By John Considine
It seems to me that there is an increase in the number of published papers using NBA Last Two Minute reports. These reports document the correctness of play-by-play calls and non-calls in the last two minutes of basketball games. It means that one has four distinct classes: correct calls; correct non-calls; incorrect calls; and incorrect non-calls. There is one major advantage of using incorrect calls and incorrect non-calls to examine referee bias. They exclude player behaviour and, therefore, allow us focus on the decisions of the officials.
If the errors made by officials are statistically correlated with variables like race, height, nationality, star-status and so on, then we can suggest that the officials are biased. A common thread in two 2023 papers is the testing of racial bias amongst referees. An NBER working paper by Naci Mocan and Eric Osborne-Christenson finds racial bias for one subset of calls dependent on the racial make up of the referee crew and the players involved. A Scientific Reports paper by Konstantinos Pelechrinis does not find support for racial bias – although this paper only examines a particular subset of the fouls.
The Pelechrinis paper is primarily concerned with home bias in the decisions of referees. This is also the subject of a fine 2022 Journal of Sports Economics paper by Hua Gong. The title of the paper is ‘The Effect of the Crowd on Home Bias: Evidence from NBA Games During the COVID-19 Pandemic’. As the title suggests, this paper is like many others that seek to use the lack of crowds during the pandemic to identify how crowds might influence officials. The evidence in this paper is that there is no referee bias in favour of home teams.
What I find more interesting is the error rate of officials. The average percentage of incorrect calls is 3.9% and the average percentage of incorrect non-calls is 6.0%. The higher rate of incorrect non-calls compared to incorrect calls is in keeping with the literature. There may be a slight concern over conflict of interest in that these reports are generated by the NBA. However, they are probably more informed than some of the crowd sourced “correctness” of calls data used in other papers.
If the errors made by officials are statistically correlated with variables like race, height, nationality, star-status and so on, then we can suggest that the officials are biased. A common thread in two 2023 papers is the testing of racial bias amongst referees. An NBER working paper by Naci Mocan and Eric Osborne-Christenson finds racial bias for one subset of calls dependent on the racial make up of the referee crew and the players involved. A Scientific Reports paper by Konstantinos Pelechrinis does not find support for racial bias – although this paper only examines a particular subset of the fouls.
The Pelechrinis paper is primarily concerned with home bias in the decisions of referees. This is also the subject of a fine 2022 Journal of Sports Economics paper by Hua Gong. The title of the paper is ‘The Effect of the Crowd on Home Bias: Evidence from NBA Games During the COVID-19 Pandemic’. As the title suggests, this paper is like many others that seek to use the lack of crowds during the pandemic to identify how crowds might influence officials. The evidence in this paper is that there is no referee bias in favour of home teams.
What I find more interesting is the error rate of officials. The average percentage of incorrect calls is 3.9% and the average percentage of incorrect non-calls is 6.0%. The higher rate of incorrect non-calls compared to incorrect calls is in keeping with the literature. There may be a slight concern over conflict of interest in that these reports are generated by the NBA. However, they are probably more informed than some of the crowd sourced “correctness” of calls data used in other papers.
RSS Feed
