Minimum Sample Size Required to Achieve Power 1—β in Testing for the Significance of a Difference in Independent Proportions

Abstract

In testing for the significance of the difference in independent proportions, it is possible to estimate the size of n requited to reject H_o at a given α level at some specified level of β risk as long as one has some basis for estimating the experimental error and the magnitude of the experimental effect. This paper provides tabled values for estimating n under these conditions.