Dear all,
I think I have a problem with the estimation of my flexible factorial model. The design matrix is mostly black (see attached screenshot). And it takes a lot of iterations to estimate the model.
How could I fix this?
I have a 2x2 design with the between-subject factor group (obese vs. lean) and the within-subject factor condition (two different conditions in a task). In the factorial design I model the main factor of subject, group & condition, as well as the interaction between condition and group. I am mostly interested in the interaction term, since my main hypothesis is that brain activity will differ among groups and conditions.
However, to me it seems that the model is too complex and cannot be estimated correctly. Could that be the case?
If so, would it be valid to only estimate the above-mentioned main effects and compute their interaction in the contrast manager afterwards? How would that be any different from the previous approach?
I actually tried this already and it worked, but I am not sure if this is a valid approach.
I double checked the code and the error should not be due to coding. Also the design matrix looks fine after the model specification (see 2nd attachment).
Thank so much and I am happy to give more details in case I missed relevant information.
Best,
Nadine
MSc Neuroscience
PhD student, Decision-making in obesity: neurobiology, behavior and plasticity
Max Planck Institute for Human Cognitive and Brain Sciences
Stephanstraße 1a
04103 Leipzig, Germany
Room No.: A 113
Phone: +49 341 9940 2410
|