Linear and nonlinear models of vulnerability to depression: Personality and postpartum depression in a high risk population

Although nonlinear effects have sometimes been detected in depression research, they have been seldom considered at the empirical and conceptual levels. The current study explored the nonlinear relations between Self-Criticism and Dependency personality vulnerabilities to depressive symptoms (Blatt, 1974; 1991) in both low and high stressful life conditions. A prospective investigation of post-deliveiy depressive symptoms in high- and low-risk pregnancy (HRP and LRP) samples was conducted. It was hypothesized that HRP (i.e., gestational diabetes mellitus), as a highly stressful life context, would interact with personality vulnerabilities to depression, resulting in nonlinear threshold effects on outcomes. Hierarchical Multiple Regression analyses (HMR) indicate nonlinear relationships between depression and both Self-Criticism and Dependency in the HRP group (N = 100). In contrast, the relationship among these variables remained linear in the LRP group (N = 109). Multiple groups Structural Equation Model (SEM) indicated significant stress × personality interaction; for the high risk group the relations between the personality variables and depression are cubic but only linear for the low risk group and these differences are significant. Overall, the results are consistent with the notion of stress thresholds and the findings underscore the general need to examine non-linear effects in research on personality and depression. In addition, the findings support the assumed importance of the balance between the personality dimensions of relatedness and self-definition.