A Meta-Analysis Examining the Academic Achievement of Individuals with ASD in Reading, Writing and Mathematics

Background:  Media reports and popular opinion often portray individuals with ASD either as academically impaired or perpetuate the stereotype of ‘autistic genius’. Unfortunately, little empirical research examines the academic profile of individuals with ASD. One notable exception was a research synthesis by Schaefe-Whitby & Richmond-Mancil (2009) where the results were somewhat mixed. On one hand, they report that individuals with ASD have deficits in comprehension, written expression, and problem solving, yet they also state that their reading, math and writing skills are commensurate with IQ. The current meta-analysis of all studies examining academic achievement between 1990 and the present will clarify whether there is one academic profile for individuals with ASD, the size of strengths or weaknesses across domains and the consistency of these differences.

Objectives:  1. To use meta-analysis to determine the size, direction, and consistency of differences between academic achievement scores and PIQ, among individuals with ASD. 2. To explore whether the effect sizes within each of five academic areas (writing, reading comprehension, decoding, math computation, and math reasoning) are heterogeneous.  3. To describe the predicted range within which their academic abilities vary (across each of the five areas) for individuals with ASD.

Methods:  We conducted five meta-analyses of 8 to 18 studies of individuals with ASD that included one or more measures per academic area. Standardized mean differences (SMDs) between the academic achievement scores and PIQ scores were calculated using a random effects model and Hedge’s g SMD. The statistical significance of the grand SMD for each academic area was tested with a Z-test, and a Q-test was used to determine whether there was heterogeneity in effects. The range of the population SMD was estimated using the tau-squared method.

Results:  

1. The models revealed that academic skills of individuals with ASD were commensurate with PIQ for decoding and math reasoning. While there were reliable deficits between PIQ and reading comprehension (g = -0.4), math computation (g = -0.4) and writing (g = -0.6), they have little clinical significance.

2. The differences between academic ability and PIQ were heterogeneous suggesting that the overall finding that academic skill does not differ significantly from PIQ is unlikely true for all individuals or groups of individuals with ASD.

3. The range within which the difference between PIQ and academic achievement varied was within ±1SD for most of the skills studied. However, the range was generally negatively skewed meaning that a larger proportion of individuals with ASD had weaker skills than predicted by PIQ as opposed to stronger than expected ability. 

Conclusions:  Academic achievement scores for individuals with ASD vary within the normal range for most of the skills studied. However, with the exception of decoding ability, individuals with ASD were more likely to have difficulties mastering the academic skill than to excel in a given area, and the area of greatest concern was written expression. These meta-analyses support previous findings that the most defining feature of ASD is variability, and that ASD is not associated with any particular academic profile.