Introduction: Insecure attachment is a risk factor and predictor for mental disorders and suicide. The connection between insecure attachment and resilience-a protective factor which contributes to mental health-is of particular interest, although pathways are poorly understood. This study presents a comprehensive, integrative framework, investigating the mediating effects of social support (external resource) and self-efficacy (internal resource) on the relationship between insecure attachment and resilience.
Methods: The sample comprised N = 339 healthy adults. Insecure attachment was analyzed using a refined version of existing dimensional approaches, assessing attachment anxiety and avoidance. Also, resilience, perceived social support, and self-efficacy were measured. Two distinct serial multiple mediator models were specified and tested to examine the differential direct and indirect pathways associated with anxiety and avoidance.
Results: Significant indirect effects via social support, self-efficacy, and their sequential combination were observed for both attachment anxiety and avoidance (all p < 0.05). The association between anxiety and resilience was partially mediated (direct effect: p < 0.001), whereas the association between avoidance and resilience was fully mediated (direct effect: p = 0.72).
Discussion: The results highlight differences in the underlying mechanisms of the dimensions: While the psychopathology of avoidance is straightforward, considering anxiety additional factors seem crucial. Social support and self-efficacy effectively buffer negative consequences of burden. The findings suggest a shift in focus within intervention research-from hard-to-modify attachment to the more accessible targets self-efficacy and social support.
Keywords: insecure attachment; mental well-being; protective factors; resilience; serial mediation analysis.
Copyright © 2026 Schäfer, Schöniger, Brunner, König, Zwiky, Küttner, Hamann, Scheiwe, Andreev, Ptasczynski, Fleischer, Korna, Wittig, Kamrla, Selle and Redlich.