Chapter 11 - Effect of Sample Size on the Performance of Resource Selection Analyses

Abstract

Assessing resource selection is one of the main objectives of many wildlife radiotelemetry studies. Methods for determining resource selection vary from the widely used χ2 analysis for categorical data to complex statistical modeling using continuous data. This chapter conducts an evaluation with elk location data generated from an automated tracking system, which allowed collection of a substantially larger dataset than what could be collected with conventional methods of radiotelemetry. These data are used to estimate baseline patterns of resource selection for a population of elk and use resampling methods to simulate the effects of varying numbers of animals and varying numbers of locations per animal on the accuracy of resource selection estimates. The chapter calculates the percentage of correct conclusions (accuracy) for 1000 simulations for elk selection of six resource types (aspect, distance to open roads, distance to cover, distance to forage, % canopy closure, and vegetation) under varying levels of resampling for the two methods of resource selection. In general, the accuracy of resource selection increased with an increasing number of animals and an increasing number of observations per animal for all resource variables.