An Examination of the "Marriage Effect" on Desistance from Crime among U.S. Immigrants

This project satisfies requirements as stated in the National Institute of Justice (NIJ) FY 2011 W.E.B. Du Bois Fellowship Program Solicitation. The primary aim of the proposed research is to address gaps in the literature by examining the relationship between marriage and offending among a racially/ethnically diverse sample of U.S. immigrants. Specifically the research asks whether the marriage relationship among immigrants is similar to the marriage effect documented in the life course literature. This research will also examine whether the marriage effect among immigrants is conditioned by generational and nativity status. This is the first longitudinal study of the influence of major life turning points - specifically marriage - on immigrant offending trajectories from adolescence through young adulthood. This research uniquely recognizes that the "immigrant" category is influenced by multiple layers of heterogeneity. Therefore, the grantee pays attention to understanding the influence of marriage on persistence and desistance from crime by disaggregating the immigrant sample by generational status and nativity group.
These questions will be examined using the National Longitudinal Survey of Youth (NLSY97) (1997-2008). It is a longitudinal data set interviewing a sample of 8,984 nationally representative youth living in the U.S. since 1997. There have been 12 waves of data collection for youth ages 12-28 years old. The survey oversampled on African Americans and Hispanics (N=2,000) and includes self reported measures of delinquency. Immigrant status could be determined for 7,918 of the sample with 6,418 being Native born, N=532 first generation and N=1946 second generation. Key data measures that will be used in this study include a count measure of the number of criminal involvements in each year disaggregated by crime type (i.e. violent crime, property crime, and drug crime). Other variables included in the analysis will consist of family and neighborhood structural variables and demographic correlates of offending including sex, race, intelligence and the early onset of delinquency. The independent variable of interest is marriage, which is a time-varying (measured at each wave at data collection) dichotomous (0,1 = married) variable. Approximately 37 percent (N=2,767) of the sample reported ever being married, and 30 percent (N=2,269) are currently married. The average age of first marriage is 22 years old.
Analysis will happen in two stages - first basic descriptive analyses to examine differential distributions of marriage across race and ethnicity will be conducted. Then two level hierarchical linear models will be used to analyze within individual and between individual changes over the course of the study. Group-mean centering will be used for the time-varying covariates (e.g. marriage). A Poisson extension of hierarchical linear modeling (HLM) will be used in the data analysis to allow for over dispersion in each model.
This research has implications for policy and practice. There is currently much empirical work underway in the area of understanding the nexus between immigration and crime. Legislation abounds in the last decade to address the "immigrant" problem in the U.S. This work has policy relevance because if policies (i.e. deportation) foster the breakdown of marriages and families, this research may speak to revisiting those policies.
ca/ncf