Dyadic Person Similarity Predicts Similarity in Face Judgments

Here we share our data and code for our project Dyadic Person Similarity Predicts Similarity in Face Judgments.

Research shows that perceivers consistently extract information from faces to judge whether others are attractive, trustworthy, or dominant. However, there is also substantial variability among perceivers when making these social judgements. We seem to use our own dispositions as frameworks for making judgements about others, so could it be that two people who are similar in their dispositions also tend to make more similar social judgements? A consistent understanding of others may serve as a foundation for our social relationships, be it platonic or romantic. The present study investigated whether dyadic similarities in participants’ own dispositions are related to similarities in their social face judgements. 307 participants based in the UK rated 24 faces on six social traits. Participants also rated themselves on 13 social traits and completed a 60-item personality questionnaire. We computed dissimilarities between pairs of participants for face judgements, self-perceived trait ratings, and personality ratings, resulting in three separate dissimilarity matrices. Each entry in each matrix depicted pairwise distances in ratings between participants. Using representational similarity analysis, the three matrices were then correlated with each other. Results showed that both the self-perceived trait and personality matrices were significantly correlated with the face judgement matrix respectively. These results indicate that participants with more similar personalities and social trait dispositions were more likely to make similar social face judgements. Importantly, these associations were stable even when controlling for age, gender, ethnicity, and geographical location. These findings show that people who are more similar to each other also perceive the world in a similar manner and may form the basis for how we gravitate towards others, build connections and form relationships.

The study was approved by the Psychology Research Ethics Committee at City, University of London.

We pre-registered our study, specifying sample size, inclusion, and exclusion criteria (https://doi.org/10.17605/OSF.IO/ZR92G).

Here we share the raw anonymised data and the code we used for main analyses. The README file explains the folder contents, including what each script does. Raw data consists of csv files for each participant. Code runs in MATLAB.