Causal Inference Methods for Intergenerational Research Using Observational Data

Abstract

Identifying early causal factors leading to the development of poor mental health and behavioral outcomes is essential to design efficient preventive interventions. The substantial associations observed between parental risk factors (e.g., maternal stress in pregnancy, parental education, parental psychopathology, parent–child relationship) and child outcomes point toward the importance of parents in shaping child outcomes. However, such associations may also reflect confounding, including genetic transmission—that is, the child inherits genetic risk common to the parental risk factor and the child outcome. This can generate associations in the absence of a causal effect. As randomized trials and experiments are often not feasible or ethical, observational studies can help to infer causality under specific assumptions. This review aims to provide a comprehensive summary of current causal inference methods using observational data in intergenerational settings. We present the rich causal inference toolbox currently available to researchers, including genetically informed and analytical methods, and discuss their application to child mental health and related outcomes. We outline promising research areas and discuss how existing approaches can be combined or extended to probe the causal nature of intergenerational effects.