contrasts using Mixed

Hi programmers-

I've got a repeated measures design with 4 time points. I need to run a
MIXED analysis evaluating the specific comparisons of each time point to
time 1.  In other words, T1 vs. T2, T1 vs. T3, and T1 vs. T4.

I thought the MIXED was similar to GLM and would run the following:

MIXED outcome BY independent WITH ASSESSPT...
...
/CONTRAST (ASSESSPT) = SPECIAL (1 1 1 1
                                                           -1 1 0 0
                                                           -1 0 1 0
                                                           -1 0 0 1)


but it does not.  Can anyone suggest syntax that might address this?

Thank you!!

=====================
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

[Bruce Weaver]

Charla Lopez wrote

Hi programmers-

I've got a repeated measures design with 4 time points. I need to run a
MIXED analysis evaluating the specific comparisons of each time point to
time 1.  In other words, T1 vs. T2, T1 vs. T3, and T1 vs. T4.

I thought the MIXED was similar to GLM and would run the following:

MIXED outcome BY independent WITH ASSESSPT...
...
/CONTRAST (ASSESSPT) = SPECIAL (1 1 1 1
                                                           -1 1 0 0
                                                           -1 0 1 0
                                                           -1 0 0 1)


but it does not.  Can anyone suggest syntax that might address this?

Thank you!!

It sounds like you are including TIME as a single categorical variable with 4 distinct values, is that right?  Have you tried entering  3 indicator variables (or dummy variables, as some folks call them) instead?  I.e.,

compute T2 = (time EQ 2).
compute T3 = (time EQ 3).
compute T4 = (time EQ 4).
exe.

MIXED Y BY { categorical variables } WITH T2 T3 T4 ...

The coefficients for T2, T3, and T4 will give tests versus the omitted reference category, T1.

You can check that all is working as expected by looking at the overall model statistics for this model versus one that uses your single TIME variable--e.g., they should give the same deviance.