The GAMA was standardized with a sample of 2,360 adults ages 18 to 96 that was stratified to closely match the 1990 U.S. Census data. The internal consistency of the GAMA IQ score was strong; across 11 age intervals the reliability coefficients ranged from .79 to .94 (the average was .90). The internal consistency coefficients for the individual subtests are adequate for Analogies (.81) and Sequences (.79) and low for Matching (.66) and Construction (.65) (Naglieri & Bardos, 1997). Two studies have reported the test-retest reliability of the GAMA (Lassiter & Matthews, 1999; Naglieri & Bardos, 1997). The test-retest data in the GAMA manual are based on a sample of 86 adults who were group-administered the GAMA twice in a 2- to 6-week interval (mean interval 25 days). The test-retest coefficient for GAMA's Full Scale IQ was .67. Stability coefficients for the four subtests were generally poor: .55 for Matching, .65 for Analogies, .74 for Sequences, and .38 for Construction. Lassiter and Matthews replicated the test-retest study with a sample of 25 college students (77% male). The subjects were administered the GAMA twice in a 22- to 48-day interval (mean interval 35 days). The stability coefficients were even lower than those reported in the GAMA manual and were quite unimpressive (.62 for Full Scale IQ and .31-.54 for subtests).
The concurrent validity of the GAMA has been assessed by its correlation with several other measures of global cognitive ability, such as the WAIS-III, WAIS-R, WJ-R, KAIT, K-BIT,
Shipley-Hartford, and Wonderlic Personnel Test. A sample of 60 college students (83% male) were administered the gAmA and WAIS-III in counterbalanced order (Lassiter, Bell, Hutchinson, & Matthews, in press). Overall, the GAMA IQ was significantly lower than the mean WAIS-III FS-IQ, V-IQ, POI, and VCI, but was significantly higher than PSI. There were no significant differences between the GAMA IQ and WAIS-III P-IQ and WMI. GAMA IQ correlated .59-.78 with WAIS-III IQs, with the highest value (as expected) obtained with P-IQ. The relationship between the GAMA and P-IQ was significantly stronger than GAMAs relationship with the V-IQ. GAMA IQ correlated .80 with POI and .41-.50 with the other three indexes. Lassiter et al. also found that, for individuals with Low Average to Average IQ on the WAIS-III, the GAMA IQs were within 4 points of the WAIS-III FS-IQ and P-IQ. However, for individuals with Superior or Very Superior WAIS-III FS-IQs, the GAMA IQ mean was approximately 8 to 14 standard-score points lower than the WAIS-III FS-IQ or P-IQ. Thus, the GAMA may not provide an accurate estimate of WAIS-III FS-IQ and P-IQ for individuals who are functioning in the Superior to Very Superior range of intelligence on the WAIS-III.
A sample of 80 college students with learning difficulties was used to examine the validity of the GAMA in comparison with the WAIS-R (Lassiter, Leverett, and Safa, 2000). Similar to the findings with the WAIS-III and college students without learning difficulties, the GAMA correlated most strongly with the WAIS-R P-IQ (.69) and least with the V-IQ (.36). The GAMA-WAIS-R FS-IQ correlation was .60. The GAMA IQ more accurately predicted the ability of individuals with learning difficulties who scored within the Average and High Average ranges on the WAIS-R. In contrast, the GAMA underestimated the WAIS-R IQs of those who fell in the Superior to Very Superior ranges of intelligence.
The relationship between the GAMA and Woodcock-Johnson—Revised Tests of Cognitive Ability (WJ-R) was examined in a sample of 62 male college students (Nagy, Lassiter, & Leverett, in press). Three of the WJ-R's composite scores were compared to the GAMA IQ, including Processing Speed, Visual Processing Speed, and Auditory Processing. There was no significant relationship between WJ-R Auditory Processing and the GAMA IQ, but the WJ-R Processing Speed and Visual Processing Composites were moderately correlated with the GAMA IQ (r = .55 and .49, respectively). The results of this study indicate that the GAMA can accurately be viewed as a measure of processing speed and visual ability, but does not measure other aspects of cognitive ability such as auditory processing.
Maher, Lassiter, Matthews, and Bell (in press) studied the relationship of the GAMA to the KAIT in a sample of 77 college students. Two of the KAIT scales showed moderate correlations with the GAMA: Fluid IQ (r = .47) and Composite IQ (r = 44). Although the correlation between KAIT Crystallized IQ and GAMA IQ was significant, it was of very small magnitude (r = .28). Thus, similar to findings on the WAIS-III and WAIS-R, the GAMA appears to have a small relationship to measures of verbal or crystallized ability, contrasted with moderate relationships to fluid abilities and nonverbal reasoning skills. Also, parallel to findings on the WAIS-III and WAIS-R, the GAMA underestimated the KAIT Composite IQ of individuals who scored above average or better on the KAIT.
Naglieri and Bardos (1997) reported that the GAMA had significant correlations with the KBIT Composite, Matrices subtest, and Vocabulary subtest (r = .70, .72, and .54, respectively). Leverett, Matthews, Bell, and Lassiter (in press) reported a moderate correlation between the GAMA IQ and Wonderlic Personnel Test (r = .46). Overall, the relationship with the GAMA to each of these brief measures of cognitive ability further supports the validity of the GAMA as a measure of general intelligence.
Was this article helpful?
Whenever a doctor informs the parents that their child is suffering with Autism, the first & foremost question that is thrown over him is - How did it happen? How did my child get this disease? Well, there is no definite answer to what are the exact causes of Autism.