Working Paper: NBER ID: w20422
Authors: Abhijit Banerjee; Arun G. Chandrasekhar; Esther Duflo; Matthew O. Jackson
Abstract: Can we identify the members of a community who are best- placed to diffuse information simply by asking a random sample of individuals? We show that boundedly-rational individuals can, simply by tracking sources of gossip, identify those who are most central in a network according to "diffusion centrality," which nests other standard centrality measures. Testing this prediction with data from 35 Indian villages, we find that respondents accurately nominate those who are diffusion central (not just those with many friends). Moreover, these nominees are more central in the network than traditional village leaders and geographically central individuals.
Keywords: Social Networks; Information Diffusion; Centrality; Gossip
JEL Codes: D13; D85; L14; O12; Z13
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| individuals can accurately nominate those who are most central in a network based on gossip (D85) | nominees are more central according to diffusion centrality measures (D79) |
| nominations are significantly correlated with diffusion centrality even after controlling for leadership status and geographic position (D79) | individuals learn about the network's structure effectively (D85) |
| a one-standard deviation increase in diffusion centrality (R12) | predictive power of the model (C52) |
| individuals utilize gossip to identify central figures accurately (Z13) | nominations are not simply naming prominent figures (D72) |