Cause or consequence?: Exploring causal links between attachment and depression

The association between women's attachment representations and episodes of depressed mood during pregnancy and early postpartum months was explored using statistical techniques to control for measurement error that may overestimate associations between the constructs. Two hundred and thirty-five women completed attachment and depression questionnaires prenatally and when their infants were 6 months old. They also completed assessments of depression four times in the early postpartum months. The results were consistent with expectations. Prenatal attachment scores were associated with both prenatal and several postnatal assessments of depressed mood, and correlations did not differ over time despite a significant decrease in depressed mood from pregnancy to 6 months postpartum. Next, Structural Equation Model was used to simultaneously predict attachment and depression over time and control for reliability and measurement errors. Stability of both attachment and depression was consistent with stability coefficients reported in several previous studies. Furthermore, there was support for the proposal that internal working models of attachment are both a cause (self-model/anxiety) and a consequence (other-model/avoidance) of depression. The self-model of attachment was negatively associated with depression over time, and depression was negatively associated with the other-model over time.