A Latent-Variable Modeling Approach to Assessing Reliability and Validity of a Creativity Instrument

Abstract ABSTRACT: A multiple-choice paper-and-pencil test was developed based on the constructs from the Torrance Test of Creative Thinking (TTCT). The rationale of this new instrument was to shorten the amount of time required for the administration and scoring of creativity tests. This instrument was translated into Spanish and administered along with the TTCT and the Villa and Auzmendi Creativity Test (VAT) to 2,270 students in Spain. Teachers' ratings of student creativity were also obtained. Using a traditional concurrent validation approach from an earlier study, we found low but significant correlations between the scores of the four subscales of the new test and the same subscales of the VAT and TTCT, as well as the students' academic achievement measures. In terms of reliability, mostly moderate (internal consistency) reliability coefficients were found for the new test. As an alternative to the traditional approach, this study employed structural equation modeling (SEM) with multiple indicators to examine the validity and reliability of the new creativity test. The results indicated that substantial improvements in the validity and reliability estimation over the traditional approaches can be made by using SEM with multiple indicators.