Selection bias in social facilitation theory? Audience effects on elite biathletes' performance are gender-specific

Abstract Social facilitation proves robust in conditioning tasks (e.g., running), yet in coordination tasks (e.g., rifle-shooting) some studies report performance deterioration. Recent Biathlon World Cup data offered the unique opportunity to test this task-specificity (conditioning = cross country skiing, coordination = rifle-shooting). Audience restrictions due to COVID-19 allowed to compare athletes' performance in the absence (2020) and presence (season 2018/2019) of an audience. Gender-specific regulations (e.g., course length) necessitated the inclusion of gender as additional factor. Results of 83 (sprint competition) and 34 (mass start competition) biathletes revealed that task-specific social facilitation is moderated by gender: in the presence of an audience male biathletes showed performance improvements in the conditioning task and performance deteriorations in the coordination task; female biathletes showed the reverse pattern. This gender dependency may have gone unnoticed in the past due to sample selection bias (