Inclusion of a measure of event dependence in our study afforded better traction on understanding the causal relations between alcohol consumption and life event occurrence. We found that events rated as independent of respondents’ behavior were associated with greater alcohol consumption among early drinking men. This makes it unlikely that the causal direction was for heavier drinking to create greater stress exposure. Rather, our findings are consistent with the notion that early drinking men are more likely to drink in response to independent life events compared with men with later drinking onset. However, we are unable to address the direction of causality between dependent life events and past-year drinking because our study design did not provide closely timed data on event occurrence and consumption. It is possible that early initiators have drinking patterns in adulthood that more often contribute to occurrence of dependent events. There is some support for this in our sample, as evidenced by the higher levels of dependent events among early initiators (see Table 1).
