Tests for Multiple Proportions in a One-Way Design

Description:

Suppose researchers are designing a study to compare the infection rates for three bone marrow transplant treatments. They wish to determine the sample size needed to achieve 90% power to detect a general difference in proportions among the 3 treatments. The current treatment has an infection rate of about 31%. It is much easier to obtain patients under the current treatment, so there will be twice as many patients in the current treatment group, as compared to the other treatment groups.

The researchers wish to compare the required sample sizes for a variety of assumed proportions for each treatment. The proportion sets are entered as columns in the spreadsheet.

The sample sizes are given as the total count. The group sample sizes are obtained by multiplying the total count by the group allocation proportion.

The proportion sets, as specified in the input table, are listed below the sample size table.

The needed sample sizes for each proportion set may also be compared using the sample size chart. The X-axis is simply the “set number” of each proportion set.