The intercorrelations between DAS subtests and scales (see Tables 9.1-9.3, pp. 198-200, DAS Introductory and Technical Handbook) indicate that, in the total group, correlations among the 20 subtests (including the three achievement subtests but excluding the Pattern Construction-alternative method) range from a low of .07 (Picture Similarities and Recall of Objects-Delayed) to a high of .68 (Recall of Objects-Immediate and Recall of Objects-Delayed). Among the twelve subtests that compose the core battery at the 3 levels, the highest intercorrelations are between verbal subtests: Lower Preschool and Upper Preschool Verbal Comprehension and Naming Vocabulary (.61 and .64 respectively) and School-Age Word Definitions and Similarities (.64). The lowest subtest intercorrelations are between Lower Preschool Block Building and Picture Similarities (.28); Upper Preschool Naming Vocabulary and Pattern Construction (.28); and School-Age Word Definitions and Similarities with Recall of Designs (both .38).

In the School-Age group, each of the two subtests that make up a cluster correlated higher with each other than they did with any of the other subtests from other clusters. The Verbal Cluster subtests correlate more highly with each other (r = .64) than they do with either the Nonverbal Reasoning subtests (M r = .48) or the Spatial subtests (M r = .40). The Nonverbal Cluster subtests correlated higher (.58) with themselves than with the Verbal Cluster subtests (M r = .48) or the Spatial Cluster subtests (M r = .49). Additionally, the Verbal Cluster has more in common with the Nonverbal Reasoning Cluster, and the Nonverbal Reasoning Cluster has more in common with the Spatial Cluster than do the Verbal and Spatial Clusters with each other. The implication of this interaction will be discussed later in the chapter, but it does support the hypothesized integrative function required by the Nonverbal Reasoning (fluid intelligence) subtests.

Average correlations between all subtests and the GCA range from .18 to .81. As expected, those subtests that, because of high g loading, are used to create the GCA correlate better with the GCA than did the lower g-loaded (diagnostic) subtests.

Average correlations between the 17 individual subtests (excluding achievement subtests) and the GCA range from .22 to .82. On the Preschool battery, Early Number Concepts has the highest correlation with the GCA (.82), followed by Verbal Comprehension (.79), Similarities (.71), Pattern Construction (.67), Picture Similarities and Copying (.65), Block Building (.58), Matching Letter-Like Forms (.51), Recognition of Pictures (.47), Recall of Digits (.44), and Recall of Objects (.22). On the School-Age battery, Sequential and Quantitative Reasoning has the highest correlation with the GCA (.79), followed by Matrices (.76), Similarities (.75), Word Definitions (.74), Recall of Designs and Pattern Construction (.71), Recall of Digits (.33), Recall of Objects (.31), and Speed of Information Processing (.25). Overall, the average correlation between the core subtests and the GCA is .74, compared to the average of .39 for the diagnostic subtests. These findings support the use of the DAS as a measure of general mental ability. Although the subtests vary in their correlations with the GCA, and in their respective reliabilities, the estimation of general intellectual functioning as measured by the GCA appears justified.