The literature on the dimensional structure of Rotter's Locus of Control Scale presents a confused picture, with solutions ranging from two to nine factors reported. In the present study, Revelle and Rocklin's (1979)'Very Simple Structure' (VSS) method was used to determine the optimum number of interpretable factors to extract from the two-factor model represented the optimum solution, and that there was no justification for extracting more than two factors for the purpose of subscale construction. Oblique rotations, as compared with orthogonal Varimax rotation, did not improve the goodness-of-fit of the two-factor solution. When data from males and females were analysed separately, there was no evidence of differences in factorial structure. The findings illustrate how the VSS technique, by providing a quantitative basis for deciding how many factors to extract from a data set, can be used to clarify inconclusive results obtaind from the application of conventional psychometric 'rules of thumb' in factor analysis.