deviation coding in anova with unbalanced n

Hi, if I have an oneway Anova with:
X Y
1 1
1 2
1 2
1 3
2 5
2 7
3 2
3 3
3 4

The mean of Y for each X are: 2, 6, 3
The pooled global mean is 3.222 and the non-pooled global mean is
(2+6+3)/3=3.667

If a do an Anova with deviation contrast, taking the last category as
reference, I obtain the estimates: -1.667 and 2.333, which correspond
to the differences between each mean category and the non pooled
global mean: -1.667=2-3.67

I would like to make an Anova with deviation contrast which gives the
differences between each category and the Pooled global mean, instead
of the non-pooled. SPSS, Stata, SAS work by default with the non-
pooled global mean, but there is the posibilility of directly writting
the contrast coefficient.
Does anyone knows what are the contrast coefficient to obtain the
estimates with respect to the pooled global mean?

Thank's in advance

J. Blas Navarro
Methodology Unit - Psychology
Autonomous University of Barcelona
Spain

.