1
$\begingroup$

I have two within-subjects factors A and B. Each participant undergoes all combinations of A and B. Now I am interested whether A or B or their combination can explain my DV (C) which is coded 0 and 1. So I create several models, say:
1) C ~ A * B + (1|Subject)
2) C ~ A + (1|Subject)
3)C ~ B + (1|Subject)
As 2) and 3) are not including all cases, do I have to aggregate for these models, or do I just include the same data as for model 1. In other words, do I have to create means across all levels of B for 2) and across all levels of A for 3)?

$\endgroup$

1 Answer 1

1
$\begingroup$

If your research question is whether A, B and/or their combination are associated with C, and all combinations were applied to each subject, it does not make sense to exclude either A or B from your model. An appropriate model is C ~ A * B + (1|Subject), and the research questions can be answered with reference to the parameter estimates for A, B and their interaction: A:B

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.