Some Statistical Problems Associated with Biological Assessment- Implications and Solutions

One of the methods used to define biological resources is biological surveys. Results from biological surveys are often used to make comparisons between populations of organisms above and below discharges, or between reference sites and suspected impacted sites. One of the metrics from these biological surveys is count data that are based on replicate samples taken from sites or stations. These data present challenges for the analyst because they are often clumped, with zero values and potential outliers. In this paper we present some approaches that will increase the ability of the analyst to apply the appropriate models to these data. Access to the programs described are given along with contacts for help in using the compiled programs. KEY TERMS: biological assessments; statistics; clumped distributions; negative binomial