This page was translated automatically from Turkish. Read the original

This page is an AI-assisted Turkish summary of the study presented at the 26th World Congress of Psychiatry. The study belongs to the presenting authors; please refer to the original source when citing it.

EV0526

Examining emotional intelligence among health sciences students at Kuwait University: a cross-sectional study

EXPLORING EMOTIONAL INTELLIGENCE AMONG ALLIED HEALTH SCIENCES STUDENTS AT KUWAIT UNIVERSITY: A CROSS-SECTIONAL STUDY

Ability-model-based self-report and trait emotional intelligence measures diverge in a subset of students, revealing a stable discordant profile that remains undetected in single-measure designs.

The study aimed to investigate whether ability-model-based self-reported emotional intelligence (SSEIT) and trait emotional intelligence (TEIQue-SF) diverge within individuals among students at Kuwait University’s Faculty of Health Sciences, and under what conditions this question can be answered reliably. Of the 561 responses collected, 559 remained after duplicate removal; after excluding 70 respondents (12.5%) who provided invariant responses, 489 participants were included in the analysis. The correlation between the SSEIT (α = 0.96) and the Kuwaiti Arabic TEIQue-SF (α = 0.83) was r = 0.35. Latent profile analysis was performed on standardized scores with a variance floor imposed, and model selection was based on BIC; outcome variables were examined using BCH adjustment, bootstrapping with 2,000 resamples (1,471 valid), and Holm/Bonferroni corrections for multiple testing. BIC selected four profiles (entropy 0.65; average posterior probability (AvePP) 0.70 to 0.92): High emotional intelligence (n = 140, 28.6%), Average (n = 167, 34.2%), Low (n = 158, 32.3%), and Discordant (n = 24, 4.9%); aBIC and AIC continued to decrease up to k = 6. The Discordant profile combined very low self-perceived ability (-1.82 z) with nearly average trait emotional intelligence (-0.09 z), and its mean TEIQue-SF score (4.30) exceeded those of the Average (4.26) and Low (4.02) profiles; this profile was recovered in 49 of 50 multiple starts and in 92% of valid bootstrap samples. Among the outcome variables, workshop participation was concentrated in the Low profile (+24.9 percentage points for Low versus High; +20.5 percentage points for Low versus Average), whereas confidence in managing stress was highest in the Discordant profile (+0.90 for Discordant versus Average; +0.48 for High versus Average); these contrasts remained significant after Holm correction. No association with grade point average (GPA) was detected (H = 6.68; p = 0.083), and workshop participation was independent of department and year of study (p = 0.107 for both). The authors emphasize that emotional intelligence self-report measures from different traditions can diverge sufficiently to identify distinct subgroups. They note that unconstrained latent profile analysis in the unscreened sample produced a degenerate class consisting of exactly 74 respondents with TEIQue-SF = 4.00, that correcting this eliminated the GPA finding that appeared significant in the naive analysis, and that screening, variance floors, BCH estimation, and multiplicity control should become standard practice. Because the design was cross-sectional, the direction of causality could not be determined.

All WCP 2026 summaries