College Football Attendance in the Long Run: Division II

Abstract

A balanced panel (52 teams over 38 years) is used to estimate fixed- and random-effects models for average season attendance. All variables are either stationary or cointegrated. Independent variables measure economic conditions, demographic characteristics, and team performance. Contrary to expectations, attendance is an inferior good while travel cost (real gas price per mile driven) is insignificant. Greater undergraduate enrollment increases attendance. Attendance decreases with rising county population in both models – one at ten percent probability. More wins in the current season and a greater number of playoff appearances in the last ten years increase attendance. Lifetime winning percentage is insignificant.