 |
 |
Dear SPMers, Is my understanding correct that the `Flexible Factorial Design' is based on a full model? Then the results should depend on the factors specified in the design irrespective of the specified main effects and interactions as these only influence the appearance of the design matrix. Surprisingly I got slightly different results (df and t values) for a certain contrast when changing the specification of main effects / interactions although keeping the factor inputs constant. What could cause the difference in dfs? Shouldn't the results be identical assuming a full model? I'd be very grateful for any help. Many thanks in advance. Mira