How Accurate Are Teachers at Estimating Cognitive Abilities of Children Who Are on the Autism Spectrum?

Background:  Clinical experience indicates that teachers often over- or under-estimate the intellectual abilities of children with Autism Spectrum Disorders (ASDs) who are in mainstream education. No previous study has explicitly measured objective IQ and compared this with teacher reports. Evaluating teachers’ accuracy in estimating cognitive abilities in children with an ASD may provide useful information in planning school based interventions.

Objectives:  We aimed to examine how accurate teachers are at estimating intelligence in a large heterogeneous sample of children with an ASD.  We aimed to test the hypothesis that both cognitive and behavioral child characteristics, and attainments in reading, numeracy and oral expression, would influence their teacher’s estimation of general intellectual ability.

Methods: Data from school reports and clinical assessments were analyzed for a sample of 106 children (mean age 10.0 years, mean full scale IQ 88) who met DSM-IV-TR criteria for either Asperger syndrome (n=52), autism (n=35), or PDD-NOS (n=19). Each participant attended a different mainstream school. Their teachers were asked to estimate IQ as being: i) above average (111+); ii) average (91-110); iii) low average (81-90); iv) well below average (<80). Results were compared to actual IQ scores, measured using the WISCV-IV. Behavioral difficulties, also rated by teachers, were measured using the Strengths and Difficulties Questionnaire, the Social Communication Disorders Checklist, and the Conners 3rd edition, as potential confounding variables.

Results:  30/106 teachers felt unable to estimate the child’s intelligence, but this did not characterize any subgroup on the basis of their intellectual or behavioral characteristics. When teachers did offer an IQ estimate, there was a weak association between estimated and actual full-scale IQ with most children receiving an inaccurate IQ estimate (n=45/76, 59%).  The great majority (n=39/45) of these inaccuracies involved overestimating the child’s abilities with underestimations being unusual (n=6/45). Accuracy was highest for children with above average abilities (90%; n=9/10) and was lowest in children with below average abilities (average IQ=42%, n=11/26; low average=30%, n=4/13; well below average=25%, n=7/27). There was no correlation between estimates and the child’s verbal-performance IQ discrepancy.

Overestimation of IQ was associated with a lack of conduct problems (p=.03), but not with prosocial behavior. Neither peer difficulties nor ADHD severity were related to the accuracy of IQ estimate. For children with abilities in the average range, overestimation was significantly associated with above average reading skills (p = .02). A similar trend was found in those with below-average intelligence (p=.06). Children whose IQ was underestimated tended to have a poor working memory (p = .07)

Conclusions:  These findings indicate that the majority of teachers in mainstream education overestimate the cognitive abilities of children with an ASD. This discrepancy is explained in part by child attributes, such as a lack of problematic behavior at school and good reading attainments. Children with poor working memory skills may appear to be less able. Overestimation of abilities may lead to unrealistic expectations, and potentially to disappointment and frustration on the part of children and their parents.