0
$\begingroup$

I am using sumcoding for my treatments in my design matrix $X$, as such I leave one treatment out and estimate these as follows X_missing = 0- sum(other treatments).

I am using confint to estimate the confidence interval of the fixed treatment effects, but since X_missing is calculated manually I can't really calculate the confidence interval of the missing treatment. Is there a way to work around this?

Improve this question
$\endgroup$

1 Answer 1

Reset to default
0
$\begingroup$

I've eventually chosen for the most simple solution: set a different treatment as a reference in my sumcoding.

Improve this answer
$\endgroup$

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.