Assortative Mating for Body Weight: New Evidence from Multilevel and Fixed Effects Models

Julie H. Carmalt, Cornell University

Significant and positive interspousal BMI correlations have been used as evidence of assortative mating for body weight (e.g., active mate selection based on phenotypic preferences). However, resemblance in spousal BMI may also result from passive matching processes such as social homogamy or convergence. Using data from the Marital and Relationship Survey (N = 433 married couples), this study utilizes HLM and couple fixed effects regression to (a) provide new descriptions of spousal BMI association (e.g., How different, on average, are spouses’ BMI scores? Are husbands, on average, relatively heavier than wives? Do these outcomes vary by the mean weight of the couple?), and (b) exploit within-couple variation in BMI to determine whether resemblance in spousal BMI is a consequence of active initial assortment or passive matching processes. Preliminary results from models that control for within-couple unobserved heterogeneity suggest that factors other than initial assortment explain resemblance in spousal BMI.