OBJECTIVE: Examine bidirectional associations between weight perception and weight change over time among adults.

METHODS: Data are from adult employees (N = 623) across 12 U.S. public high schools participating in a cluster-randomized multilevel weight gain prevention intervention. Data were collected at baseline, 12 months, and 24 months. Perceived weight status (very/somewhat underweight, just right, somewhat overweight, very overweight) were obtained via self-administered surveys. Weight (kg) was measured by trained staff. Change in weight was calculated as the difference between baseline weight and weight at each follow-up time point. Structural equation models were used to assess bidirectional associations of perceived weight status and change in weight over time. Models were adjusted for study condition, gender, age, race/ethnicity, education level, and previous time point.

RESULTS: The sample was 65% female with a mean age of 44.6 (SD = 11.3). Nearly two thirds of the sample consisted of people with overweight (38.8%) or obesity (27.3%). Structural equation models indicated that baseline weight predicted subsequent perceived weight status (beta = 0.26; P < 0.001), whereas baseline perceived weight status did not predict subsequent change in weight, adjusting for previous time point and covariates.

CONCLUSIONS: Results do not support bidirectional causality between weight perception and weight change in an adult sample.