Executive dysfunction is thought to be primary to autism. We examined differences in executive function between 20 adults with autism and learning disability and 23 individuals with learning disabilities outside the autistic spectrum. All participants were matched for chronological age and full-scale IQ, and were given a battery of tasks assessing fluency, planning, set-shifting, inhibition and working memory. Analyses of the individual tasks revealed very few significant differences between the two groups. However, analyses of composite scores derived for each executive domain revealed that the group with autism showed impaired performance on the working memory and planning tests. Together, these two measures were sufficient to classify participants into their diagnostic groups significantly better than would be expected by chance (75% of the autism group; 65% of the control group). Executive impairments were neither universal nor exclusive to the autism group, and we suggest that an alternative cognitive theory may better explain the cognitive profile we found.