Patterns of visual-spatial performance and 'spatial ability': dissociation of ethnic and sex differences

Is there a common basis for the ethnic and sex differences that are characteristically obtained on psychometric tests of spatial ability? Three experiments approached this question by observing subject differences in the recognition and reconstruction of visual-spatial displays. The pattern of performance on these experimental tasks was compared with that on a traditional spatial ability test. In the first experiment, two samples of 40 students, balanced for sex, from Zimbabwe and Scotland respectively, attempted a forced-choice recognition task for meaningful scenes. Both ethnic groups and both sexes showed equivalent performance. The same subjects then undertook a task involving the reproduction of an arrangement of blocks into two-dimensional plan and elevation views. On this task, involving spatial reorientation, the Zimbabweans made over three times as many errors as the Scots. In a third experiment the requirement for spatial reorientation was added to the original recognition task and this was performed by a further 40 subjects. A significant difference between ethnic groups now emerged and this effect covaried with spatial ability. Again, however, no sex difference was observed. The overall pattern of results points to spatial reorientation as a major factor in the cross-ethnic differences. The absence of a sex difference on the experimental tasks contrasts with its appearance in both samples on the spatial ability test and represents a puzzling obstacle to our current understanding. This dissociation of sex and ethnic differences provides evidence against the hypothesis that they stem from the same source.