Gender differences and correlations between learning perceptions and communication performance in ChatGPT-simulated OSCE nurse–patient communication

Abstract

This exploratory study examined gender differences in nursing students’ learning perceptions and communication performance during ChatGPT-simulated OSCE communication. Twenty-eight students (12 male, 16 female) participated, completing OSCE evaluations and perception questionnaires. Mann–Whitney U tests showed that male students had significantly higher perception scores than female students (p < 0.05), but Spearman’s rho found no correlation between perceptions and performance. Z-scores revealed that male students tended to overestimate their performance, whereas female students’ learning perceptions were more aligned with their actual scores. Additionally, the number of dialogue turns was positively correlated with communication scores, though neither variable showed significant gender differences. Despite the small sample size, these findings suggest that gender may influence perceptions in AI-assisted learning environments, supporting the inclusion of gender and engagement metrics in AI educational design.