Abstract / Summary
Background. ADHD and early anger problems are linked to later social difficulties, but whether childhood hyperactivity and temper predict adult romantic relationships beyond adult personality is less clear. Methods. Analyses in two US cohorts, each planned and written down before the data for that outcome were analysed. In the NLSY79 Children and Young Adults (CYA; 11,545 children of 4,941 mothers), mother-rated hyperactivity and temper at ages 4–14 were related to adult relationship unhappiness, separation or divorce and days of lost temper, with and without adjustment for adult neuroticism (and, for hyperactivity, disagreeableness). Infant temperament was related to later hyperactivity. In Add Health (N = 4,922), a self-reported lifetime ADHD diagnosis was related to relationship quality and partner violence reported at the same interview (Wave IV, 2008–09) and to marital dissolution by Wave V (2016–18), with and without adjustment for neuroticism and trait anger. Results. Each SD of childhood hyperactivity predicted +5.9 percentage points in later separation or divorce (95% CI 4.3 to 7.5; 34.7% base rate among the ever-married), almost unchanged after adjustment for adult neuroticism and disagreeableness (+5.8). Childhood temper predicted separation or divorce (+5.9 points; +3.7 after adjusting for hyperactivity and adult neuroticism) and more days of lost temper in adulthood (+0.17 SD). Infant difficulty and activity predicted later hyperactivity. An ADHD diagnosis was associated with lower relationship quality and more violence toward and from the partner at the same interview, and predicted +13.1 points in later dissolution; estimates shrank by 12% to 34% after adjusting for personality and stayed in the same direction, although relationship quality and dissolution were then no longer significant after correction. In checks added later, the CYA results held without covariates and, at about half the size, with a single rating per child; in Add Health only violence from the partner held both without covariates and with survey weights, and the dissolution estimate (64 diagnosed respondents) was +7.8 points (−5.2 to 20.7) with weights. Relative age did not predict diagnosis. Conclusions. In these two US cohorts, childhood hyperactivity and temper rated by mothers were early markers of adult relationship risk that brief adult personality measures (neuroticism, and for hyperactivity also disagreeableness) captured only in part. For a self-reported ADHD diagnosis the evidence was firm only for violence from the partner. The childhood results are consistent with giving attention and activity problems their own place in models of couple risk. Analysis scripts and aggregate output: https://doi.org/10.17605/OSF.IO/4X7RF. Part of the 2WO project. Individual-level data cannot be redistributed; they are available from the data providers.