Posted by B. Douglas Ward on March 26, 2001 at 16:21:24:
In Reply to: Re: ANOVA2 problem posted by jian on March 23, 2001 at 17:40:24:
Jian:
Suppose you execute 3dANOVA2 without the -bucket command, e.g.:
3dANOVA2\
-type 1 \
-alevels 2 \
-blevels 2 \
-dset 1 1 'mc_601KM_mech_LL@1+tlrc[0]' \
...
etc.
...
-ftr Trtmnt \
-fa Inten_effect \
-fb Hand_effect \
-fab IntenHand_effect \
-amean 1 Int_low \
-amean 2 Int_high \
-bmean 1 Hand_left \
-bmean 2 Hand_right \
-adiff 1 2 LowHigh \
-bdiff 1 2 Left_right
Then the output will consist of 10 separate 2-subbrick datasets. You can then
use program 3dbucket to combine selected subbricks into a bucket dataset of
your choice. For example:
3dbucket \
-prefix 3dANOVA2.results \
'Int_low+tlrc[0]' \
'Int_high+tlrc[0]' \
'Hand_left+tlrc[0]' \
'Hand_right+tlrc[0]' \
'LowHigh+tlrc[0]' \
'Left_right+tlrc[0]' \
'Trtmnt+tlrc[1]' \
'Inten_effect+tlrc[1]' \
'Hand_effect+tlrc[1]' \
'IntenHand_effect+tlrc[1]'
The above will create bucket dataset 3dANOVA2.results+tlrc having 10 subbricks.
The first 4 subbricks contain the factor level means; the next 2 subbricks
contain the differences in factor level means; the last 4 subbricks contain
the F-statistics for treatment, main effects, and interaction.
If the resulting bucket dataset is still too large, then you can assemble
the subbricks into several smaller bucket datasets.
Doug Ward