Multiple criteria systems are recommended as best practice to identify culturally, linguistically, economically diverse students for gifted services, in which schools often incorporate measures of creativity. However, the role of creativity in identification systems and its recruitment of diverse student populations is unclear. The Torrance Test of Creative Thinking (TTCT) is the most widely used norm-referenced creativity test in gifted identification. Although commonly used for identifying talent, little is known on the variability in composite scores on the TTCT-Figural and student demographics (i.e., race/ethnicity, sex, socioeconomic status, English language learning status). This study evaluated student demographic subgroup differences that exist after the initial phase of an identification process (i.e., universal screening, referrals) and examined the relationship among student demographics (i.e., race/ethnicity, free/reduced lunch status, English language learning status, sex), cognitive ability, academic achievement, and creativity, as measured by the TTCT-Figural Form A or B, to the probability of being identified for gifted programs. In a midsized school district in the state of Texas, findings indicate several demographic differences for students who were referred or universally screened across the measures of cognitive ability, academic achievement, and creativity. However, there were lower differences when using the TTCT-Figural. Results of a hierarchical generalized linear regression indicate underrepresented groups showed no difference in the probability of being identified after controlling for measures of cognitive ability, academic achievement, and creativity. Though, cognitive ability and academic achievement tests were more predictive of identification compared to the TTCT-Figural. Implications and recommendations for future research are discussed.
