing several identification strategies, I show that ability grouping does not affect students’ performance and choices in the context of comprehensive educational system as the ones of England and Sweden. These results indicate that policies aimed at changing the learning environment of schools are not very efficient for tackling IEO. Instead, it would probably be more fruitful to implement policies aimed at reducing inequality in early-life circumstances as well as policies aimed at reducing opportunity costs in access, and guiding students to make more efficient choices.