Abstract In a series of experiments Dunbar (1993; 1995; 1998) has demonstrated that social group size is related to brain size in non-human animals. In two experiments, the present study aimed to extend findings by Dunbar and Spoors (1995) using their measure of social group size in humans. This measure postulates that there are three different closeness levels in social group size and that the levels scale up in a ratio of 3:1, level 1 being the smallest. Moreover, we aimed to place studies in this field in the context of individual differences. Experiment 1: Examining head size, general intelligence and personality, it was found that the three levels of social group size suggested by Dunbar (Zhou et al., 2005) are indeed distinct. However, they did not scale up in the ratios predicted. Dunbar’s measure of social group size was found to be related to head size and general intelligence, but not as consistently as reported in Dunbar’s studies (Dunbar & Spoors, 1995; Hill & Dunbar, 2003; Zhou et al., 2005). Moreover, social group size was found to be related to personality. Experiment 2: Adding social cognition as a variable, it was found that social group size correlates with head size through social and general cognition, but personality is also a significant factor in social group size. These findings led us to suggest that although the relationship between head size and social group size is mediated by cognition, as suggested by Dunbar (Hill & Dunbar, 2003), this does not provide conclusive evidence for the claim that the brain expanded ‘for’ social ability, as personality was also a major factor in social group size. We suggest, therefore, that the field should focus more on individual differences and, furthermore, that the field is in need of more nuanced studies.