Applying a simple statistical test to compare two different particle counts

Often one wants to compare a count, x1, from one sample with a count, x2, from another sample, to test whether they came from populations having the same means. Ractliffe (1964) [Appl. Statist. 13, 84-86] and Shuie and Bain (1982) [Appl. Statist. 31, 130-134] explored the behavior of the test statistic μ =(x2-x1)/ √ (x2 + x1), which can be used for this purpose. We demonstrate how to apply this test statistic to experimental design and analysis in the situation where two samples of particles are to be counted and compared. To summarize: |μ| > 2 indicates (at the 95% confidence level) the variates x1 and x2 were not drawn from populations having the same means. © 1991.