This paper presents some elementary applications of Bayesian statistics to problems faced by wildlife biologists. Bayesian confidence limits for frequency of occurrence are shown to be generally superior to classical confidence limits. Population density can be estimated from frequency data if the species is sparsely distributed relative to the size of the sample plot. For other situations, limits are developed based on the normal distribution and prior knowledge that the density is non-negative, which insures that the lower confidence limit is non-negative. Conditions are described under which Bayesian confidence limits are superior to those calculated with classical methods; examples are also given on how prior knowledge of the density can be used to sharpen inferences drawn from a new sample.
|Title||Some Bayesian statistical techniques useful in estimating frequency and density|
|Authors||D. H. Johnson|
|Publication Subtype||Federal Government Series|
|Series Title||Special Scientific Report - Wildlife|
|Record Source||USGS Publications Warehouse|
|USGS Organization||Northern Prairie Wildlife Research Center|