Evaluation of a six-item second-order academic engagement model in Ecuadorian university students

Introduction Academic engagement is a key determinant of students’ wellbeing and academic success. Although the Utrecht Work Engagement Scale for Students (UWES-S-9) is one of the most widely used instruments to assess this construct, previous studies have reported inconsistent evidence regarding its factorial structure, particularly in Latin American populations. This study examined the psychometric properties of the UWES-S-9 and evaluated whether a more parsimonious higher-order model could better represent academic engagement among Ecuadorian university students. Methods A cross-sectional study was conducted with 1,732 university students from three Ecuadorian universities. The sample was randomly divided for exploratory and confirmatory analyses. Exploratory factor analysis (EFA), confirmatory factor analysis (CFA), measurement invariance across sex, age, and university type, internal consistency, convergent validity, and nomological validity through associations with external psychological constructs were assessed. Results Among eight competing models, the six-item three-factor model and its second-order counterpart showed the most favorable relative fit indices (CFI = 0.992, TLI = 0.980, SRMR = 0.022), although the RMSEA remained elevated (RMSEA = 0.105), indicating mixed evidence of overall model fit. The scale showed excellent reliability ( ω = 0.905; CR = 0.958), adequate convergent validity (AVE > 0.68 for all dimensions), and full configural, metric, scalar, and strict invariance across sex, age, and university type. Associations with psychological capital, academic satisfaction, perceived stress, psychological inflexibility, and symptoms of anxiety and depression occurred in the theoretically expected directions, providing evidence of nomological validity. Conclusion The refined six-item second-order UWES-S provides a brief and reliable representation of academic engagement in Ecuadorian university students. Its hierarchical structure offers a theoretically coherent and parsimonious representation of the construct, with favorable incremental and residual fit indices, although the elevated RMSEA warrants caution in interpreting overall model fit. These findings support the six-item model as a promising candidate for research and educational assessment; however, independent replication is needed before it can be considered an established alternative to the UWES-S-9.