The purpose of this study was to examine the validity of the grades 6-8 version of DISCOVER, a performance-based assessment, and investigate its effectiveness in identifying gifted minority students. Questions examined the alignment between DISCOVER and Gardner's (1983) theory of multiple intelligences (MI) and assessed gender and ethnic differences. The sample consisted of 395 middle school students, predominantly Mexican Americans and Native Americans belonging to lower socioeconomic classes, from schools in Arizona. Results supported a good fit between DISCOVER and MI theory. That is, students identified as gifted in one intelligence were not necessarily identified in other intelligences. The 2 x 3 MANOVA (gender by ethnicity) showed no significant interaction or main effect for ethnicity. However, a main effect for gender was found with males outperforming females in math. No overall gender or ethnic differences in identification were revealed. In total, 12.4% of the participants were identified, suggesting that using DISCOVER might diminish the long-standing problem of minority underrepresentation in gifted programs.