Hi everyone,

as my computer crashed a few days ago, I lost the source to an article I was reading referring to how large the sample size deviation in repeated measures can be between treatment and control group. If I remember correctly it cited Stevens (might have been Student though) on how starting at a 0.5 divergence in sample sizes there would, among other issues, almost always be sphericity violations in place. As I'm dealing with a borderline case, I would really like to quote that one, but can't find it for the life of me.

So if anyone could help me out on that one => eternal gratitude. Thanks,
statcat