Although more exploratory in nature, the analyses examining the effects of AO and delay on changes in drinking and problems across the college years also yielded interesting results. Individuals with an earlier AO and shorter delay remained at higher risk for heavy drinking and problems by senior year of college, but those with a later AO or a longer delay appeared to start to “catch up” over time. Although these results are consistent with a recent study examining the risks associated with an early AI (Henry et al., 2011), they may also reflect a statistical artifact, whereby individuals with an earlier AOor shorter delay had more room toincrease their drinking over time. Consistent with this possibility, a longer delay predicted greater increases in heavy drinking and problems for participants with an earlier AO (by age 16 years) but not for those with a later AO (after age 16 years), more than likely as a result of the restricted range for this variable among those with a later AO. Future longitudinal studies assessing relations between AO, delay, and alcohol outcomes over
