In vivo occipital asymmetry (OA) measurements have been used to infer functional asymmetries such as language dominance. We investigated the degree of correlation between OA measurements derived by the same rater from different scans of the same subject. We used magnetic resonance (MR) and computerized axial tomography (CT) to study correlations between MR:MR, CT:CT and CT:MR. The highest intrasubject correlation was for MR:MR (r = 0.79). The CT:CT value was similar (r = 0.78) and the CT:MR correlation was slightly lower (r = 0.72). The findings indicate that the reliability of occipital asymmetry measurements is modest at best, setting a low ceiling on the valid use of this variable to infer other indexes.