Closed dill closed 8 years ago
Two issues: covariates and adjustment terms Fair enough; message could state: Model contains covariate term(s) no adjustment terms will be included.
k+a default I don't feel too strongly about this; I understand Tiago's point. At present, the behaviour of ds() mimics the behaviour of DisWin. That may be a legacy, but I'm content for the two pieces of software to conform on this particular matter.
Tiago says:
Should the message be less criptic if it just said somethin like
"Warning: when covariates are included adjustment terms are not fitted"
I find this a weird message, since in the code there's no adjustments, so the user would not necessarily be expecting these to be fitted any way?
If it was me, the default would be no ajdustments. You want adjustments you include them manually!
Have you thought about not having defaults for parameters "key" and "adjustments", this would force people to know what they are doing and avoid the need for these warnings