Background. We explored whether the cognitive performance of gender identity disorder patients (GID) was comparable to that of their biological sex or skewed towards that of their gender identity. Method. We tested four potentially sex-sensitive cognitive factors (rotation, visualization, perception, and verbalization) as well as two neutral factors (logic and arithmetic) in GID patients from Norway (GID-N, n=33) or the USA (GID-US, n=19) and in a control group (C, n=29). The testing was undertaken prior to cross sex hormone treatment. Four-way ANOVA was applied in the final analysis of the cognitive performance and its dependency on different predictors (age, biological sex, education, group). Results. In both GID groups as well as in the control group (C) males excelled in visualization and rotation, also when controlling for potential confounders (biological sex, group, age and education). No female advantage was detected. Furthermore, no interaction between biological sex and group assignment was revealed in the samples. Conclusion. In this study the cognitive pattern of GID patients is consistent with that of their biological sex and not that of their gender identity. (C) 2003 Elsevier Science Ltd. All rights reserved.