We study social learning in a social network setting where agents receive independent noisy signals about the truth. Agents naïvely update beliefs by repeatedly taking weighted averages of neighbors' opinions. The weights are fixed in the sense of representing average frequency and intensity of...