Action, Function, & Structure; Interpreting Network Effects on Behavior in Rural Malawi

A long series of ethnographic and sociological studies on kinship systems and information flows in developing societies has portrayed networks as varying structurally, serving multiple functions, and expressing themselves in different types of interaction. Little of this earlier work has informed empirical research in demography or development-related research. In stead, the latter operationalize social networks in relatively narrow ways, allowing for little overlap between multiple networks, and focusing on a subset of potential causal mechanisms. In an effort to pull the empirical literature closer to its qualitative forbearer, we use data from the Malawi Diffusion and Ideation Change Project to test how conversation networks and transfer networks overlap. We offer some predictions regarding how these overlapping networks might individually or jointly influence distinct outcome including ownership of livestock, planning innovative crops and HIV testing. Our sample of women from Malawi, interviewed in 3 rounds across a 6-year period, also enables us to question the inter-temporal stability of network effects. Our findings highlight: (a) how networks based on different actions appear nonetheless consistent with diverse behavioral outcomes; (b) how there is relatively little overlap between conversational and transfer networks; and (c) how there is considerable instability in temporal effects of conversational networks.