Research on externalizing psychopathology has relied heavily on White samples to generate core knowledge, with few studies examining variability in its components, including grandiose narcissism, across racial/ethnic groups. This preregistered (https://osf.io/n4s3f/) study addressed the following research questions: (1) Is there evidence for measurement invariance of the Narcissistic Admiration and Rivalry Questionnaire (NARQ) across racial/ethnic groups?; (2) Are there racial/ethnic group differences in (a) mean levels of the two NARQ subscales: admiration and rivalry, and (b) correlations between NARQ subscales and self-esteem?; (3) Do variations in ethnic identity commitment account for any observed group differences in the mean levels and correlations? The sample consisted of 1,248 U.S. adults between ages 18 and 30, evenly divided among those self-identifying as Black, Latine, and White. Both Confirmatory Factor Analysis and Exploratory Structural Equation Modeling indicated the intended two-factor structure across groups demonstrated configural, metric, scalar, and residual invariance. Group mean differences were most pronounced for the admiration subscale, with Black adults scoring higher than White (d = 0.63) and Latine adults (d = 0.46-0.47). Black-White mean differences in admiration were reduced by half when accounting for group differences in ethnic identity commitment. As admiration captures assertive self-promotion, participants identifying strongly as Black may endorse greater agency as an adaptive response to marginalization. Black-Latine mean differences were unaffected by group differences in ethnic identity commitment. This article highlights the importance of, and provides a framework for, psychometric examinations before interpreting group mean differences. (PsycInfo Database Record (c) 2025 APA, all rights reserved).