Note on the inverse and direct methods of estimation in R D!Gordon's problem
成果类型：
Article
署名作者：
Pearson, ES
刊物名称：
BIOMETRIKA
ISSN/ISSBN：
0006-3444
发表日期：
1939
页码：
181186
关键词：
摘要：
In this paper it is pointed out that the reliability of the estimate of mean density of a bacterial population obtained by the dilution method may be expressed in a different way from that followed by R. D. Gordon in the preceding contribution. Many statisticians would probably prefer to obtain confidence or fiducial limits for the unknown density [rho], rather than to make an appeal to Bayes'' theorem, using the concept of inverse probability. It is pointed out how a preliminary solution on these lines was given in 1935 by Matuszewski, Neyman and Supinska who provided a chart from which limits for [rho] associated with a confidence coefficent of 0.95, could be read off. To use this chart, the maximum likelihood estimate of [rho] must be found; this may be obtained in certain cases from Halvorson and Ziegler''s 1933 tables.