The General Ability Measure for Adults (GAMA; Naglieri & Bardos, 1997), is a 66 item measure of nonverbal intelligence that may be a good candidate for the evaluation of nonverbal ability across cultures as it reduces potential confounding variables such as linguistic background, and exposure to formal education.  The current study investigated differential item functioning (DIF) on the GAMA using stratified samples of young adults (19 – 22 years old) from the United States (n = 276), Mexico (n = 131), and Greece (n = 217).  This age group was selected as many career or placement choices are made during this age period involving a measure of general cognitive ability (i.e. military).

DIF analysis was performed using the Mantel–Haenszel procedure (MH; Holland & Thayer, 1988). The 66 test items were dichotomously coded and the GAMA total raw score was used as a matching criteria to control for participant ability level. Thin matching was used, in which each score level (total raw score of 1 to 66) serves as its own ability category. The matching criteria was further purified using a two step iterative process. In each analysis the United States was used as the reference group.

The MH Chi Squared Test of Conditional Independence was used to test for significant differences, followed by transformations into Delta MH (Dorans & Holland, 1993). Using the three level DIF item performance by the Educational Testing Service classification system one item in the Greek sample and none of the items in the Mexican sample contained a large amount of DIF.

These findings suggest that the GAMA is highly compatible across samples of young adults from Greek and Mexican populations and any standardization efforts in Greece and Mexico should not anticipate significant differences in performance for this age group.