Using a multinomial logit model, the discrete-choice model for human behavior is developed for the analysis of resource selection by animals. This model has the advantage that resource availability can vary between animals and over time. Furthermore, the characteristics of the animals under study can be incorporated into the model. Methodology for estimating the multinomial logit model from a sample of alternative units is presented.

Key Words
Habitat selection; Independence of irrelevant alternatives; Multinomial logit models.

Marti L. McCracken is Research Associate, Mathematical Institute, University of St. Andrews, St. Andrews, Fife, KY16 9SS, U.K. (E-mail: marti@dcs.st-andrews.ac.uk). Bryan F. J. Manly is Professor of Statistics, Department of Mathematics and Statistics, University of Otago, P.O. Box 56, Dunedin, New Zealand (E-mail: bmanly@maths.otago.ac.nz). Madeleine Vander Heyden is Faculty Research Assistant in Wildlife Science, Oregon Cooperative Wildlife Research Unit, 104 Nash Hall, Oregon State University, Corvallis, OR 97331 (E-mail: Madeline.VanderHeyden@orst.edu).