Implicit Biases in Refereeing: Lessons from NBA Referees

A 2022 study by researcher Konstantinos Pelechrinis, published on the arXiv preprint server, applied statistical analysis to examine implicit biases in NBA officiating. The research used extensive datasets, including official NBA play-by-play data and the league's Last Two Minutes reports for close games from the 2015 season onward. The goal was to quantify biases that might occur in the split-second decisions referees must make, moving beyond anecdotal evidence with robust statistical methods.

The analysis focused on three potential biases: home-vs-away, bias towards individual players or teams, and racial bias. The results were nuanced. The study confirmed a statistically significant home team bias, which was particularly pronounced during playoff games. Interestingly, this bias was observed to have reduced following the COVID-19 pandemic, when games were played in empty or neutral-site "bubble" environments, suggesting crowd influence may be a factor.

Crucially, the research found no statistical evidence for two other hypothesized biases. While the data indicated that certain individual players consistently received more favorable calls than chance would predict, there was no corresponding evidence of a negative bias against specific players or teams. Most significantly, the analysis found no evidence of racial bias in the officiating decisions examined, a finding that challenges some common perceptions.