Division of Epidemiology and Biostatistics, Department of Internal Medicine, University of New Mexico, Albuquerque, NM 87131, USA. e-mail: EBedrick@salud.unm.edu
MODEL SELECTION CRITERIA FOR LOGLINEAR MODELS
Article first published online: 7 DEC 2010
© 2010 Australian Statistical Publishing Association Inc.
Australian & New Zealand Journal of Statistics
Volume 52, Issue 4, pages 439–449, December 2010
How to Cite
Bedrick, . E. J. and Crandall, . W. K. (2010), MODEL SELECTION CRITERIA FOR LOGLINEAR MODELS. Australian & New Zealand Journal of Statistics, 52: 439–449. doi: 10.1111/j.1467-842X.2010.00593.x
Acknowledgments. We thank the editorial staff and referees for many suggestions that improved the manuscript.
- Issue published online: 27 DEC 2010
- Article first published online: 7 DEC 2010
- contingency table;
- multinomial model;
- Poisson counts
Considerable work has been devoted to developing model selection criteria for normal theory regression models. Less attention has been paid to discrete data. We develop two loglinear model selection criteria for Poisson counts. These criteria are based on an estimated bias adjustment of the Akaike information criterion. We observe in a simulation study that the corrected statistics provide good model choices and relatively accurate estimates of the mean structure.