BACKGROUND The aim of this study was to explore changes related to sex differences on the Wechsler Intelligence Scale for Children - Revised (WISC-R) subtest performance over a 7-year interval in middle-aged adults with intellectual disability (ID). Cognitive sex differences have been extensively studied in the general population, but there are few reports concerning individuals with ID. Sex differences are of current relevance to actively debated issues such as cognitive changes during menopause and risk for Alzheimer's disease. Given that hormonal effects on cognition have been observed in the general population, particularly in areas such as visuospatial processing, and individuals with Down's syndrome (DS) have been reported to be hormonally and reproductively atypical, we analysed our data to allow for the possibility of an aetiology-specific profile of sex differences for these adults. METHODS The WISC-R subtests were administered in a longitudinal study, as part of a more comprehensive test battery, at least twice within 7 years. Participants were 18 females with ID without DS [age at first test time (time 1): mean = 40.5; IQ: mean = 59.3], 10 males with ID without DS (age at time 1: mean = 42.4; IQ: mean = 59.4), 21 females with DS (age at time 1: mean = 37.9; IQ: mean = 51.6), and 21 males with DS (age at time 1: mean = 40.3; IQ: mean = 54.3). All participants were in the mild to moderate range of ID and were displaying no changes suggestive of early dementia. RESULTS Females, regardless of aetiology of ID, exhibited a robust superiority on the coding subtest, which parallels the widely reported difference among adults in the general population. Additionally, there was a decline in overall performance during the 7-year study interval, particularly on the verbal subscale subtests, but there was no evidence of sex-differentiated decline. There were also marginal sex by aetiology interactions on the object assembly and block design subtests, suggesting that males with unspecified ID might perform better than their female peers, but among adults with DS, females might do better than males. CONCLUSIONS This study supports the presence of cognitive sex differences in the population with ID as indicated by female superiority on the WISC-R coding subtest. Extending this observation to adults with ID has implications for explanations of female advantage on this task, which now have to account for its presence among individuals with a broader range of intellectual capabilities, more atypical developmental histories and more varied genotypes than previously considered. Trends towards sex by aetiology interactions on the two visuoconstructive subtests, while marginal, were sufficient to warrant continued consideration of the idea of a distinct profile of sex differences for adults with DS and to justify looking at the effects of sex separately within different aetiologies of ID.