Abstract: The general decline of fertility in the Western world is still not completely understood. The use of social networks to explain the demographic transition has provided new insights on the question. Interpersonal relations can explain the diffusion of attitudes and behaviors throughout populations. This approach however raises the issue of agency and networks: where do networks come from, and how do they evolve in time? This dissertation examines the theory of social networks in relation to the demographic transition. It proposes the view that networks are significantly driven by the agency of Egos; it proposes to a lesser extent that biology and evolutionary traits affect the formation and evolution of social networks. Using data from the Malawi Diffusion and Ideational Change Project, it shows that the evolution of networks over time is driven in some part by the actions and desires of the Ego. Some patterns of evolution also conform to the predictions of the evolutionary hypothesis, even if they cannot be considered strong enough to confirm it. This dissertation therefore concludes that social networks do play an important role in the transition from high to low fertility, but that the diffusion of attitudes and behaviors is not a passive process. To the contrary, agents will construct networks and use them to achieve certain fertility goals. Ultimately, public policies implemented to affect fertility through diffusion should account for, and make use of, these patterns of network evolution.