Profiles of self-reported emotional intelligence and cognitive emotion regulation strategies among university students: a latent profile analysis

Self-reported emotional intelligence (EI) is associated with psychological adjustment, but most studies of its links with emotion regulation (ER) are variable-centred. Students could differ in the configuration of EI dimensions (e.g., high emotion appraisal with low perceived regulation) or mainly in overall level. Using latent profile analysis (LPA), this study tested whether Wong and Law Emotional Intelligence Scale (WLEIS) profiles among Saudi university students differed by shape or by level, and how the profiles related to cognitive ER strategies. In this cross-sectional study, 412 undergraduates at King Khalid University, Saudi Arabia, completed the WLEIS and the Cognitive Emotion Regulation Questionnaire-short form (CERQ-short) during the 2025–2026 academic year; 38 of 450 responses with missing questionnaire or demographic data were excluded without imputation. LPA used the four WLEIS dimensions. Model selection considered information criteria, entropy, likelihood-ratio tests, classification accuracy, and parsimony. The BCH method compared cognitive ER strategies across profiles while accounting for classification uncertainty, with Bonferroni correction across nine omnibus tests. Approximate η² and descriptive reconstructed pairwise differences, 95% CIs, and standardized mean differences (d) were used to characterise magnitude. A three-profile solution was selected: lower ( n = 91, 22.1%), intermediate ( n = 221, 53.6%), and higher ( n = 100, 24.3%) WLEIS-score profiles (entropy = 0.91; average posterior probabilities = 0.93-0.96). Profile means were ordered consistently on all four dimensions, indicating level rather than shape differentiation. After correction, differences remained for catastrophising (η² ≈ 0.13), positive reappraisal (0.12), refocus on planning (0.09), putting into perspective (0.09), self-blame (0.08), positive refocusing (0.07), and rumination (0.06; all adjusted p < .001), but not for acceptance or blaming others (adjusted p = 1.000). Higher-versus-lower profile differences ranged from |d| ≈ 0.76 (rumination) to 1.13 (catastrophising). In this sample, heterogeneity in self-reported EI was mainly quantitative: profiles represented levels of largely co-varying WLEIS dimensions rather than distinct EI types and showed graded differences in cognitive ER. No configural EI subtypes emerged, suggesting that continuous WLEIS scores may suffice for many purposes. Cross-sectional self-report data cannot establish causality or the effectiveness of profile-based support.