Association between late-night tweeting and next-day game performance among professional basketball players

Objectives: In sports, decreased sleep duration is generally associated with poorer performance compared to adequate or enhanced sleep duration. Yet, these findings have primarily been taken from small numbers of athletes performing outside of real games or competitions. It remains unknown how acute decreased sleep duration impacts real-game performance among professional athletes. Here, we merged 2 publicly available datasets to jointly measure late-night social media activity (a proxy for sleep deprivation) and next-day game performance. Setting: Professional basketball competition. Participants: 112 players from the National Basketball Association. Measurements: Time-stamped social media activity and in-game individual performance statistics. Results: Late-night tweeting (compared to not late-night tweeting) is associated with within-person reductions in next-day game performance, including fewer points scored and fewer rebounds. However, we also observe less time played per game following late-night tweets and decreases in the negative outputs of turnovers and personal fouls. The critical measure of shooting accuracy – which is not time dependent – provides the clearest evidence of a performance penalty following late-night tweeting activity (between 11:00 PM and 7:00 AM); players successfully make shots at a rate 1.7 percentage points less following late-night tweeting. Conclusions: Our findings suggest that acute sleep deprivation, as measured via late-night Twitter activity, is associated with changes in next-day game performance among professional National Basketball Association athletes. More broadly, the use of late-night social media activity may serve as a useful general proxy for sleep deprivation in other social, occupational, and physical performance-based contexts.