Giovanni Lascano-Arias

Background. Relational self-esteem—self-worth derived from significant relationships—is central to psychosocial functioning. However, brief measures validated with appropriate methods for ordinal items are scarce in Ecuadorian populations.

Objective. To examine the factorial structure, internal consistency, convergent validity, and gender-based measurement invariance of the Relational Self-Esteem Scale (RSE), using a cross-validation approach.

Design. Instrumental, cross-sectional study with n = 1.478 Ecuadorian adults. The sample was split for cross-validation: 40% (n = 591) for EFA and 60% (n = 887) for CFA. EFA used a polychoric matrix, ULS extraction, and oblimin rotation; factor retention was guided by parallel analysis. CFA employed WLSMV. Convergent validity was tested against personal self-esteem (RSES), collective self-esteem (CSE), and mental well-being (MHC-SF). Multigroup measurement invariance by gender was assessed (configural, metric, scalar, strict).

Results. Parallel analysis and confirmatory factor analysis supported a unidimensional structure with robust factor loadings and an adequate overall model fit. The scale demonstrated high internal consistency and reliability. Convergent validity was established through significant positive correlations with measures of self-esteem, core self-evaluations, and mental health. The instrument showed full measurement invariance across genders, meeting all criteria from configural to strict levels.

Conclusion. The RSE shows a unidimensional structure, adequate reliability, convergent validity, and gender invariance in Ecuadorian adults. Findings support using the total score as a brief, valid indicator of relational self-esteem in this context.