So I have some data from a couple of different experimental conditions.

I’m interested in both differences in the mean and variance in each condition.

So I can do some anovas to compare means, and levene’s test to compare variance.

However, is there any way of assessing the power of a test for homogeneity of variance (eg levene’s)?
The reason I ask is that in general variance looks similar across my conditions, but how do I know if I’ve got enough data to be sure the variance doesn’t change?

42

I’d ask 10 cats and see how many prefer it.

Am seeing a stats chap tomorrow morning. Will ask him.

He said it would be complicated :/

Sorry cant be more help!

Oh well, thanks for asking!