Measurement of Working Memory on the Wechsler Adult Intelligence Scale: Should We Subtract Arithmetic?

Deficits in working memory are often used as evidence to support the diagnosis of a neuropsychological disorder or acquired brain injury. Subtests that comprise the working memory index (WMI) of the WAIS-IV are assumed to all measure the same construct; however, deficits in basic arithmetic skills may unduly influence performance on the Arithmetic subtest, particularly among students with neurodevelopmental disabilities. The current study examined the relationship between working memory scores and measured mathematical ability in a sample of 605 postsecondary students seeking psychoeducational assessments to verify their need for academic accommodation. Different combinations of the three WAIS-IV working memory subtests’ scores resulted in significantly different composite WMI scores. Across diagnostic groups, subtest scores were highest on Letter-Number Sequencing and Digit Span and lowest on Arithmetic. Furthermore, math achievement scores were more strongly correlated with Arithmetic scores than with Digit Span or Letter-Number Sequencing scores. Students with learning disabilities performed significantly more poorly on Arithmetic relative to students with attention deficit hyperactivity disorder. Given the risk of clinicians falsely diagnosing a working memory impairment, we therefore advise clinicians to substitute out the Arithmetic subtest, which requires math skills, when evaluating young adults for neurodevelopmental conditions. This will avoid them being falsely identified as having a working memory impairment when in fact their working memory is intact.