Research suggests that test design can interact with symptoms of autism spectrum disorder (ASD) such that different standardized cognitive tests may be more or less likely to evoke behaviors that interfere with assessment of functioning. This study compares the results of two recommended and commonly used cognitive assessments (Wechsler Intelligence Scales for Children --- Fourth Edition and Differential Ability Scales --- Second Edition) in a sample of children diagnosed with ASD and considers the influence of autistic symptomatology, adaptive behavior, and behavior during test administration on cognitive scores.

Convergent validity was established for these instruments. However, participants achieved significantly higher overall scores on the DAS-II. Nearly half of the sample received different classification labels on the two assessments. All of these children had higher DAS-II scores. More than a quarter of the sample achieved scores with non-overlapping 95% confidence bands. Behavior during test administration did not systematically vary between tests and was not related to cognitive scores, with the exception of a negative association between off-topic behavior and overall cognitive scores. Autistic symptomatology was not associated with cognitive scores, while adaptive behavior was positively associated with scores. Neither was associated with the magnitude of difference between overall scores on the two assessments. The difference between overall scores was found to be attributable to a relative weakness in processing speed, which is assessed on the WISC-IV but not the DAS-II.

Clinically, this study suggests that cognitive assessment is a valuable part of a comprehensive assessment for children with ASD as these tests appear to be minimally impacted by construct-irrelevant variance. However, choice of assessment should be considered carefully given the systematic differences in overall scores produced in this population. Similarly, results suggest that patterns of responding in children with ASD on select subtests deviate from the standardization sample, possibly due to symptoms of ASD. Examiners should be familiar with these issues and interpret tests accordingly. Future research should investigate the effects of examiner experience and role of non-standardized aspects of the assessments protocols.