Objectives. Previous factor analytic studies of the WAIS, WAIS-R and WAIS-III used standardization samples, (representing the general population), clinical populations (e.g. psychiatric) and 'non-clinical' groups (e.g. older adults). These studies endorsed the reliability of the scales in such samples and supported theoretical models of intelligence. The WAIS-III (1997) includes four Indexes based on factor analysis, which provide clinically useful information to practitioners, but have not been validated in a low IQ population. However, the WAIS-III is often administered to individuals with suspected or proven low IQ, as it offers service providers, legislators and the Department of Health with a reliable measure of IQ. The aim of this study was to investigate the factor structure of the WAIS-III in a low IQ sample. Methods. WAIS-III data was collected from assessments carried out in routine clinical practice from individuals with a full scale IQ of 74 or below (n = 105). The data were subjected to factor analysis, using two types of factor analysis: principal axis factoring and principal components analysis. Orthogonal and oblique rotations were applied to the analyses. Results. Only one robust solution could be extracted. This contained two factors, analogous to traditional verbal and performance sub-scales. Conclusions. This study does not support the four-factor solution which underlies WAIS-III index scores in a low IQ population.