Subtest specificity refers to the proportion of a subtest's variance that is both reliable (that is, not due to errors of measurement) and distinct to the subtest. Although individual subtests on the DAS overlap in their measurement properties (that is, the majority of the reliable variance for most subtests is common factor variance), all possess sufficient specificity to justify the interpretation of a specific subtest functions. [On any instrument, subtests with inadequate specificity should not be interpreted as measuring specific functions.] These DAS subtests, however, can be interpreted as (a) good or (b) fair measures of g. McGrew and Murphy (1995) consider a test’s specificity to be high if it is (a) .25 or more, and (b) it is greater than the proportion of error variance. All specificity values for the DAS subtests exceed .25, the lowest being .30 (Similarities) and the highest being .82 (Speed of Information Processing). In every case, all subtest specificities substantially exceed the proportion of error variance.

Elliott (1997) examined and compared the DAS subtest specificities with those found on a number of other popular cognitive test batteries (WPPSI-R, WISC-III, WISC-R, K-ABC, SB:FE, and the WJ-R COG). He found that, when compared to these other measures, the DAS had approximately one third more reliable subtest specificity than the other batteries. While other tests had approximately 35 to 37 percent of specific variance, the DAS averaged 47 percent for the Preschool battery and 50 percent for the School-age battery. As he notes "These results show the DAS to have about one third greater specificity than other batteries, and strongly suggests that the original development goal of a battery with reliable, specific, individually interpretable subtests has been achieved" (p. 195).