<p>Robust averaging is an analytic feature of our perceptual systems that adaptively downweights outlying information during information processing. Here, we test whether individuals demonstrate robust averaging for a critical source of social information—facial affect—whether it is altered by psychotic-like experiences, and whether it is associated with social connection (the positive sense of relatedness from relationships and perceived/received support and inclusion). Participants completed a novel face averaging task in which they judged whether face arrays that varied as a function of reliability (variance of the faces), strength (emotional intensity of the faces), and valence (positive or negative), were on average more positive or negative. Afterwards, participants completed self-report measures of psychotic-like experiences and social connection. Two analytic approaches revealed the presence of robust averaging for emotional faces whereby inlying faces (i.e., those closer to the mean emotion expression of the face array) were given greater weight compared to outlying faces on trial-by-trial decisions. This effect was specific to high variance trials. There were no associations between robust averaging and social connection or psychotic-like experiences. These findings suggest individuals use robust averaging as an adaptive strategy to summarize social information, although any clinical and behavioral implications of individual differences remain to be clarified.</p>