Spatial ability of first year university students is measured and evaluated in this paper. We used standard Mental Cutting Test (MCT), where a body is given by perspective view and correct cross section has to be chosen. While gender differences in MCT are reported by several papers including our earlier results, much less known are the reasons of these differences. Here we show that typical mistakes (answers to problems which are close to be correct) can be one of the possible reasons, since female students made typical mistakes in some cases more frequently than males.