The second sample, the divorce analytic sample, was a subset of the marriage analytic sample, and it excluded those married less than 3 years (and not divorced) at their assessment (or most recent assessment, for those with multiple assessments). We implemented this cut-off to ensure that participants had passed through the period of highest risk for divorce before being classified as not divorced, which is typically in the early years of marriage (Cherlin, 1992; Mayol-Garcia et al., 2021). In total, the divorce analytic sample included 3360 EA participants (from 925 extended families; 1938 (58%) female; Mage = 42.30 years, age range = 21–84) and 920 AA participants (from 392 extended families; 532 (58%) female; Mage = 42.94 years, age range = 21–79).
