Chi-square test
Find sample size: If you know the two proportions, use this side to find how many subjects you need. [Enter decimals, not percentages: 0.25 for 25%]	Find effect size: If you know the proportion and number of subjects for Group 1, use this side to find detectable proportions for Group 2. [Enter decimal, not percentage: 0.25 for 25%]
Group 1 proportion: Group 2 proportion: Click here for sample size: You will need subjects in Group 1 You will need subjects in Group 2	N for Group 1: Group 1 proportion: Click here for effect size: You can detect Group 2 proportion p2 < You can detect Group 2 proportion p2 >
For different power or significance level, change the fields below: Alpha: Prob(reject H0 when H0 is true) 0.1 0.05 0.02 0.01 0.005 0.001 0.0005 0.0001 Power: Prob(reject H0 when H1 is true) 0.80 0.90
For unequal sample sizes, change the ratio below: Group2/Group1:
The calculations here use equations 4.18 & 4.19 in J.L. Fleiss, et al., Statistical Methods for Rates and Proportions, 4th ed., and are consistent with their Table A.4. The results here may differ from other webpages or software because most of them ignore the continuity correction. The Fleiss book discusses the importance of the continuity correction; see also Levin and Chen, American Statistician 53:62-66. Choose another study design v. 03/03 Contact: sfh2@columbia.edu