1
$\begingroup$

Following my question here, I am also looking at the difference between males and females and I have conducted linear regression in a general linear model setup for this purpose.

My effect size for this part of the project is $\eta^2$ (which is 0.25).

$\eta^2$ is equivalent to $R^2$ (which is the appeal for me in this case).

I am wondering how do I calculate the confidence interval of $\eta^2$. Does anyone know of any online calculators that can I can use?

$\endgroup$
  • $\begingroup$ what does it mean, "$\eta^2$ is equivalent to $R^2$" ? $\endgroup$ – Stéphane Laurent Nov 27 '14 at 9:55
7
$\begingroup$

I could well be wrong about this, but...I think I would take the square root, yielding eta, which is akin to multiple R. Then I would use the standard method for finding a confidence interval for a correlation, which involves transforming to a Fisher's Z, obtaining upper and lower confidence limits based on Z and N, and transforming back. Then I'd square these limits to get my upper and lower limits for eta-squared. (One online calculator that could help is Vassar's.)

$\endgroup$
  • $\begingroup$ This is really quite a clever solution, +1. I went looking for my meta-analysis book this morning, but I couldn't find it... $\endgroup$ – gung May 11 '12 at 21:16
  • $\begingroup$ And intuitive too! $\endgroup$ – Amarald May 11 '12 at 21:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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