Bayesian Methods in Applied Anthropology and Human Behavioral Ecology.

Abstract

In this dissertation, I explore the use of Bayesian models and Hamiltonian Markov Chain Monte Carlo methods in the analysis of data relevant to applied anthropological and human behavioral ecological studies. The anthropological data in such studies are often incomplete, censored, or uncertain, limiting the usefulness of both the data itself and standard statistical tools. However, recent advances in Bayesian model specification and Hamiltonian Monte Carlo parameter estimation methods allow for integration over uncertainty due to missing, censored, and partially known data, and can thus dramatically improve the quality of analysis in pure and applied anthropological studies. The first chapter in this dissertation uses a multi-level Bayesian survival analysis model and a Gamma-Poisson Bayesian network to analyze spatial and temporal patterns in life-history trade-offs in the Agta. The second chapter uses a Bayesian Ornstein-Uhlenbeck process-based adaptive phylogenetic model to investigate the relationship between stratification and the evolutionary origins of female genital modification in Africa. The third chapter uses Bayesian methods to investigate the cultural evolutionary dynamics of the transmission and maintenance/attenuation of female genital modification in a case study of the African diaspora and indigenous populations of Colombia. Copies of dissertations may be obtained by addressing your request to ProQuest, 789 E. Eisenhower Parkway, P.O. Box 1346, Ann Arbor, MI 48106-1346. Telephone [phone]; e-mail: [email protected]