This dissertation attempts to answer the question "Why do some adolescents pursue college while others do not?" In attempting to answer this question my focus is on what I call college-encouraging preferences and perceptions, which are preferences and perceptions that make adolescents more likely to pursue college. The model I develop engages the rational choice literature in both economics and sociology, but it deals primarily with considerations outside the scope of traditional rational choice models. I deal with preferences and perceptions but not only those relating to pecuniary costs and benefits. Also, unlike most rational choice perspectives, I focus on interpersonal variation in preferences and perceptions and how this variation affects college entry decisions. I analyze two preferences and two perceptions: preferences for academic activities, preferences for various labor-market outcomes, perceptions of the ability to complete college, and perceptions of the effect of education on labor-market outcomes. Using both propensity-score matching and regression estimators, I find that preferences for academic activities increase educational expectations, preferences for labor-market outcomes that education improves increase educational expectations and college entry, and that the subjective probability of college completion conditional upon college entry increases college entry. Regarding perceptions of the effect of education on labor-market outcomes, I turn away from the maximization assumption of traditional human capital approaches and develop a simple satisficing model. Consistent with predictions of the model, results show that the more education an adolescent believes is required to enter the occupation they expect, the more likely they are to enter college. Examination of reverse causality (i.e., educational decisions affect occupational decisions) found only weak effects. Analysis of the determinants of preferences and perceptions shows that cognitive skill and parental education are both positively related to most college-encouraging preferences and perceptions. Blacks, Hispanics, and Asians have more college-encouraging preferences and perceptions than whites of comparable cognitive skill and family background. Multilevel models also suggest that high schools influence the preferences and perceptions of the students within them.