Sex differences in spatial cognition: an evolutionary approach

This thesis is concerned with investigating causal explanations for the observed patterns of sex differences in spatial cognition. Researchers in this area have traditionally reported a male advantage in many spatial tasks, particularly mental rotation. However, a number of more recent results suggest that the pattern of sex differences in spatial ability is more complicated: females show a spatial memory advantage on some tasks involving large arrays of objects, and navigational tasks seem to reveal sex differences in attentional bias toward different types of spatial information. After reviewing three main types of causal explanations, it is concluded that the evolutionary approach has at least two compelling advantages. Firstly, it is inherently multidisciplinary, and readily incorporates both biological-proximal and social-developmental explanations, as these mechanisms are integral to the expression of evolved traits. Secondly, it not only fits the available data, but provides a framework for generating novel theories and predictions; indeed, Silverman and Eals (1992) used an evolutionary model of sex differences in spatial ability – the 'hunter-gatherer' model, based on the sexual division of labour hypothesized to underlie hominid foraging and navigational requirements – to successfully predict the previously unobserved female advantage on certain object array tasks. Thus, the primary aim of this thesis is to expand the hunter-gatherer model by further examining the computational demands placed on spatial ability in ancestral humans, and by making increasingly focused predictions about the cognitive skills underlying sex differences in spatial performance. Following a detailed discussion of these computational demands, it is argued that male-typical hunting as a foraging strategy is associated with long range surveybased wayfinding, which requires the allocentric encoding of spatial relationships and enhanced sensitivity to coordinate information such as distance, based on the use of external geometric reference points. By contrast, it is argued that female-typical gathering as a foraging strategy is associated with short range route-based wayfinding, which requires the ability to encode relational egocentric-categorical information about a large number of objects, but does not require enhanced sensitivity to categorical information. In order to examine this proposed model, several different spatial tasks are used and devised, with the aim of measuring sex differences in bias and sensitivity toward different types of spatial information. In all cases the computational demands of the task are related to the computational demands hypothesized to underlie hunting and gathering. Examples of manipulations of established tasks include variations on mental rotation (including measurements of the effects of angular distance), object array tasks, object recognition, categorical and coordinate sensitivity tasks, the judgement of line angle and position, and the tilt illusion. Examples of novel lower-level tasks designed to measure the skills underlying mental