The heritability of depressive symptoms: multiple informants and multiple measures

Background: Earlier research suggests large rater differences in heritability estimates of children's depressive symptoms in the context of significant age and sex-limitation effects. Methods: With data from an ongoing, population-based twin-family study, we estimated genetic and environmental influences on children's depression with models allowing for sex-specific effects. Our assessments of twins included self-reports and ratings made by their classmate peers, their parents and their teachers, allowing estimates of genetic and environmental effects with data from different informants. Model-fitting used maximum likelihood estimation of log-transformed data from a sample of 1,366 11- and 12-year-old twin pairs. Results: Estimates of additive genetic effects were significant for both boys and girls across all four informants, and, standardized to reflect the percentage of phenotypic variance accounted for, those estimates ranged from .28 to .71. Significant effects from common environmental sources were found in ratings of teachers and parents, where, typically, one individual rated both co-twins, but neither in peer nominations nor self-ratings. The correlation of teacher and parent ratings was modest, and bivariate model-fitting found no correlation in either genetic or shared environmental effects between teacher and parental ratings. At this young age, sex-limitation effects were found only in teacher ratings, where genetic effects were greater in girls than in boys. Conclusions: Results underscore the utility of using multiple informants to measure children's behavior and provide a foundation for follow-up of these twins in later adolescence.