• Identify any significant differences between the DAS GCA and each cluster (Verbal, Nonverbal Reasoning, and Spatial)

Determining whether the GCA is an accurate overall representation of the person's abilities requires that the examiner determine how discrepant each cluster (Verbal, Nonverbal Reasoning, and Spatial) is from the GCA itself. The DAS provides examiners easy access to the discrepancy requirement for each comparison. The DAS Introductory and Technical Handbook Table B.1. and Table B.4. (pp. 290-292) show the difference required for statistical significance at three levels (.15, .05, and .01) for the entire age appropriate samples (Preschool and School-Age) as well as significance at each of 16 age categories. Rounded mean values at the .05 significance level are also found on the protocol Summary Page. For School-Age children, the mean value for each of the three comparisons is approximately 9 points, while for Upper Preschool children, the Verbal and Nonverbal mean values are 9 and 8 respectively.

• Identify the frequency of any observed significant differences

If discrepancies do exist between the cluster scores and the GCA, examiners must determine the frequency of such occurrences. Significant discrepancies, although important in developing hypotheses about a child's performance, are often found with surprisingly high frequency in samples of children. It should be noted that the GCA-Cluster comparisons for the DAS standardization sample typically demonstrated that, if one disregards the direction of the differences, approximately 25% of the population is expected to display at least one significant GCA-Cluster discrepancy. For example, although School-Age children on average require a 9-point difference between the GCA and any one of the cluster scores, a 9-point difference is expected to occur in over 25% of the cases [See DAS Introductory and Technical Handbook Table B.2. and Table B.3. (p. 291). All the percentages in these two tables reflect those children from the standardization sample having a difference in either direction]. Although the DAS Introductory and Technical Handbook does not provide frequency data for specific direction of difference (e.g., GCA>V vs. GCA < V), one can estimate such differences by halving the size of the percentage shown in these tables. For example, Table B.3. indicates that a child obtaining a difference of 15 points between the GCA and the Verbal ability score would be similar to about 10% of the standardization sample. If examiners were interested in only GCA > Verbal Ability, a difference of 15 points might be expected in approximately 5% of the population.

• If there is at least one difference that is significant and unusual, interpret clusters rather than the GCA.