Confidence Intervals for Percentiles: An Application to Estimation of Potential Maximum Biomass of Trout in Wyoming Streams

Abstract

Until now, percentiles (other than the mean) of normally distributed data have seen restricted use due to the lack of an easy-to-use tool for estimating confidence intervals. Such a tool is developed in this article, using relevant statistical theory and taking advantage of modern statistical computing software. The development includes intervals for percentiles regression lines. An application to the estimation of maximum potential biomass of trout (Oncorhynchus spp., Salmo trutta, and Salvelinus fontinalis) in Wyoming streams is presented. The availability of easily computable confidence intervals for percentiles will allow wider use of these statistics.