Academic journal article Contemporary Economic Policy

Social Interactions and College Enrollment: Evidence from the National Education Longitudinal Study

Academic journal article Contemporary Economic Policy

Social Interactions and College Enrollment: Evidence from the National Education Longitudinal Study

Article excerpt


Many groups of adolescents continue to enroll in college at low rates. There are particularly large gaps between high- and low-income students and between minority and majority groups (Perna and Swail 2002: NCES 2005). While the rate of return for an additional year of schooling is usually estimated to be between 8% and 10%, the return has been shown to be highest for individuals in the middle of the socioeconomic status distribution (Ashenfelter and Rouse 2000). (1) Research has also documented that there are little or no differences in the returns to schooling across racial groups (Barrow and Rouse 2005). The high rewards associated with college and relatively low attendance by some groups has directed many researchers to attempt to explain the disparities in college enrollment.

In most economics research, investigators appeal to a number of constraints faced by students in explaining low college participation rates, in particular financial constraints, lack of information, lack of academic ability, and attending low-quality schools. However, credit constraints and other financial factors among high school graduates have been shown to be of limited importance in explaining individuals' choices of attending college. (2) Pre-college informational interventions have also experienced limited success (Perna and Swail 2002), which implies that information asymmetries across groups are not the whole story in explaining college enrollment disparities. While Cameron and Heckman (1998, 2001) emphasize the importance of ability differences that appear through early differences in family resources, there are still large unexplained differences in college enrollment for individuals with similar measured ability levels. The remaining unexplained differences in college enrollment suggest that pursuing alternative explanations may provide additional insight into the current disparities in college enrollment.

One explanation of researchers' inability to account for low college enrollment rates of some groups is that most research does not include all the relevant aspects of an adolescent's college decision process. In their seminal economic analysis of adolescent college enrollment decisions, Manski and Wise (1983) suggest five factors that are most important in determining enrollment: (1) academic aptitude, (2) family income, (3) cost and aid. (4) quality of high school and decisions of peers, and (5) labor market conditions. This list of important factors has been used in decades of research of college enrollment by economists and other educational researchers (Avery and Hoxby 2004; Cameron and Heckman 1998; Haveman and Wolfe 1995; Rouse 1994). An examination of economic research on the question of college enrollment, however, also indicates that one factor is noticeably missing from most analyses--the decisions of peers.

There are several reasons why the decisions of peers have been omitted from previous research on college enrollment. First, most data sets contain limited information on the decisions of peers. Second, most scholars appeared to believe that peers' choices were not as important as the other factors in determining college enrollment. Third, there is no consensus on the relevant group of peers--it could include friends, neighbors, siblings, classmates, etc. Finally, there are well-known methodological issues of empirically examining interdependent decisions (Manski 1993).

In this paper, I examine whether interdependent choices, called social interactions in the economics literature, can help explain the dramatic differences in college enrollment choices among individuals with similar characteristics (Brock and Durlauf 2001a, 2001b; Manski 2000). In order to begin to estimate credible social interactions in college enrollment choices, this paper proposes an instrumental variables approach to overcome the difficulty of disentangling different types of social effects--the so-called "reflection problem" identified by Manski (1993, 2000). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.