Using the Language of elite athletes to predict their personality and on court transgressions

Abstract

Personality is predictive of many behaviors, but personality questionnaires cannot be readily administered to persons of interest. The language people use to express themselves can often predict personality and so should, in theory, provide a surrogate marker for predicting behavior. We used social media (Twitter) language from a sample of 252 NBA players to estimate their Five Factor personality scores, and then, used these scores to try and predict on-court transgressive behavior. A machine learning model was able to predict players’ tendency to commit technical fouls (predictive performance: r =.18); with the most important contributors to the model including neuroticism, extraversion, and conscientiousness. These findings show that personality can predict individual choices and behaviors in specific contexts; furthermore, by assessing the degree to which our digital footprint can be used to derive actionable predictions of behavior, the current findings could inform discussions concerning regulations of data privacy.