Emotion-specific regulation components differentially predict profiles of adolescent psychosocial dysfunction

= 16.06) completed measures of emotion-specific ER across sadness, fear, and anger (CERSA), psychosocial adjustment (SDQ), pathological traits (PID-5-BF), and disability (WHODAS-12). Person-centered cluster analysis identified dysfunction profiles, and ordinal LASSO penalized regression selected stable ER predictors of cluster membership, with both steps embedded within a full-pipeline bootstrap. Three stable clusters emerged, reflecting a severity gradient: low-, moderate-, and high-dysfunction. This solution was independently replicated via latent profile analysis sensitivity analysis (84.4% agreement, weighted κ = 0.86). LASSO regression with full-pipeline bootstrap inference retained eight stable predictors from 21 candidates, with seven threshold-specific effects supported by unconditional 95% confidence intervals excluding zero. General dysregulation across all three emotions was the strongest maladaptive predictor. In addition to general deficits, specific strategies showed context-dependent effects: Rumination-Anger was associated with greater dysfunction, whereas Reappraisal-Anger and Support Seeking-Sadness predicted lower dysfunction severity. Some associations were specific to particular severity transitions (low to moderate or moderate to high dysfunction) rather than uniform across the dysfunction spectrum. Our findings indicate that general dysregulation may represent transdiagnostic vulnerability, whereas specific strategies show emotion-context dependency, supporting functionalist theories that adaptiveness depends on matching regulatory strategies to an emotion's function.