Why did humans become as intelligent as they are? The Social Brain Hypothesis claims that general ability is largely a side-effect of the demands of human social group size, and that IQ is a principle contributor to effective social behaviour. The present study tests this hypothesis using a novel methodology: an extant compulsory residential network is examined in order to maintain the universality of structural confounds. Several possible surrogates of social complexity are proposed and their relationships with IQ, personality and empathy-systemizing behaviour are investigated. We find that IQ does predict social group size and complexity in males but not in females. Integration of the E-S (Empathizing-Systemizing) Theory framework revealed that ‘brain type’ is a significant factor in social network modelling. Different brain types involve different predictors of social complexity, lending support to the cognitive styles approach. We conclude that empathy and systemizing behaviour appear to be key in mediating social network formation and cohesion in humans. We conclude that the framework of the Social Brain Hypothesis is too general to be of real use to the study of individual differences in humans; it is essential to identify key brain regions involved in social network maintenance. People with different brain types approach the world in qualitatively different ways, utilising different cognitive processes in order to achieve the same goals. This account may imply multiple stable evolutionary strategies that were in competition during the evolution of the human brain.