Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I need to calculate the sample size required for an observational study in which incidence of disease is 19-29%. Population affected is 600,000 people.

The study has two samples similar on baseline characteristics treated with two different drugs.

Statistical analysis will be chi-square and Fisher's test. I have to demonstrate non-inferiority of one drugs with respect of the other one. I don't have any other info.

Can anyone help me to calculate the sample sizes required in order to obtain power of 0.80 with an alpha 0.05?

Thanks a lot in advance

share|improve this question
Surely one of the applets at will solve your problem. – whuber Dec 14 '10 at 18:35
@Roland How could you compute a sample size in a power study from a non-inferiority trial if you don't know the expected difference (or equivalently, the % of patients responding to the new treatment compared to the actual one)? – chl Dec 15 '10 at 20:35

G*Power is a commonly-recommended program for sample size calculations. I've only dabbled with it a couple of times in the past, but it's more than capable of handling the situation you describe.

share|improve this answer
(+1), excellent program. To spare the reader an extra click, you could link to the new version (G*Power3) which has a GUI for Windows and MacOS X: – caracal Dec 14 '10 at 20:50
Oops - thanks for that, caracal. I just grabbed the link without noticing the content of the page. – Matt Parker Dec 15 '10 at 2:34

Are all your participants infected? If you also have a placebo group and your df= 3 then according to Cohen (1992) at power = .8 and alpha = 0.5, for a chi square analysis will need 40 (39) participants if your effect is strong, 107 if it is a medium effect zise and 964 if its only a small effect. For only 2 groups the sizes are 26;87;725 for strong medium and small effect sizess respectively. see: for the article

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.