R help - Jun 2003 - equivalence test

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,
	is it possible to do an equivalence test on paired quantitative datas 
in R? Is there a way to calculate sample size for such tests?

I've tried to find some documentation on that subject but I was 
unsuccessfull. 
I'll be happy with any links on equivalence test. If such a test does'nt
exist in R, i'll do it manually if I find a method to do so.

Best regards

Blaise
- -- 
Le doute me ronge. Et si tout n'?tait qu'illusion ? Si rien
n'existait ?
Dans ce cas, j'aurais pay? ma moquette beaucoup trop cher. Woody Allen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQE++anfs13e5HkHBycRAgeDAKC896MUrcUmoWrMHdu7bv9a/VJgewCffGeu
dmvi6Wy4jenHAj0la4ePHdU=tdNF
-----END PGP SIGNATURE-----

>Hi,
>	is it possible to do an equivalence test on paired quantitative datas 
>in R? Is there a way to calculate sample size for such tests?
>
>I've tried to find some documentation on that subject but I was 
>unsuccessfull. 
>I'll be happy with any links on equivalence test. If such a test
does'nt
>exist in R, i'll do it manually if I find a method to do so.
>
>Best regards
>
>Blaise
The sample size for an equivalence trial with normally distributed
outcomes can be calculated by

n <- 2 * s^2 / delta^2 * (qnorm(alpha/2) + qnorm(beta/2))^2

s = standard deviation
delta = clinically important difference
alpha = Type I error
beta = Type II error


-- 
Ross Darnell