Prospective Relations between Parents’ Depressive Symptoms and Children’s Attributional Style

Abstract

Children of parents with depression are at increased risk for developing psychopathology. The purpose of the current longitudinal study was to examine the dynamic relations between parents’ depressive symptoms and children’s cognitions, specifically their attributions for the causes of life events. Participants were 227 parent–child dyads with one parent (M_age = 42.19, SD = 6.82; 76% female) and one child (M_age = 12.53, SD = 2.33; 53% female) per family. Parents either were diagnosed with a current major depressive disorder (n= 129; 72.9% female) or were lifetime-free of mood disorders (n= 98; 79.6% female). The Beck Depression Inventory-II was used to obtain a dimensional measure of parents’ depressive symptoms, and the Children’s Attributional Style Questionnaire–Revised was used to assess children’s attributions of negative and positive events. Evaluations were conducted 5 times across 22 months. We used latent difference score (LDS) modeling to examine the relations between changes in parents’ depressive symptoms and changes in children’s attributional style over time. The final model provided a close fit to the data: χ²(30) = 35.22, p = .24; comparative fit index = .995, root mean square error of approximation = .028, 90% confidence interval (CI) [.000, .060], standardized root mean square residual = .024. Parents’ levels of depressive symptoms significantly predicted the worsening of children’s attributions (i.e., becoming more pessimistic) over the 22 months, whereas children’s attributions did not significantly predict changes in parents’ depressive symptoms at the next time point. Preventive interventions should aim to both reduce parents’ depression and teach children strategies for examining the accuracy of their beliefs regarding the causes of life events.

Acknowledgments

We thank all those who contributed to completion of this project, including Steven Hollon, Ph.D.; Robert DeRubeis, Ph.D.; Richard Shelton, Ph.D.; Jay Amsterdam, Ph.D.; Sona Dimidjian, Ph.D.; Margaret Lovett, M.S.; Cynthia Flynn, Ph.D.; Russell Hanford, Ph.D.; Virginia Burks, Ph.D.; and Tory Creed, Ph.D.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Institute of Mental Health grants (R01MH57822, R01MH57834, R01MH057977) and training grant (T32 MH018921).