An analysis of the reliability and validity of the Naglieri Nonverbal Ability Test (NNAT) with English language Learner (ELL) Mexican American children

Abstract

The purpose of this study was to investigate the reliability and validity of the results of the Naglieri Nonverbal Ability Test (NNAT; Naglieri, 1997a) with a sample of English Language Learner (ELL) Mexican American children and to compare the performance on the NNAT of 122 ELL Mexican American children with children from the standardization sample. The rationale for conducting this study was the need to identify culturally sensitive and technically adequate nonverbal measures of ability for the fastest growing minority group within America’s public schools today, Mexican American children. The NNAT was administered to participants with parental consent. Statistical analyses of the scores did yield positive evidence of internal consistency for the Nonverbal Ability Index (NAI) total score of the NNAT. However, when individual clusters were analyzed, Pattern Completion, Reasoning by Analogy, and Serial Reasoning did not yield positive evidence of internal consistency. Only Spatial Visualization approached the reliability standard deemed acceptable for tests of cognitive ability. The mean differences of the NNAT scores between two independent groups were also assessed in the present study. Results of the statistical analyses did not yield statistically significant differences across age and grade factors between the scores of the ELL Mexican American sample and the standardization sample. Finally, the proposed factor structure of the NNAT was compared with the factor structure found with the ELL Mexican American sample. Goodness-of-fit test statistics indicate that the proposed four-factor structure does not fit well with the data obtained from this sample of ELL Mexican American students. Furthermore, although the NNAT is considered to be a unidimensional test of general ability, nine factors were extracted upon analysis, providing evidence that the items on each of the four clusters do not function together as four distinct dimensions with this ELL Mexican American sample. Given that the individual clusters that collectively combine to yield the NAI total score are not based on any particular model of intelligence, interpretation of specific strengths and weaknesses should be discouraged. Finally, the NNAT’s overall score should be interpreted with caution and may best be used in conjunction with multidimensional ability and/or intelligence measures.