Physical exercise and interviewer-rated appearance: longitudinal evidence from the China family panel studies, 2010–2022

Abstract Physical appearance is an important form of social capital, yet limited longitudinal evidence exists on whether it can be shaped by modifiable behaviors such as physical exercise. Using seven waves of data from the China Family Panel Studies (CFPS) collected between 2010 and 2022, this study examines whether changes in exercise behavior are associated with changes in interviewer-rated physical appearance. We construct an unbalanced individual-level panel and estimate fixed-effects models to exploit within-person variation over time while accounting for time-invariant unobserved heterogeneity and survey-year effects. Physical appearance is measured using interviewer evaluations on a 1–7 scale, and physical exercise is primarily measured by weekly exercise frequency. The results show that exercise frequency is positively and significantly associated with appearance ratings: each additional exercise session per week corresponds to a modest increase in interviewer-rated appearance. This positive association remains stable across sequentially adjusted model specifications. However, when total weekly exercise duration is used as an alternative measure, the estimated coefficient is no longer statistically significant. The estimated association is somewhat larger among women than among men, although the gender difference is not statistically significant. In addition, lagged exercise frequency is not significantly associated with current appearance ratings, suggesting that the appearance-related benefits of exercise may be more immediate than persistent. This study contributes to the literature by improving identification relative to prior cross-sectional research through panel fixed-effects estimation and by using interviewer-rated appearance. The findings provide longitudinal evidence that physical exercise is associated with external appearance evaluations in a manner that is relevant to real-world social interactions.