"Matthias" <[EMAIL PROTECTED]> wrote in message
news:<a4heq0$ki3$[EMAIL PROTECTED]>...
> Hello,
>
> would be nice if someone can give me some advice with regard to the
> following problem:
>
> I would like to compare the means of two independent numerical sets of data
> whether they are significantly different from each other or not. One of the
> two underlying assumption to calculate the T-Test is not given (Variances
> are assumed to be NOT equally distributed; but data is normally
> distributed). What kind of (non?)parametric-test does exist - instead of the
> T-Test - to calculate possible differences in the two means?
> I'm using SPSS for further calculations.
>
> Thank you for your time and help,
>
> Matthias
two comments: one, you do not say how different are the variances.
the test statistic is robust to some violation of this assumption. i
do not know what the output looks like for SPSS, but in SAS the output
includes the p value for both the equal variance and unequal variance
calculation. you will see that they are typically close unless the
variances are "really" different. what is really different depends on
the shapes of the distributions, skewness, etc. it sounds as if your
data are pretty symmetric, so i bet you can tolerate variances that
differ by a factor of 3.
secondly, if you really want a nonparametric alternative, check out
the wilcoxon rank sum test.
JJD
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at
http://jse.stat.ncsu.edu/
=================================================================
