Proportion Difference Power / Sample Size Calculation

(Revised 10/30/2009 - Also display results of uncorrected ("classical") calculation.)


This screen computes the sample size required to detect a difference between two proportions.

Note: Before using this page for the first time, make sure you read the JavaStat user interface guidelines for important information about interacting with JavaStat pages.

Significance Level (alpha):
(Usually 0.05)
Power (% chance of detecting):
(Usually 80)
Group 1 Population Proportion:
(Between 0.0 and 1.0)
Group 2 Population Proportion:
(Between 0.0 and 1.0)
Relative Sample Sizes Required (Group 2 / Group 1): (For equal samples, use 1.0)

Sample Size Required

Group 1

Group 2

Total

"Classical" Calculation:

With Continuity Correction:


Note: This page incorporates a continuity correction to the usual sample-size formula based on the normal approximation to the binomial distribution. This correction increases the sample size (for each group) by an amount approximately equal to 2/abs(p1-p2), where p1 and p2 are the population proportions for the two groups. For a good discussion of this, see: Statistical Methods for Rates and Proportions by Joseph L. Fleiss (2nd ed., 1981, John Wiley & Sons, NY), chapter 3. This web page produces values consistent with those in Table A3 of that book.


Return to the Interactive Statistics page or to the JCP Home Page

Send e-mail to John C. Pezzullo at