Determinants of Attendance at Division I Women's College Basketball Games: The Superstar Effect

This paper provides the first comprehensive empirical analysis of the determinants of attendance in NCAA Division I women's basketball. We use a national game-level dataset from the 2023-24 season and a separate longitudinal panel covering 2015-16 through 2023-24. Using OLS, Tobit, and GLM, we show that attendance-measured as a fraction of arena capacity-is strongly associated with team success, recent postseason performance, and overall program quality. Game timing and local sports competition exert smaller but statistically meaningful effects. The largest single-game attendance impacts are driven by four prominent school-player pairings: Iowa/Caitlin Clark, UConn/Paige Bueckers, LSU/Angel Reese, and USC/JuJu Watkins. To isolate superstar effects from underlying school strength, we apply Synthetic Difference-in-Differences methods to the longitudinal panel for Iowa (Clark), USC (Watkins), UConn (Bueckers), and LSU (Kim Mulkey and Reese). We find large and robust causal effects for Clark and Watkins: Clark increased Iowa's home attendance by approximately 5,470 fans per game between 2021 and 2023, while Watkins raised USC's attendance by about 3,130 fans per game in 2023. In contrast, we find no statistically significant attendance effect for Bueckers at UConn in 2023. LSU experienced a major structural break in 2021 following Mulkey's arrival, and the incremental effect of Reese in 2022 cannot be separately identified due to staggered treatment dynamics that distort SDiD counterfactuals. Overall, the results indicate that superstar athletes can generate large, immediate demand shifts, though effects vary substantially across program contexts and baseline demand levels.