In this dissertation I examine the relationship between subjective beliefs about the outcomes of educational choices and the generation of inequality of educational opportunity (IEO) in post-industrial society. Taking my departure in the rational action turn in the sociology of educational stratification, I argue that students facing significant educational transitions form their educational expectations by taking into account the foreseeable, yet inherently uncertain, consequences of potential educational pathways. This process of expectation formation, I posit, involves evaluations of the relation between the self and educational prospects; evaluations that are socially bounded in that students take their family's social position into consideration when forming their educational expectations. One important consequence of this learning process is that equally talented students tend to make different educational choices according to their family background. IEO thus appears to be mediated by the expectations students hold for their futures. Taken together, this research agenda argues that both researchers and policy-makers need to consider the expectation-based origin of educational inequalities if educational reform is to promote educational and social mobility in post-industrial society.

I pursue my research agenda in five chapters. In the introductory Chapter I I situate my research contributions in the tradition of the sociology of educational stratification. This chapter also outlines how the expectation-based explanation of IEO complements explanations stressing family resources as an important cause of IEO; it carefully defines "expectation," the core concept underlying the dissertation; it places the methodological contributions of the dissertation in the debate over the role of causal inference in social science; and it discusses the potential of the findings of the dissertation to inform educational policy.

In Chapters II and III, constituting the substantive contribution of the dissertation, I examine the process through which students form expectations for their educational futures. Focusing on the causes rather than the consequences of educational expectations, I argue that students shape their expectations in response to the signals about their academic performance they receive from institutionalized performance indicators in schools. Chapter II considers the impact of educational tracking on expectation formation among high school students in the U.S., whereas Chapter III analyzes the role of the grade point average in the expectation formation process among socially disadvantaged high school students in the U.S. The empirical analyses in both chapters strongly suggest that students rely on information about their academic performances when considering their educational prospects. The two chapters thus highlight that educational expectations are subject to change over the educational career, and that educational systems play a prominent role in students' expectation formation.

Chapters IV and V constitute the methodological contribution of the dissertation. Chapter IV develops a general method for decomposing total effects into its direct and indirect counterparts in nonlinear probability models such as the logistic response model. The method forms a solution to the issue of comparing coefficients across nested nonlinear probability models with different covariates; an issue that has troubled scholars of sociological methodology for decades. In Chapter V I productively apply the decomposition method to the issue of analytically conceptualizing and empirically quantifying the extent to which family background influences educational decisions among equally talented students. This influence, often referred to as the secondary effects of social stratification, is explained in terms of family background differences in the expected payoffs to schooling; that is, differences in students' educational expectations that arise by virtue of their families' positions in the stratification hierarchy. Applying the method to a cohort born in 1954 in Denmark, I find that roughly 60 percent of class differences in university completion can be attributed, although indirectly, to the secondary effects of social stratification. This analysis thus supports the conclusion, drawn for numerous countries around the globe, that educational expectations and their constituent subjective beliefs play a crucial role in the generation of IEO.