Jui-Chung Allen Li
Lawrence L. Wu
This paper resolves apparent discrepancies concerning trends in the intergenerational transmission of divorce. Wolfinger (1999) reports that divorce transmission has weakened substantially, due perhaps to declining stigma or selectivity among those who divorce. While plausible, his findings contrast sharply with others (McLanahan and Bumpass 1988; Teachman 2002), who report no trend in divorce transmission. Our analyses of data from the National Survey of Families and Households , using life table methods and a Cox model stratified by marriage cohort, provide no support for a decline in divorce transmission. We note that the General Social Survey data used by Wolfinger lack information on marital duration, permitting analysis only for whether respondents have divorced by interview. Apparent declines in divorce transmission thus could be an artifact of longer exposures to risk, and hence higher probabilities of divorce by interview, for earlier marriage cohorts relative to later cohorts. Our analyses of the GSS and NSFH reveal that this artifact is indeed present. We conclude that there has been no trend in divorce transmission in recent decades in the United States.