Personality, social structure, and suicide mortality rates: a comparative dada mining approach

Suicide is a growing problem in the United States, yet research often emphasizes either individual psychological factors or environmental influences, typically favoring the former, with limited studies integrating both. Given the complex and multifaceted nature of suicidal behavior, a more comprehensive approach may yield greater insights. This study compares three state-level datasets using Random Forest models. The first dataset, from 2008, includes state-level estimates of the Five Factor Model’s personality traits and various sociodemographic variables. The second dataset excludes personality data to assess the impact of its absence, while the third replicates the second dataset but with updated predictors for 2022. Several predictors, such as divorce rates, gun ownership, permit laws, per capita income, and educational attainment, were consistently important predictors across all models. Additionally, the inclusion of personality data explained the most variance in suicide mortality rates. Notably, predictor importance shifted from 2008 to 2022, emphasizing the dynamic nature of factors associated with suicide mortality over time.