The relationship between siblings’ college choices: Evidence from one million SAT-taking families

Joshua Goodman,Michael Hurwitz,Jonathan Smith,Julia Fox

doi:10.1016/j.econedurev.2015.05.006

Joshua Goodman, Michael Hurwitz + Show 2 more

Open Access

https://doi.org/10.1016/j.econedurev.2015.05.006

Copy DOI

Abstract

Recent empirical work has demonstrated the importance both of educational peer effects and of various factors that affect college choices. We connect these literatures by highlighting a previously unstudied determinant of college choice, namely the college choice made by one's older sibling. Data on 1.6 million sibling pairs of SAT-takers reveals that younger and older siblings’ choices are very closely related. One-fifth of younger siblings enroll in the same college as their older siblings. Compared to their high school classmates of similar academic skill and with observably similar families, younger siblings are about 15–20 percentage points more likely to enroll in 4-year colleges or highly competitive colleges if their older siblings do so first. These findings vary little by family characteristics. Younger siblings are more likely to follow the college choices of their older siblings the more they resemble each other in terms of academic skill, age and gender. We discuss channels through which older siblings’ college choices might causally influence their younger siblings, noting that the facts documented here should prompt further research on the sharing of information and shaping of educational preferences within families.

Highlights

Researchers have been modeling students’ college enrollment decisions for decades (e.g. Fuller, Manski and Wise, 1982)
There is substantial evidence that students don’t have complete information or, at the very least, have different amounts of information (Dillon and Smith, 2013; Hoxby and Turner, 2013). This implies that the variation in college choice is left largely unexplained, or statistically speaking, the R-squared is quite low
This paper explores the potential to improve upon existing models and, more generally, to better understand college choice by documenting the similarities in college enrollment patterns between younger and older siblings

Summary

Introduction

Researchers have been modeling students’ college enrollment decisions for decades (e.g. Fuller, Manski and Wise, 1982). Fuller, Manski and Wise, 1982). It has proven a difficult modeling problem for at least three reasons. There are thousands of colleges, each with countless attributes, most of which are unobservable to the econometrician. There is substantial evidence that students don’t have complete information or, at the very least, have different amounts of information (Dillon and Smith, 2013; Hoxby and Turner, 2013). Combined, this implies that the variation in college choice is left largely unexplained, or statistically speaking, the R-squared is quite low.

Methods

Results

Conclusion