Why do individuals select romantic partners who use drugs, are criminals, or have mental health problems, a choice that eventually puts them and their children at increased risk for negative developmental outcomes? Results are presented from a systematic literature review on assortative mating for antisocial behavior and on the subsequent influence partners have on each other. All cross-sectional, retrospective studies except one supported assortative mating over partner influence. In contrast, all prospective studies supported partner influence. Given that prospective data are generally better than retrospective data, partner influence is recognized here as an important finding, previously hidden or discounted in the literature because of its reliance on retrospective, cross-sectional designs. Theoretical perspectives, social homogamy, heterotypic assortment, and methodological issues are also examined.