作者: F. Yates , P. M. Grundy
DOI: 10.1111/J.2517-6161.1953.TB00140.X
关键词: Statistics 、 Bias of an estimator 、 Random variate 、 Mathematics 、 Sample (statistics) 、 Error variance 、 Selection (genetic algorithm) 、 Horvitz–Thompson estimator 、 Estimation 、 Weighting
摘要: In selection with probability proportional to size x from within strata without replacement, the usual method of gives rise bias in estimate total a variate y derived by weighting units weights 1/x. By means numerical examples it is shown that amount this usually quite trivial. If, however, unbiased estimates are required, true probabilities different can be calculated easily for samples 2, and considerably more labour 3. The ordinary formula estimation error also investigated, reasonably accurate. Horvitz Thompson have given an estimator variance, but inefficient new given. A revising measures so original 2. Thompson's solution problem does not appear give satisfactory approximations cases met practice. successive members sample arbitrary sets chosen solely shall measures, which has been advocated various quarters, criticized.