Implicit bias can emerge in the form of ​undifferentiated impressions,​ where perceivers’ trait ratings for a target are biased to be strongly correlated with each other (Oh et al., 2020). This happens because the perceivers use one trait cue to infer the rest of the target’s traits. I explore how biases can be revealed in intercorrelation structures of trait impressions.

In one study, I tested intercorrelations of trait impressions for evidence of a gendered-racial bias, where people might perceive Black women as more undifferentiated than White women and Black men. To do so, I used principal component analyses to compare intercorrelations between 14 trait ratings for Black women, Black men, White women and White men. While there were no significant differences in mean trait ratings between targets, analyses revealed that ratings of Black women were more intercorrelated than all other race/gender groups. Moreover, the first principal component for trait ratings for Black women explained more variance than it did for trait ratings for all other race/gender groups. These results suggest that impressions of Black women are more biased than impressions of White women, and that bias that is not revealed in mean trait ratings can be revealed in the form of undifferentiated trait ratings.

Redemption can be difficult for past-offenders to earn because the public often infers that they are likely to reoffend and are hesitant to give them a second chance. In this line of research, I have explored how appearances - specifically, incidental changes in appearance - influence moral judgments about transgressors. Across several studies, I have found evidence that people consistently infer that past-offenders who look different than they did before have changed for the better, are more remorseful, more kind, and more trustworthy than their unchanged counterparts. 

Interestingly, participants often spontaneously commented that they do not make judgments based on appearance, yet proceeded to do so. Moreover, changes in personality and other characteristics such as career or food preferences did not have the same effect, although changes in personality could arguably be considered more valid information to base judgments on. This suggests that new information can have a stronger influence on perceivers’ impressions if it is information that perceivers do not intentionally incorporate into their judgments. In future work, I intend to investigate how people think about appearances as authentic reflections of essences, or true selves. 

Gossip can quickly and effectively stabilize cooperative norms in organizations and groups. In most situations, gossip sparks reputation concerns that make people keen to cooperate. But do people care if their rivals gossip about them? Traditionally, the premise of gossip’s key role in cooperation has been tied to reputational concerns, which—in theory—can only function within established ingroups. 

In this work, we found that gossip increases cooperation across group boundaries: people who were interacting with strangers that were not part of their ingroup were more likely to act prosocially if they thought they were going to be gossiped about. Moreover, we found that gossip had no effect on cooperation within ingroup members: people interacting with others who were already part of their established ingroup were no more or less likely to act prosocially in response to the threat of gossip. Accordingly, this work suggests that being part of an established group encourages cooperative norms within the group; but in the absence of shared group identity, gossip serves an alternative social tool for enforcing cooperation and evoking group-iness. 

In another line of work, I am investigating how people derive social connection and meaning in life from the groups they are a part of—past and present (Kim, Austin, Cian, & Adams, 2023). We wondered whether ancestral social ties – as in those documented by DNA kits – afford the same feelings of social connection and meaning in life as current social ties. Using a cross-lagged design, we compared how people’s past social ties (i.e. ancestral heritage) versus current social ties (i.e. current social network) can change feelings of social connection and meaning in life over time. Our research suggests that learning about ancestral social ties increases people’s experiences of social connection, but not any more than present-day social ties.