Our study benefited from a prospective design that minimized the possibility that adolescent drinking contributed to likelihood of marital disruption, given that most participants (92.7%) had not consumed a full drink of alcohol prior to the experience of divorce/separation. Our prospective follow-up of young sample also reduced likelihood of retrospective recall bias, especially the forward telescoping bias, the tendency to report events to have occurred closer to the assessment than is true (Johnson & Schultz, 2005). We were also able to obtain a measure of current parental drinking during adolescents at a point proximal to the time of drinking initiation, which is an improvement on retrospective reports of parental drinking. Our parental reports provided a measure of child life events (including divorce/separation) that was at least partially independent of adolescents’ self-report on alcohol use. Parent reports may not adequately identify events that are salient to the child, however. In addition, events were reported retrospectively, as was the age experienced. Reports of experiences that do not rely heavily on judgment and interpretation tend to have greater validity in general; however, recall of the timing of adverse experiences may be biased (Hardt & Rutter, 2004).
