Investigating Sports Commentator Bias within a Large Corpus of American Football Broadcasts

TL;DR

This study analyzes a large corpus of American football broadcasts to investigate commentator bias, especially racial bias, using automated annotations and linking with racial metadata, supporting prior social science findings.

Contribution

It introduces the FOOTBALL dataset with extensive transcripts and annotations, enabling large-scale computational analysis of commentator bias over six decades.

Findings

Supports prior social science conclusions on racial bias

Identifies confounding factors in bias analysis

Provides a large, annotated dataset for future research

Abstract

Sports broadcasters inject drama into play-by-play commentary by building team and player narratives through subjective analyses and anecdotes. Prior studies based on small datasets and manual coding show that such theatrics evince commentator bias in sports broadcasts. To examine this phenomenon, we assemble FOOTBALL, which contains 1,455 broadcast transcripts from American football games across six decades that are automatically annotated with 250K player mentions and linked with racial metadata. We identify major confounding factors for researchers examining racial bias in FOOTBALL, and perform a computational analysis that supports conclusions from prior social science studies.