IQ estimate smackdown: Comparing IQ proxy measures to the WAIS-III

Abstract

Brief assessments of general cognitive ability are frequently needed by neuropsychologists, and many methods of estimating intelligence quotient (IQ) have been published. While these measures typically present overall correlations with the Wechsler Adult Intelligence Scale (WAIS) Full Scale IQ, it is tacitly acknowledged that these estimates are most accurate within 1 standard deviation of the mean and that accuracy diminishes moving toward the tails of the IQ distribution. However, little work has been done to systematically characterize proxy measures at the tails of the IQ distribution. Additionally, while these measures are all correlated with the WAIS, multiple proxy measures are rarely presented in one manuscript. The current article has two goals: (1) Examine various IQ proxies against Wechsler Adult Intelligence Scale (Third Version) scores, showing the overall accuracy of each measure against the gold standard IQ measure. This comparison will assist in selecting the best proxy measure for particular clinical constraints. (2) The sample is then divided into three groups (below, average, and above-average ability), and each group is analyzed separately to characterize proxy performance at the tails of the IQ distribution. Repeated measures multivariate analysis of variance compares the different proxy measures across ability levels. All IQ estimates are represented in tables so that they can be examined side by side.

Evaluating the use of the Child and Adolescent Intellectual Disability Screening Questionnaire (CAIDS-Q) to estimate IQ in children with low intellectual ability

“…There will also be occasions where estimates of IQ will be used for estimating and describing population characteristics (Moss & Hogg, 1997) or to identify those potentially at risk, such as screening children in educational settings (Sonnander, 2000). Clinical researchers may also utilise IQ estimates (Spinks et al, 2009) in order to stratify participants appropriately or match groups in terms of participants' intellectual functioning. Here, conducting full intellectual assessments on large populations is unlikely to be practicable because of the time and resources required.…”

“…A number of authors have discussed the relative advantages and disadvantages of these methods in general (e.g. Kaufman & Kaufman, 2001;Spinks et al, 2009), but with relatively less attention on their utility when used with people with an intellectual disability. Research with adults suggests that such measures overestimate the IQ of those with an assessed FSIQ of below 85 (Spinks et al, 2009).…”

“…Kaufman & Kaufman, 2001;Spinks et al, 2009), but with relatively less attention on their utility when used with people with an intellectual disability. Research with adults suggests that such measures overestimate the IQ of those with an assessed FSIQ of below 85 (Spinks et al, 2009). There has, however, been very limited research examining the performance of such tools with children with an intellectual disability.…”

“…In terms of BPS intellectual disability severity classifications, Murray & McKenzie (2014) found that although the scale could provide a heuristic for estimating which category young people aged 12-18 would be placed in, it could not do so with a degree of Collectively these studies would suggest that the CAIDS-Q could have utility in situations where a quick estimate of FSIQ is required, for example in clinical research, but where restricted or no access to an appropriately qualified psychologist precludes the use of short form intellectual assessments. Further impetus for exploring this question formally comes from the fact that, in general, proxy measures of IQ perform poorly with people with below average intellectual functioning (Spinks et al, 2009; but see ).…”

Assessing Cognitive Improvement in People with Down Syndrome: Important Considerations for Drug-Efficacy Trials

“…These assessments are standardized in the typically developing population and factor out validity and reliability issues that crop up during psychological testing ), yet are difficult to interpret in people with intellectual disabilities because they invariably fall on the extreme end of the performance distribution below the first percentile (Couzens et al 2004;Hessl et al 2009;Nelson and Dacey 1999). At the tail end of the scales, estimates of performance are significantly less accurate, and there is narrower resolution to detect meaningful change (Couzens et al 2004;Silverman et al 2010;Spinks et al 2009). Normative testing misses a key point in understanding how cognitive improvement might occur in an individual with DS; this improvement might be better compared to other people with DS, less so to the function of the everyday person.…”