Mentoring is a critical factor in fostering persistence and professional identity among undergraduate students from underrepresented minority (URM) backgrounds in STEM. However, the systematic assessment of mentor effectiveness, particularly within diverse institutional contexts, remains a challenge. The Mentoring Competency Assessment (MCA), though widely used, was originally validated on a homogeneous sample and subsequent revisions (MCA-21) relied on traditional psychometric methods (PCA/CFA), leaving concerns about its applicability and complexity in highly diverse settings. This study addressed the need for a culturally relevant and psychometrically rigorous assessment by examining the factor structure of the MCA-21 among research mentors ( N = 323) at a Hispanic-Serving Institution (HSI). We utilized Exploratory Graph Analysis (EGA), a network-based method that identifies latent structures without imposing rigid model assumptions, to assess how mentoring competencies cluster in this unique context. EGA results revealed a more parsimonious three-factor structure, significantly differing from the original six-factor model. This abbreviated structure suggests that effective mentorship at an HSI may be best captured by clusters focusing on (1) Research Skill Development, (2) Identity and Belonging, and (3) Career Alignment. These findings indicate that the structure of mentoring competencies is context-dependent, potentially integrating culturally responsive practices into core domains. The resulting three-factor structure provides a validated, context-specific, and less burdensome instrument for HSI settings (CLI = 0.951). This refinement enhances the utility of the MCA for evaluating mentor effectiveness, improving the precision of mentor training programs, and ultimately fostering equitable professional development for URM students in diverse academic environments.
Aguilera et al. (Thu,) studied this question.