Dear list,

I have two factors each with 3 levels.

Say these are A (a1,a2,a3) and B (b1,b2,b3).

I am interested to know whether a3b3 is different from all other combinations of factor levels.

I tried, based on what I reviewed in this list and other places, to set some contrasts using /TEST in SPSS.

For example, the difference between a3b3 and a1b1 can be written as:

mixed IS by A B

/fixed = A B A*B

/TEST = 'mean diff' intercept 0 A 0 -1 1 B 0 0 0 A*B 0 0 0 0 -1 0 0 1 0.

I've estimated the differences between a3b3 and all other combinations of levels.

However, I am interested in knowing whether a3b3 differs from all other combinations of levels combined. How can I do this in SPSS? Do I sum or average over all the other contrasts and test the difference between a3b3 and this sum/average? I tried this but it doesn't seem to work.

Should I be doing something different?

Any thoughts greatly appreciated.

Andra

=====================
To manage your subscription to SPSSX-L, send a message to

[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD