turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- Sample size calculation for three groups

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-12-2013 03:33 AM

Hello everybody,

I would like to perform a sample size calculation for a study in parallel design, 3 arms.

I do not use the GLMPOWER procedure but use the following code:

proc power ;

twosamplemeans

alpha=0.017

meandiff=mean difference

stddev=estimated stddev

power=.8

npergroup=.

;

run ;

In fact, I adjust the alpha level using a Bonferonni correction (0.05/3=0.017).

Is this approach a good one? If not, do you have any suggestions?

Thanks a lot in advance for your answer.

Best,

Accepted Solutions

Solution

12-12-2013
09:03 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-12-2013 09:03 AM

Take a look at Example 71.1 One-Way ANOVA in the POWER Procedure, or look at PROC GLMPOWER Example 44.1. The latter has an explicitly worked example for calculation of sample size.

Steve Denham

All Replies

Solution

12-12-2013
09:03 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-12-2013 09:03 AM

Take a look at Example 71.1 One-Way ANOVA in the POWER Procedure, or look at PROC GLMPOWER Example 44.1. The latter has an explicitly worked example for calculation of sample size.

Steve Denham

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-18-2013 07:38 AM

Thanks a lot Steve!

Do you think the approach I describe is also correct?

Best,

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-18-2013 01:34 PM

The Bonferroni correction is almost always too conservative, and consequently will lead to an increased sample size needed to detect a fixed difference with given power.

Steve Denham