Abstract: PURPOSE: People spontaneously infer personality traits (e.g., lazy, extraverted) from body shapes when encountering strangers. In previous research, computer-generated bodies were derived from laser scanning of real humans to visualize and quantify body trait impressions. Research with American individuals indicated a multivariate trait space with dimensions relating to valence (good vs. bad) and agency (active vs. passive). The trait impressions were further predicted by body shape parameters. Do body-trait inferences vary by culture or are they universal? Our study examined whether the body-trait inference applied to Chinese participants.
METHODS: A total of 140 computer-generated bodies (70 females, 70 males) were rated by Chinese participants based on 30 personality traits. We visualized the structure of the body-trait space using correspondence analysis and predicted trait ratings based on body shape parameters using multiple linear regression.
RESULTS: Body-trait space showed that the first dimension was valence (Conscientiousness/Openness/Neuroticism), which was driven by body weight. The second dimension was Extraversion/Agreeableness, which was driven by whether a body appeared strong or typical. Additionally, body shapes predicted trait ratings with above-chance accuracy at both the trait profile and individual trait levels.
CONCLUSIONS: The study is the first systematic examination of body trait impressions in Chinese individuals. It highlights that forming body impressions is a universal phenomenon, but the structure of trait spaces, determinant body features, and prediction accuracy can vary by culture.

Key words: first impressions, data-driven approach, cultural differences, body perception