topic Re: Power Calculation and Number of Subjects Needed for a Study in Statistical Procedures

Power Calculation and Number of Subjects Needed for a Study

Yughaber — Tue, 09 Nov 2021 17:37:21 GMT

Hello,

I am in great need of some guidance towards power calculation. I need to calculate the number of minimum subjects needed for a specific research study and the power of that study. I looked into proc power but am still a bit unsure of the whole thing and would appreciate some insight.

The research is pretty simple and basic. I have two groups undergoing the same surgery but patients in each group are being treated with a different drug (drug A and drug B respectively for each group), I would like to see if drug A is better than drug B. I am using a t-test for this comparison of means.

Thanks!

Re: Power Calculation and Number of Subjects Needed for a Study

Reeza — Tue, 09 Nov 2021 17:43:39 GMT

Looks like two sample means, possibly one sided?
https://support.sas.com/resources/papers/proceedings15/SAS1911-2015.pdf

Re: Power Calculation and Number of Subjects Needed for a Study

SteveDenham — Tue, 16 Nov 2021 14:32:12 GMT

One thing to consider is the distribution of your response variable - continuous, count, proportion - as all need to handled a bit differently in PROC POWER.

SteveDenham

Re: Power Calculation and Number of Subjects Needed for a Study

Yughaber — Wed, 17 Nov 2021 05:33:07 GMT

is there maybe a reference/resource I can learn from for this specific point?

Re: Power Calculation and Number of Subjects Needed for a Study

SteveDenham — Tue, 23 Nov 2021 17:43:14 GMT

The PROC POWER documentation is an excellent place to start for power and sample size calculations. For distributions, it is hard to beat Merran Evans Statistical Distributions.

SteveDenham