Statistical Resampling Methods in Biology: A Case Study of Beaver Dispersal Patterns

SYNOPTIC ABSTRACT

The performances of the bootstrap, generalized bootstrap and jackknife resampling methods are compared using beaver (Castor canadensis) dispersal distance data. The results show that female beavers disperse slightly farther from their natal families than males, but the difference is not significant. Compared with results from simple statistics, the jackknife gives unbiased estimates for the mean and standard error. The performances of the bootstrap and generalized bootstrap are similar for estimation of the mean, standard error, and 95% confidence interval. The bootstrap requires a slightly smaller sample size than the generalized bootstrap to achieve the same desired results. However, the generalized bootstrap appears to be more consistent in parameter estimation than is the bootstrap. For similar studies, we recommend 500 as the minimum number of bootstrap replications.

Key Words and Phrases:

• resampling statistics
• bootstrap
• generalized bootstrap
• jackknife
• dispersal pattern
• beaver
• Castor Canadensis