College football attendance: a panel study of the Football Bowl Subdivision

The impact of race relations on NFL attendance: An econometric analysis

“…Shifting focus to the control variables based on prior studies examining the demand for sport, all models found that team performance as measured by lagged win percent were positive and significant for both the home and away team. These findings indicate that NFL fans were responsive to the quality of both teams in a content, and falls in line with previous studies that note the importance of the strength of both teams in determining fan interest in a sport contest [45,63]. Next, population was insignificant in all models, while the percent of minorities living a market was positive and significant.…”

“…In order to examine the research questions, we develop a model to estimate attendance demand guided by prior theoretical and empirical studies [40,41]. To begin with, empirical studies analyzing the demand for professional sport have typically used either reported attendance [40,42], the natural log of attendance [43,44], or the percent of stadium capacity [45,46] as the dependent variable in their models. In order to develop these dependent variables, attendance and capacity data was collected for every regular season game from the 2012 through 2017 NFL seasons.…”

“…When there are a significant number of sellouts in a sport league, it can result in biased estimates from a regression because it does not account for the potential that teams could have had greater demand if they were not constrained in terms of stadium capacity [62]. For example, in their examination of football attendance in the U.S., [45] utilized a Tobit regression because 25 percent of the games in their sample were sell outs, thus resulting in the need to censor right-hand side observations. For the sample within this study, home teams on average filled over 97 percent of their stadium capacity, thereby suggesting that use of an Ordinary Least Squares (OLS) regression is not a suitable method of estimation.…”

“…For the sample within this study, home teams on average filled over 97 percent of their stadium capacity, thereby suggesting that use of an Ordinary Least Squares (OLS) regression is not a suitable method of estimation. The common approach within the academic literature to address this issue is using a Tobit regression with the upper limit set at stadium capacity for each game, as this censors observations of sold out games [45,63,64]. Furthermore, as the Breusch-Pagan test was statistically significant (p < 0.01), there is need to account for the presence of heteroscedasticity [65].…”

College Football Attendance: A Panel Study of the Football Championship Subdivision

“…Although American football is widely regarded as the most popular spectator sport in the USA, there have been relatively few economic studies of attendance at college football games. Two studies (Price and Sen, ; Falls and Natke, ) focused on attendance at the highest level of the National Collegiate Athletic Association (NCAA) Division I‐A or the Football Bowl Subdivision (FBS). Price and Sen () used cross section game‐level data from a single season (1997), while the other used panel data across 7 years (2003–2009).…”