Harsh parental discipline and offspring emotional and behavioural problems: Comparing findings from the random-intercept cross-lagged panel model and the monozygotic twin difference design

Objective. To elucidate associations between parental harsh discipline and child emotional and behavioural problems in monozygotic twins aged 9, 12 and 16 and to compare distinct approaches to causal inference.Method. Child reports of 5,698 identical twins from the Twins Early Development Study (TEDS) were analysed. We tested three types of longitudinal structural equation models: a cross-lagged panel model (CLPM), a random intercept CLPM (RI-CLPM) and a monozygotic twin difference version of the CLPM (MZD-CLPM). Results. Given the study aim to infer causation, interpretation of models focussed primarily on the magnitude and significance of cross-lagged associations. Behavioural problems resulted in harsher parental discipline across all models. In the CLPM, we found bidirectional effects between parental discipline behavioural problems at age 9 and 12. Point estimates of all other associations between parental harsh discipline and child emotional and behavioural problems were in the same direction but magnitude varied across models. In the MZD-CLPM, twin differences in harsh parental discipline at 9 predicted twin differences in emotional problems at 12. In the RI-CLPM, emotional problems at 12 predicted a reduction in harsh parental discipline at 16 within person. Conclusions. Findings can be interpreted as corroborating (but not definite) evidence in favour of a causal effect of child behavioural problems on later experienced harsh parental discipline. Yet, in light of the triangulated methods, results also illustrate divergence in the MZD-CLPM and RI-CLPM outcomes, and underline the importance of a well-defined research question, careful model selection and refining causal conclusions on within-person processes.