Hi Everyone,

I am trying to specify a two-factor repeated-measures ANOVA. I've gone through the documentation and the UCLA help page, but I believe that I am mispecifying my error term. NB: I found the syntax Stata uses for this to be uncharacteristically awkward. I'm comparing my calculated model against a model calculated in SPSS that is known to be accurate. The data I am using come from Andy Field's most recent textbook on SPSS, and deal with a 3x3 repeated measures design. In other words, each person experiences each of the nine conditions exactly once.

Here is an example of the data:

Code:
       id   attitude   drink   imagery  
  1.    1          1       1         1  
  2.    1          6       1         2  
  3.    1          5       1         3  
  4.    1         38       2         1  
  5.    1         -5       2         2  
  6.    1          4       2         3  
  7.    1         10       3         1  
  8.    1        -14       3         2  
  9.    1         -2       3         3  
 10.    2         43       1         1  
 11.    2         30       1         2  
 12.    2          8       1         3  
 13.    2         20       2         1  
 14.    2        -12       2         2  
 15.    2          4       2         3  
 16.    2          9       3         1  
 17.    2        -10       3         2  
 18.    2        -13       3         3
id = participant id
attitude = dv
drink = factor 1
imagery = factor 2

Frustratingly, I can get all relevant F-statistics to match the "known model", but Stata is indicating that I've explained all variability in the model which shouldn't be the case.

This is how I'm specifying my model:

Code:
anova attitude drink id / drink#id imagery/ imagery#id drink#imagery/ drink#imagery#id, repeated(drink imagery)
Can anyone point me in the right direction to explain what exactly is happening?

Thanks a lot!

Cheers,

David.