A Discrete-Choice Model of a College Football Recruit’s Program Selection Decision

Abstract

We formalize a discrete-choice model of program selection from the view of college football recruits. With a conditional logit model, we correctly predict the recruit’s preferred college for 65% of the 19,815 individuals, besting a purely random guess method (21%). We focus on the role football will play in postcollegiate careers of high-rated, mid-rated, and low-rated prospects in choosing a school. High-rated and mid-rated recruits value historical on-the-field success, historical head coach success, stadium capacity, media exposure, facilities, and coaching expenditures. Academic considerations and more recent on-the-field success, however, are more dominant factors for low-rated recruits.