Sample size considerations

[Byrd Sellers, Johnita]

Hello everyone,

 

I have one last question regarding sample size calculations that does not specifically have to do with SPSS.  I have a non-randomized, prospective study where we will be comparing treatment within four different facilities.  Is there any literature that deals with how best to go about sample size calculations for such a study?  Any assistance that you can give regarding this matter will be greatly appreciated.

 

Thanks in advance,

Johnita

[Maguin, Eugene]

Johnita,

 

To the earlier post.  You’ve decided you want to have a 25% reduction in each facility. Different facilities probably have different rates so, yes, different sample sizes. But, maybe the facility rates are quite similar.

 

Yesterday’s message. I’ve been wondering about this and yesterday I was doing some power computation of my own and read the section on proportions in the PASS documentation. Probably, the usual situation is to compute the power for a comparison of two or more groups. Your situation is that you want to test one group against a population value. And repeat that four times, one for each facility.

 

Gene Maguin

 

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Byrd Sellers, Johnita
Sent: Tuesday, December 06, 2011 6:26 PM
To: [hidden email]
Subject: Sample size considerations

 

Hello everyone,

 

I have one last question regarding sample size calculations that does not specifically have to do with SPSS.  I have a non-randomized, prospective study where we will be comparing treatment within four different facilities.  Is there any literature that deals with how best to go about sample size calculations for such a study?  Any assistance that you can give regarding this matter will be greatly appreciated.

 

Thanks in advance,

Johnita

 

And, a few days ago you posted this . . .

 

Thanks everyone for your responses.  I have taken everything that you have said into account.  After careful consideration, I would like to reduce EACH facility by 25%.  I’m actually measuring each facility separately and I’m looking to see if there is a 25% reduction from baseline to after treatment WITHIN each facility (separately).  For cases like this, I would assume that I would need to calculate a separate sample size for each facility.  If I do it this way, I will come up with different sample sizes for each facility.  Is it okay to have different sample sizes? Or do I need to look at each facility’s sample size and use the sample size of the location with the highest sample size?  In other words, is it okay to have unequal sample sizes?  If so, does anyone have any reference articles about this that I can read.