The effects of intimate partnering and marital failure on criminal behavior: Social causation or self-selection?

摘要

The 1990s witnessed a protracted debate concerning the significance of adult life events for individual criminal careers. This dissertation considers the implications of intimate partnering and marital failure for self-reported involvement in crime. Data are derived from a sub-sample (n = 1,621) of the National Youth Survey (Elliott et al., 1989), a seven-wave, prospective panel study of American youth ages 11–27. The current study concludes that well-bonded marriages to pro-social spouses reduce offending while marital failure increases it, and that these relationships are attributable to distinct social causation processes. It therefore replicates past findings of the “good marriage effect,” and supports the notion that adult life events do influence even engrained criminal behavioral patterns.; The current study adds to the criminological understanding of the marriage-crime relationship in three important ways. First, rather than focusing on marriage as a dichotomy, the full spectrum of potential partnering statuses is considered. This not only provides a rare look at the effects of non-marital cohabitation and marital failure (i.e., bad marriage, separation, and divorce), but also raises serious methodological questions about the way “marriage” has been measured in the past. Second, the multivariate analyses conducted include a broader range of theoretically relevant moderators and mediators than previously considered. This rich assortment of controls isolates the independent effects of various partnering statuses from the corollary effects of childbearing, as well as changes in peer networks, personal economy and drug use patterns. Third, analyses are conducted in reference to participation in both violent and non-violent offending, as well as to the frequency of general offending among active offenders. Examining the effects of partnering status on distinct crime outcomes provides a sound basis for understanding when, how, and why partnering status affects criminal involvement. All multivariate models are estimated using fixed effects panel methods that eliminate sources of time-stable heterogeneity. Theoretical perspectives are drawn from the fields of life course sociology, the sociology of the family, and sociological criminology. Implications for both criminological theory and crime control policy are discussed.