Do more stress and lower family economic status increase vulnerability to suicidal ideation? Evidence of a U-shaped relationship in a large cross-sectional sample of South Korean adolescents

Abstract

It is widely held in socio-behavioral studies of suicide that higher levels of stress and lower levels of economic status amplify suicidal vulnerability when confronted with a proximal stressor, reflecting the traditionally prevalent understanding in health psychology and sociology that associates adverse life circumstances with undesirable mental health outcomes. However, upon reflection, there are strong theoretical reasons to doubt that having more stress or being in a more stressful environment always increases suicidal vulnerability given the occurrence of a crisis. Using large nationally representative public survey data on South Korean adolescents, I show that the association between recent psychosocial crisis and suicidal ideation often gets stronger with more favorable levels of perceived stress and improving levels of family economic status. Overall, the increase in the probability of suicidal ideation from recent exposure to a psychosocial crisis is consistently the smallest around medium levels of stress or family economic status and larger at low or high levels. A supplementary exercise suggests that the identified moderation effects operate mainly in virtue of individual-level stress or family economic status in the relative absence of contextual influences at the school level. The findings present preliminary evidence of the stress inoculation hypothesis with regard to suicidal ideation. Research on suicidal vulnerability could benefit from increased attentiveness to the mechanisms through which being in an adverse or unfavorable life situation could protect against the suicide-inducing effects of proximal stressors.

Fig 1. Distribution of key variables.

Fig 2. Predicted probabilities for each category of stress and economic status.

All variables not involved in the interaction were set to zero or the reference category. Error bars indicate 90% confidence intervals.

Fig 3. Relative risk for each category of stress and economic status.

All covariates not involved in the interaction were set to zero or the reference category. Standard errors were estimated with the delta method. The error bars indicate 90% confidence intervals.

Fig 4. Risk difference for each category of stress and economic status.

All covariates not involved in the interaction were set to zero or the reference category. Standard errors were estimated with the delta method. Error bars indicate 90% confidence intervals.

Fig 5. Predicted probability by each category of stress and economic status, multilevel models.

All covariates not involved in the interaction terms were set to zero or the reference category. Random effects were set to zero. Error bars indicate 90% credible intervals.