Social support and posttraumatic growth : A meta-analysis

Background: The beneficial role of social support on posttraumatic growth (PTG) has been assumed by theoretical models and established in some studies. However, there are inconsistent findings and little knowledge on moderators. The present study aims to investigate the overall effect size of the relationship and identify factors affecting the association.

Methods: Six electronic databases were searched. Newcastle-Ottawa Quality Assessment Scale (NOS) were used to evaluate the quality of studies. Study quality, study design, trauma type, PTG measure, social support measure, continent, publishing language, sample size, gender, religion, and age were analyzed as moderators. Meta-regression was conducted with the significant differential predictors in moderator analysis.

Results: The meta-analysis included 217 samples and a total of 47,940 participants from both longitudinal and cross-sectional studies. There was a medium positive effect size between social support and PTG in random effect model, r = 0.418, p < .001. The meta-regression analysis indicated that the association between social support and PTG was stronger among caregivers (vs. other traumatized samples), Chinese, older individuals and studies with smaller sample size.

Limitations: Only survey results were included in the analysis. The retrospective self-report may limit a more objective assessment of the relations. In addition, 87 % of the studies were cross-sectional, which may influence the estimation of a valid effect size.

Conclusions: Regarding the medium positive association between social support and PTG, it is important to enhance social support for trauma survivors. It will be especially effective for caregivers, Chinese, and older people.

Highlights
•  The overall effect size of the relationship between PTG and social support requires examination via empirical evidence.

•  The meta-analysis indicates a medium positive effect size between social support and PTG in the trauma population, r = .418.

•  The association was affected by study design, publishing language, trauma type, and the age of the individuals.