Income Inequality, Not Gun Policy or Mental Illness, is the Strongest State-Level Predictor of Mass Shootings in the United States: A Multi-Method Analysis (2018–2024)

Abstract Background: Mass shootings are a growing public health crisis in the United States, yet state-level predictors remain poorly understood. Prior research has identified income inequality as a risk factor, but no study has simultaneously examined socioeconomic, mental health, substance use, educational, gun policy, and demographic predictors using multiple analytical methods. This study identifies which state-level characteristics most consistently predict mass shooting rates and evaluates the roles of gun policy and mental illness prevalence. Methods: We compiled corrected Gun Violence Archive data on 3,852 mass shooting incidents (defined as 4 or more persons shot) across all 50 U.S. states and the District of Columbia from 2018 to 2024. We assembled 23 predictor variables from six domains (socioeconomic, mental health, substance use, education/cognitive, gun policy, and demographics) using authoritative sources including the U.S. Census, SAMHSA, FBI, KFF, and NCES. Five complementary analytical approaches were employed: bivariate Pearson correlations, multiple OLS regression, LASSO regularization, Random Forest, and Gradient Boosting. A consensus ranking was derived by averaging variable importance ranks across methods. Results: The Gini coefficient (income inequality) ranked as the most important predictor across all five methods (consensus rank 1.8/21). In the parsimonious OLS model (R² = 0.461, p < .001), the Gini coefficient was highly significant (b = 14.73, p = .007), while poverty rate became non-significant (p = .380) when both were included. Mental health provider density was significantly protective (b = -0.003, p = .030). Gun ownership rate (r = 0.044, p = .762) and gun law strength (r = -0.121, p = .397) showed no significant association. Mental illness prevalence was paradoxically negatively correlated with mass shooting rates (r = -0.307, p = .026). LASSO eliminated all gun policy variables from the model. Conclusions: Income inequality, not gun ownership or mental illness prevalence, is the strongest and most consistent state-level predictor of mass shooting frequency. The paradoxical negative association between mental illness prevalence and mass shootings likely reflects confounding with state-level diagnostic infrastructure. Expanding mental health provider access represents the most actionable policy lever. These findings suggest prevention strategies should prioritize structural determinants of health alongside traditional gun violence interventions.