A meta-analysis of the reading comprehension skills of individuals on the autism spectrum.
Reading problems in autism usually trace to weak decoding or vocabulary, not the diagnosis itself.
01Research in Context
What this study did
The team pooled 36 studies on kids and adults with autism. They asked one question: how far behind are these learners in reading comprehension?
They used a meta-analysis. This means they turned every study's results into one common score. Then they averaged those scores to see the big picture.
What they found
Overall, the autism group scored half to three-quarters of a standard deviation below typical readers. That is a medium-large gap.
But the gap shrank when learners had strong word-reading skills or a rich vocabulary. The label 'autism' mattered less than these two teachable skills.
How this fits with other research
Heinicke et al. (2012) showed that small-group direct instruction with prompts works for almost every learner. Capio et al. (2013) now tells us what to prompt: decoding and semantic knowledge.
Matson et al. (1989) warned that judges rarely agree on the 'semantic base' of a text. The meta-analysis agrees—semantic skill, not fuzzy text ratings, drives comprehension.
Melchiori (2000) taught kids to recombine syllables to read new words. That stimulus-equivalence work backs the meta finding: solid decoding unlocks meaning.
Why it matters
Before you place a child in a comprehension group, test two things: can they read the words, and do they know what the words mean? If either skill is weak, teach that first. Use discrete trials for decoding and mand training for vocabulary. Re-test comprehension after a few weeks—progress may surprise you.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Run a 2-minute timed word-reading probe and a quick picture-vocabulary check; pick the lower score for your first teaching target.
02At a glance
03Original abstract
This meta-analysis examined 36 studies comparing autism spectrum disorder (ASD) and control groups in reading comprehension. Three moderators (semantic knowledge, decoding skill, PIQ) and two text types (high vs. low social knowledge) were examined as predictors of reading comprehension in ASD. The overall standardized mean difference for reading comprehension was g = -0.7 SD. The strongest individual predictors of reading comprehension were semantic knowledge (explaining 57 % of variance) and decoding skill (explaining 55 % of variance). Individuals with ASD were significantly worse at comprehending highly social than less social texts. Having ASD alone does not predict reading comprehension deficits. Instead, individual skills, especially language ability, must be considered before one can accurately predict whether a given individual with ASD will experience difficulties in reading comprehension.
Journal of autism and developmental disorders, 2013 · doi:10.1007/s10803-012-1638-1