Concurrent and predictive validity of the Universal Nonverbal Intelligence Test and the Leiter International Performance Scale - Revised

The relationship between two tests of cognitive ability measured nonverbally and the relative capability of intellectual constructs measured by those tests to predict academic achievement (operationalized by end of the year group achievement tests) was examined. One hundred elementary and middle school students were administered the Universal Nonverbal Intelligence Test (UNIT) and the Leiter International Performance Scale - Revised in counter-balanced order; reading, math, and language scores from Terra Nova were matched with 37 cases in the sample. Correlation coefficients describing the relationship among global scores within the IT and between the UNIT and the Leiter-R were statistically significant (p < .001) and ranged from .33 for the. UNIT Memory Quotient/Leiter-R Fluid Reasoning scores t .90 for the UNIT Full Scale IQ/UNIT Reasoning Quotient and UNIT Nonsymbolic Quotient/UNIT Reasoning Quotient. The coefficient between the UNIT and Leiter-R Full Scale IQ scores was .72 (Q < .001). Mean differences between the UNIT and Leiter-R Full Scale were significant with the UNIT Full Scale IQ score being approximately five points higher than the Leiter-R, t. = 4.73, )2. < .001. Effect size for the t-test was modest, .35. Based on stepwise multiple regression analyses, the UNIT Full Scale IQ predicted all three areas of academic achievement significantly better than the Leiter-R Full Scale IQ score, with the variance accounted for by the UNIT Full Scale IQ score ranging from 39 percent to 55 percent (p. < .01). The Leiter-R contributed an additional 2 percent of variance. In addition, a number of the UNIT and Leiter-R global scores were statistically significant predictors of achievement (e.g., UNIT Reasoning Quotient, Memory Quotient, Leiter-R Reasoning). Results are consistent with prior research that has found the UNIT and Leiter-R to provide comparable measures of general intelligence. However, this is the first study to suggest that the UNIT may be superior to the Leiter-R in its relative capability to predict academic achievement. School psychologists and administrators will find these results useful in choosing assessment instruments to evaluate the increasingly culturally and linguistically diverse population of students.