There is ample evidence that watching sports induces strong emotions that translate into manifold consequential behaviours. However, it is rather ill-understood how exactly spectators’ emotions unfold during soccer matches and what determines their intensity. To address these questions, we used the 2018 FIFA World Cup as a natural quasi-experiment to conduct a pre-registered study on spectators’ emotional experiences. Employing an app-based experience-sampling design, we tracked 251 German spectators during the tournament and assessed high-resolution changes in core affect (valence, activation) throughout soccer matches. Across the three German matches, multi-level models revealed that all spectators exhibited strong changes on both affective dimensions in response to Germany’s performance. Although fans experienced slightly more intense affect than non-fans, particularly during losses, this moderating effect was very small in comparison to the magnitude of the affective fluctuations that occurred independent of fan identity. Taken together, the findings suggest group emotions (collectively felt emotion irrespective of individual affiliation) rather than group-affiliation based emotions (individually felt emotion because of an affiliated group), as the dominant process underlying spectator affect during the 2018 FIFA World Cup.

Social information, including faces and human bodies, holds special status in visual perception generally, and in visual processing of complex arrays such as real-world scenes specifically. To date, unbalanced representation of social compared with nonsocial information in affective stimulus sets has limited the clear determination of effects as attributable to, or independent of, social content. We present the Complex Affective Scene Set (COMPASS), a set of 150 social and 150 nonsocial naturalistic affective scenes that are balanced across valence and arousal dimensions. Participants ( n = 847) rated valence and arousal for each scene. The normative ratings for the 300 images together, and separately by social content, show the canonical boomerang shape that confirms coverage of much of the affective circumplex. COMPASS adds uniquely to existing visual stimulus sets by balancing social content across affect dimensions, thereby eliminating a potentially major confound across affect categories (i.e., combinations of valence and arousal). The robust special status of social information persisted even after balancing of affect categories and was observed in slower rating response times for social versus nonsocial stimuli. The COMPASS images also match the complexity of real-world environments by incorporating stimulus competition within each scene. Together, these attributes facilitate the use of the stimulus set in particular for disambiguating the effects of affect and social content for a range of research questions and populations.