Evaluative Conditioning (EC) describes changes in liking that occur when a conditioned stimulus (CS) is paired with an unconditioned stimulus (US). Sperlich and Unkelbach (2024) showed that EC studies generally do not exhibit a negativity bias, such that negative information (e.g., images of aggressive animals) does not typically have a stronger impact than positive information (e.g., images of cute animals) on CS (e.g., geometric shapes) ratings. They offered a cognitive-ecological explanation for this pattern: because positive and negative USs (e.g., cute or aggressive animals) are usually presented with equal frequency in EC paradigms, no bias emerges. However, when positive and negative USs differ in frequency, valence asymmetries arise. In particular, rare negative USs produce a negativity bias (i.e., more negative ratings of CSs paired with rare negative USs than with frequent negative USs), whereas rare positive information leads to a positivity bias (i.e., more positive ratings of CSs paired with rare positive USs than with frequent positive USs).

The present research examines whether these valence asymmetries, previously observed for individual stimuli, also extend to groups. Across four pre-registered online experiments (N = 2,404), we tested whether rare group members (aliens or soccer players; CSs) paired with negative or positive animal pictures (USs) influence evaluations of other paired (Study 2) or non-paired group members (Study 1, 3, and 4). Across all experiments, the results consistently show that individual group members exert a disproportionate influence on evaluations of the group as a whole. In other words, rare stimuli exert an influence on group evaluations comparable to that of frequent stimuli.