Furthermore, a subset of the twins, from which the current study sample comes, was more intensively studied. This subsample of 1035 families was mostly comprised of randomly selected twins across all study years, while about a third were selected based on parental self-reports of elevated risk for alcohol problems. Previous investigation has shown that this modest enrichment of at-risk families did not systematically affect substance use or behavioral problems among the twins through adolescence [27]. Additional assessments of this more intensively studied subsample included one-on-one psychiatric interviews at ages 14 (mean = 14.2, SD = 0.15; n = 1852, response rate 89.5%) and 22 (mean = 22.4, SD = 0.70; n = 1347, response rate 73.0%). At the time of the age 14 interview, individuals completed an additional questionnaire which included self and co-twin assessments of emotional and behavioral problems.
