Attending college can be very expensive. Research has demonstrated that costs associated with college (tuition, fees, room and board, etc.) and subsidies that help defray those costs (grants, scholarships, loans, etc.) can influence students' enrollment decisions. We also know that some students' choices are more heavily influenced by those costs and subsidies. What is not as clear is if college costs and subsidies influence students early in the college choice process. Policymakers and practitioners need to understand which students might not aspire to attend college and if finances is one of the factors.

Using the Educational Longitudinal Study of 2002, this study assesses whether or not students in the tenth and twelfth grades are affected in different ways by costs and subsidies associated with attending college. I do this by creating dependent variables in which students are placed into categories based upon both their predisposition to attend college and the level of importance costs and subsidies have in their decision-making process (referred to as predisposition/finances nexus measures). Using multinomial logistic modeling (MNLM), I test whether students' demographic, academic, and psychosocial characteristics predict which category of the joint indicator of college aspiration and financial concern students belong.

The results indicate that, while students overwhelmingly report that they do plan to attend college after high school, there are some characteristics that explain which students do report that they do not plan to attend college because it is too expensive. In general, these students are poor, White, male, not academically strong, and do not feel as though others in their lives want them to attend college. Additionally, in the twelfth grade, the differences between students who do not plan to go to college because of finances and those who report that they do plan to attend college and finances are very important in their decision making based on parent status, academic preparation and performance, and encouragement of others. Implications based on these results, possible survey improvements, and directions for future research are discussed.