1
$\begingroup$

I'm hoping to do a study to calculate whether 1 method of measurement is more variable than another in a subgroup of patients.

I assume this will be done with an F-test, from some reading?

However, I'm unsure as to how to do a power calculation for this? Would I have to synthesize data and then repeatedly run an F-test to see if it's powered? That seems very dirty, but I can't see an appropriate function in R...

$\endgroup$
1
  • 2
    $\begingroup$ The F test for variance is usually not recommended because it's very sensitive to the normality assumption. $\endgroup$
    – Glen_b
    Apr 16, 2015 at 15:37

1 Answer 1

1
$\begingroup$

Following links are likely to help you:

From: http://en.wikipedia.org/wiki/Levene%27s_test

In statistics, Levene's test is an inferential statistic used to assess the equality of variances for a variable calculated for two or more groups.

And this page shows how to calculate power and sample size for Levene's test: http://gradworks.umi.com/14/47/1447667.html

Others:

http://www.stat.purdue.edu/~zhanghao/STAT514/handout/chapter03/PowerSampleSize.pdf

http://www.webpages.uidaho.edu/~chrisw/stat507/sccpow1.pdf

$\endgroup$
1
  • 1
    $\begingroup$ Link-only answers tend, sooner or later, to degrade through "link rot." At a minimum, please summarize the contents of your links so that when they disappear or change, your answer will still be useful. $\endgroup$
    – whuber
    Apr 17, 2015 at 17:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.