How to test for differences in quadratic means/variances?

[marika]

Hello

I have two groups of patients that underwent different surgical procedures. I would like to compare the groups to find differences between their quadratic means. The values that I am comparing are measurements between pre and post-op, and they can be negative or positive. If I use the arithmetic mean I will cancel out the magnitudes, so I want to use the quadratic means. I would also like to compare the variances between the groups, to see if there is more dispersion in the values in one group versus the other.

I am using SPSS v.20. If I was doing arithmetic means I think I would do t-tests and look at Levene's test for the variances. But I can't figure out how to do this analysis on the quadratic means. Do I need to do it by hand?

Thank you in advance!
Marika

[marika]

As a follow up, I am wondering whether I could simply square the differences, or use the absolute values, in order to get rid of the signs. And then use t-tests to compare the arithmetic means. Would either of those options be appropriate statistically?

[GretaGarbo]

But I can't figure out how to do this analysis on the quadratic means.

Neither can I. What is “quadratic means”?

[Dason]

I think it's the square root of the mean of the values squared.

[marika]

yes that is the correct definition of quadratic means (also called root mean square).
I couldn't figure out how to do the analysis, so I used absolute values. I don't know if this is statistically correct though...

[Dason]

So just to be clear if everybody in one group increases 4 units from pre to post and the other group decreases 4 units you want to consider that the exact same thing and not find any differences between the two groups?