The paradox of social capital and gendered suicide ideation: Findings from correlated random effects (CRE) models

Abstract

Objective The purpose of this study is twofold. First, it examines how social capital accessible through interpersonal contacts (e.g., family, friends, neighbors, coworkers) is related to suicide ideation during the COVID-19 pandemic. Second, it investigates whether this relationship varies across gender groups. Methods Two waves of longitudinal data (n = 8953 person-years) are drawn from the Seoul Young Adult Panel Survey fielded during the height of the pandemic (2021−2022). For estimation strategy, the correlated random effects (CRE) analysis is used, along with logit and probit regression as robustness checks. Results Adjusting for individual covariates, unobserved time-invariant heterogeneity, and regional fixed effects, results from the CRE models show a significantly positive association between social capital and suicidality (ß = 0 .004, p < .001). Specifically, greater access to network-mediated social support increases the probability of suicide intent over time. Importantly, this effect is significant only among women (β = 0 .006, p < .001). These results are consistently significant when using logit and probit models with Bonferroni-adjusted p -values. Conclusions In contrast to conventional view in the scholarship that social capital is protective of mental health risks, this study highlights its potential downside. Future research should recognize that interpersonal ties do not necessarily promote emotional or psychological well-being, especially for women. As such, the academic community should pay more detailed attention to how and the conditions under which social capital can produce gendered benefits as well as impose differential costs for males and females. Copyright © 2026. Published by Elsevier B.V.