Abstract
A number of statistics have been suggested for testing the difference between two Poisson expectations. On the basis of size and power calculations a new statistic, v= (2x+ 3/4)1/2 ‐ (2y + 3/4)1/2, where x and y are the two observed Poisson values, is recommended for testing such differences. This statistic is shown to be similar in performance to the familiar u=(x‐y)f(x+y)1/2, except in the tails of the distribution where it is superior.