A methodology for identifying the impacted groups in referee discrimination studies

Abstract

This article presents an empirical methodology that allows one to identify the group being discriminated against by sport referees. Reestimating Price and Wolfers (2010, PW) discrimination in foul-calling analysis with our methodology, we find evidence that only black players are discriminated against.

Keywords:

• discrimination
• basketball
• referees

JEL Classification:

• J15
• J44
• J71
• L83

Acknowledgement

We thank Jill Mccluskey for helpful suggestions.

Notes

¹ Two additional player fixed effects (George Zidek and Danny Young) were dropped because they were highly collinear with the other control variables. As in Price and Wolfers (2010, PW), a number of additional referee fixed effects were dropped because they were collinear with the other control variables.

² See Mittelhammer (1996, pp. 186–7).

³ There are relatively fewer observations on white players during games officiated by no white and three black referees (n = 1475) resulting in a relatively larger SE of the difference in the extreme cases (0.2447). However, the difference between three white and no black referees and one white and two black referees is also statistically insignificant.

⁴ Because we consistently subtract the differences from the extreme referee mix cases as three white and no black minus no white and no black referees for both black and white players, a decrease (increase) in the variable of interest for black players has a similar effect as an increase for white players.

⁵ Berri et al. (2006) created a winning margin index called win score and calculated it as win score = (points − field goal attempts − 1/2 free throw attempts) − turnovers + rebounds + steals + 1/2 blocks + 1/2 assists − 1/2 fouls. We analysed this productivity index per 48 minutes played.