While prior literature has largely focused on marriage effects during young adulthood, it is less clear whether these effects are as strong in middle adulthood. Thus, we investigated age differences in marriage effects on problem-drinking reduction. We employed parallel analyses with two independent samples (analytic-sample Ns of 577 and 441, respectively). Both are high-risk samples by design, with about 50% of participants having a parent with lifetime alcohol use disorder. Both samples have been assessed longitudinally from early young adulthood to the mid-to-late 30s. Separate parallel analyses with these two samples allowed evaluation of the reproducibility of results. Growth models of problem drinking tested marriage as a time-varying predictor and thereby assessed age differences in marriage effects. For both samples, results consistently showed marriage effects to be strongest in early young adulthood and to decrease somewhat monotonically thereafter with age, reaching very small (and nonsignificant) magnitudes by the 30s. Results may reflect that role transitions like marriage have more impact on problem drinking in earlier versus later adulthood, thereby highlighting the importance of life span developmental research for understanding problem-drinking desistance. Our findings can inform intervention strategies aimed at reducing problem drinking by jumpstarting or amplifying natural processes of adult role adaptation.