A female advantage is observed in the accurate recognition of mental and emotional states from the face (Hall, 1978, 1984). The psychological mechanisms that may underlie this advantage have not been addressed systematically by researchers. The present thesis discusses the potential mechanisms underlying the female advantage by considering the Extreme Male Brain (EMB) hypothesis of autism (Baron-Cohen, 2002). Several possible directions for research are presented, including sex differences in i) automaticity of processing facial expressions; ii) attention to the eyes; iii) configural versus featural processing of faces; and iv) stimulation of emotion. The first three of these directions are addressed in experimental chapters. A priming task and emotional face-word Stroop task were used to investigate sex differences in the automaticity of processing facial expressions. Sex differences in attention to the eyes were investigated in two eye tracking studies, and in two studies manipulating the eye region of emotional faces. Finally, a study with spatial frequency filtered emotional faces examined sex differences in the use of fine and coarse detail facial information. Overall, the investigations provide some evidence for greater female attention to the eye region in faces, and the possibility that this may explain an observed female advantage in facial expression recognition. Results are discussed in relation to the EMB hypothesis and sex differences in social cognition more generally. Potential directions for further research are outlined.