Abstract
Bayesian model averaging (BMA) is a powerful technique to address model selection uncertainty and recent computational advances have led to a proliferation of usage. BMA methods are of particular interest in environmental health risk assessment because of the high degree of uncertainty that typically arises in that context. In this article, we review a variety of approaches to conducting BMA and compare four implementations in a setting where there are a number of potential predictors. We then use these four methods to calculate risk assessment measures that account for the uncertainty involved in modeling environmental exposures. These methods are used to reexamine data from a study conducted by Walkowiak et al. (2001) to investigate the effects of maternal polychlorinated biphenyl exposure on cognitive development in early childhood. This case study reveals that different strategies for implementing BMA can yield varying risk assessment results. We conclude with some practical recommendations.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Bailer, A. J., R. B. Noble, and M. W. Wheeler. 2005. Model uncertainty and risk estimation for experimental studies of quantal responses. Risk Anal., 25, 291–299.
Budtz-Jorgensen, E., N. Keiding, and P. Grandjean. 2001. Benchmark Dose calculations from epidemiological data. Biometrics, 57, 698–706.
Caldwell, B. M., and R. H. Bradley. 1985. Home observation for measurement of the environment. New York, NY, Dorsey.
Carlin, B. P., and S. Chib 1995. Bayesian model choice via Markov chain Monte Carlo methods. J. R. Stat. Soc. Ser. B, 57, 473–484.
Clyde, M. 1999. Discussion of Bayesian model averaging: A tutorial by Hoeting JA, Madigan D, Raftery AE, Volinksy CT. Stat. Sci., 14, 382–417.
Clyde, M. 2000. Model uncertainty and health effects studies for particulate matter. Environmetrics, 11, 745–763.
Clyde, M. 2003. Model averaging. In Subjective and objective Bayesian statistics, (ed. J. Press, 320–333. New York, NY, John Wiley and Sons.
Clyde, M., and E. I. George. 2004. Model uncertainty. Stat. Sci., 19, 81–94.
Crump, K. S. 1984. A new method for determining allowable daily intakes. Fundam. Appl. Toxicol., 4, 854–871.
Crump, K. S. 1995. Calculation of benchmark doses from continuous data. Risk Anal., 15, 79–89.
Dellaportas, P., J. Forster, and I. Ntzoufras. 2002. Bayesian model and variable selection using MCMC. Stat. and Computing, 12, 27–36.
Draper, D. (1995). Assessment and propagation of model uncertainty. J. R. Stat. Soc., Ser. B, 57, 45–97.
Gelman, A., J. B. Carlin, H. S. Stern, and D. B. Rubin. 2004. Bayesian data analysis, 2nd ed. Boca Raton, FL, Chapman and Hall/CRC.
George, E. I., and R. E. McCulloch. 1993. Gibbs variable selection via Gibbs sampling. J. Am. Stat. Assoc., 88, 881–889.
Green, P. J. 1995. Reversible jump Markov chain Monte Carlo computation and Bayesian model determination. Biometrika, 82, 711–732.
Hoeting, J. A., D. Madigan, A. E. Raftery, and C. T. Volinsky. 1999. Bayesian model averaging: A tutorial. Stat. Sci., 14, 382–117.
Hoeting, J., A. E. Raftery, and D. Madigan. 1996. A method for simultaneous variable selection and outlier identification in linear regression. Comput. Stat. Data Anal., 22, 251–270.
Jacobson, J. L. and S. W. Jacobson. 1996. Intellectual impairment in children exposed to polychlorinated biphenyls in utero. N. Engl. J. Medi., 335, 783–789.
Kass, R., and A. Raftery. 1995. Bayes factors. J. Am. Stat. Assoc., 90, 773–795.
Longnecker, M. P., M. S. Wolff, B. C. Gladen, J. W. Brock, P. Grandjean, J. L. Jacobson, S. A. Korrick, W. J. Rogan, N. Weisglas-Kuperus, I. Hertz-Picciotto, P. Ayotte, P. Stewart, G. Winneke, M. J. Charles, S. W. Jacobson, E. Dewailly, E. R. Boersma, L. M. Altshul, B. Heinzow, J. J. Pagano, and A. A. Jensen. 2003. Comparison of polychlorinated biphenyl levels across studies of human neurodevelopment. Environ. Health Perspect., 111, 65–70.
Lunn, D. J., A. Thomas, N. Best, and D. Spiegelhalter. 2000. WinBUGS—A Bayesian modeling framework: Concepts, structure and extensibility. Stat. and Comput., 10, 325–337.
Morales, K. H., J. G. Ibrahim, C. Chen, and L. M. Ryan. 2006. Bayesian model averaging with applications to benchmark dose estimation for arsenic in drinking water. J. Am. Stat. Assoc., 101, 9–17.
Ntzoufras, I. 2002. Gibbs variable selection using BUGS. J. Stat. Software, 7, 1–19.
Raftery, A. E. 1995. Bayesian model selection in social research. Sociol. Methodol., 25, 111–163.
Raftery, A. E., D. Madigan, and J. Hoeting. 1997. Bayesian model averaging for linear regression models. J. Am. Stat. Assoc., 92, 179–191.
Ribas-Fitó, N., M. Sala, M. Kogevinas, and J. Sunyer. 2001. Polychlorinated biphenyls (PCBs) and neurological development in children: A systematic review. J. Epidemiol. Commun. Health, 55, 537–546.
Schwarz, G. 1978. Estimating the dimension of a model. Ann. Stat., 6, 461–464.
Vreugdenhil, H. J. I. 2003. Neurodevelopmental effects of prenatal exposure to environmental levels of PCBs and dioxins in children at school age. Rotterdam, The Netherlands, Optima Grafische Communicatie.
Walkowiak, J., J. A. Wiener, A. Fastabend, B. Heinzow, U. Kramer, E. Schmidt, H. J. Steingrueber, S. Wundram, and G. Winneke. 2001. Environmental exposure to polychlorinated biphenyls and quality of the home environment: Effects on psychodevelopment in early childhood. Lancet, 358, 1602–1607.
Wasserman, L. 2000. Bayesian model selection and model averaging. J. Math. Psychol., 44, 92–107.
Author information
Authors and Affiliations
Corresponding author
Additional information
Melissa Whitney’s research was supported in part by a grant provided by the National Institutes of Health (NIH 5T32ES007142 Graduate Training in Biostatistics).
Rights and permissions
About this article
Cite this article
Whitney, M., Ryan, L. & Walkowiak, J. On the Use of Bayesian Model Averaging for Covariate Selection in Epidemiological Modeling. J Stat Theory Pract 7, 233–247 (2013). https://doi.org/10.1080/15598608.2013.772037
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1080/15598608.2013.772037