UC Irvine

Classical statistical inference procedures usually assume the independence of sample units. However, the assumption of independence is often unrealistic in cross‐cultural research because societies in neighboring or historically related regions tend to be duplicates of one another across a wide variety of traits that are spread by historical fission, diffusion, or migration of peoples. A recent generalization of the usual regression model explicitly allows for networks of interdependencies among sample units as part of the model specification. Here, two new estimation procedures for this network autocorrelation model are compared to previously employed maximum likelihood procedures, and to the usual regression procedures which ignore interdependence. The results of comparisons based on simulated autocorrelation data and the reanalyses of two previously published empirical studies indicate that both of the procedures proposed here compare very favorably with the maximum likelihood approach, and both are vastly superior to the usual regression procedures when there is moderate to high autocorrelation (i.e., interdependence). [Galton's Problem, cultural diffusion, networks, cultural evolution, statistical methodology] 1984 American Anthropological Association