Analyzing Selection Experiments with Log-Linear Models

Manly's measure of selectivity is a particularly useful measure of resource selectivity by resource users because it is often simply related to biologically meaningful parameters in resource use models. This papers describes how Manly's measure is naturally suited for log—linear model analysis. If resources users can be cross—classified by factors such as age and sex, log—linear models provides a convenient method for examining how selectivities depend on these factors. The techniques are applicable when resource usages occur independently and do not affect the resource availabilities.