An Assessment of Demand Variables on Division I College Football Series (FBS) Bowl Series Attendance (2016)

Undergraduate: Rhett Jackson

Faculty Advisor: Nels Popp
Department: Exercise & Sport Science

---

This past season five new college football bowl games were created despite a decrease in bowl attendance from the previous seasons (Solomon, 2016). With attendance decreasing, bowl organizers have more pressure for creating matchups that will draw crowds and maximize revenue. This paper develops a regression model predicting demand for Division I-A college football post-season bowl games. Ten key predictor variables were identified and collected from all 113 bowl matchups between the 2013-2015 seasons; (a) Average Home Attendance, (b) Average Temperature, (c) Bowl Age, (d) Bowl City Population, Distance, Saragin Power Rankings,Winning Percentage, Local Population of Competing School(s), Previous Five Year Bowl History, and Ticket Price. Utilizing a stepwise entry process, five of the variables produced a statistically significant regression model, which predicted 60.7 percent of the variance in bowl game attendance. The significant variables included: (a) Average Home Attendnace, (b) Bowl Age, (c) Distance, (d) Bowl City Population, and (e) Saragain Power Rankings. All significant coefficients were positive, with the exception of distance travelled, which was negative, as expected. The results of the study can be used to assist bowl game committees in determining the optimal bowl matchups yielding the highest attendance for future bowl games. In fact, the final model was utilized to predict attendance for possible alternative bowl game matchups from the 2015 bowl season.