topic Re: How to use proc power without mean or power? in Statistical Procedures

How to use proc power without mean or power?

saza — Thu, 27 Jan 2022 16:15:06 GMT

Hello,

I was given a scenario that does not state a power value nor a mean or population size. How can I use proc power to solve for this scenario and problem set?

A researcher would like to examine whether there is a relationship between recent influenza vaccination (this season: yes/no) and a history of having the flu in the past 3 years (yes/no). She would like to detect a minimum difference in proportions of .10, and she expects a proportion of 0.35 in the reference group. She wants to limit falsely rejecting the null to 1%, and falsely rejecting the alternative hypothesis to 90%. What is the minimum sample size per group, given these criteria?

1) To detect a minimum difference of 0.10 (i.e., between 0.35 and 0.45) the minimum sample size per group is _____

Re: How to use proc power without mean or power?

PaigeMiller — Thu, 27 Jan 2022 16:26:29 GMT

I pointed you to an example in your previous thread.

https://communities.sas.com/t5/Statistical-Procedures/Was-told-to-figure-out-a-SAS-question-without-a-dataset-using/m-p/792695#M38846

Then others also gave examples. And now you repeat the question.

Re: How to use proc power without mean or power?

saza — Thu, 27 Jan 2022 16:34:05 GMT

I was not able to open the link you posted however am looking over it now. I have tried a variation of codes and the one I think should have worked, states "multiple answers"

PROC POWER;
ONESAMPLEMEANS
ALPHA =.05
SIDES =2
TEST =T
NULLMEAN=.35 to .45
STDDEV =.
MEAN =.10
NTOTAL =.
POWER =.;
run;

Re: How to use proc power without mean or power?

ballardw — Thu, 27 Jan 2022 16:39:31 GMT

If the question is about sample size then you must provide a power. Proc Power will only estimate 1 missing property.

If you do not know a power then you can request analysis at multiple likely values for power and either pick one or show the range: Power = .7 .8 .9 for example. You will get 3 scenarios. Similar for estimated "means" involved. Literature

Re: How to use proc power without mean or power?

saza — Thu, 27 Jan 2022 17:09:40 GMT

You are correct. I cannot have more than one missing value. Based on how you answered my question yesterday, I believe a twosamplemeans test might not be the correct route to solving this particular problem