Implicit evaluations of faces depend on emotional expression and group membership

Faces carry a lot of information influencing evaluative reactions, such as emotional expression, age, or group membership. Even though, typically, many of these aspects will be present in a face concurrently, only few studies have examined automatic evaluative reactions to faces that vary on more than one dimension. As an exception, two recent priming studies examined the concurrent influence of group membership and emotional expression. Quite astoundingly, they leave the reader with two divergent outcomes: while Weisbuch and Ambady (2008) observed an interactive influence of emotional expression and group membership on evaluative reactions, Craig et al. (2014) found that group membership did not contribute to the implicit evaluation of positive and negative emotional expressions. In order to shed light on this matter, we conducted three high-powered experiments using prime images of highly relevant in-group and out-group members expressing happiness and fear. We furthermore varied the social context of the priming task in order to give the “interaction hypothesis” a chance. However, we found no evidence for the interaction reported by Weisbuch and Ambady. In contrast to Craig et al., we found that both emotional expression and group membership independently contributed to implicit evaluations. Differences are discussed in terms of relevance of the employed groups, test power, and the time-scale of underlying processes.