WAIS-III measurement invariance: Data from Estonian standardization

Abstract

Objective A prerequisite of any psychological instrument used to compare individuals from different groups is measurement invariance (MI). It indicates that the test measures the same psychological constructs regardless of the particular grouping variable of the test-taker. Our purpose was to evaluate the MI across sex, age groups and educational levels in the recently adapted Estonian version of the Wechsler Adult Intelligence Scale – Third Edition (WAIS-III). Method We analysed the Estonian standardization sample of WAIS-III (N = 770) with confirmatory factor analysis (CFA) to establish the best baseline factor model for further analysis. Multi-group confirmatory factor analysis (MG-CFA) was applied to evaluate MI of the test and, granted this, mean differences across sex, age groups and educational levels. Results CFA supported the four-factor model. The test demonstrated partial MI across sexes; latent mean comparisons showed that men had a significantly higher mean score on the Perceptual Organization factor. Partial MI also held across age groups and, as expected, older groups had significantly lower means than younger age groups. The analyses across the educational levels failed to prove the MI as the metric invariance was not tenable. Discussion The results of this study provide evidence that the structural model underlying the Estonian adaption of WAIS-III is partially invariant across sex and age groups, hence the test functions same manner across these groups. Estonian WAIS-III was not invariant across the educational levels, which may indicate that the measure has a different structure or meaning to different educational groups.