As we know, when we analysis the factor variable in R, it always present n-1 factor variables, hiding one variables. When we use effect coding, and we want to know the coefficient, SD, SE of that one be hidden, what should we do?
I know that we can go back and change the reference level by re-code them, but is other way?
You can try removing the hidden intercept including -1 in your model:
fit1
Juan Pablo Edwards Molina ,thank you, but I don't understand what you mean? is species the hidden variable? But I have already do effect coding, I don't have this variable any more. What should I do ?
Thanks again for your answer!
By default R sets the first level of your categorical variable as the reference level. Then all the estimations are relative to it, not to a general mean. Adding - 1 to your right hand equation it will estimate directly each level factor and their statistics. It would be better to show some code of what you're trying to do..
Juan Pablo Edwards Molina Thank you very much! this is my code:
Mwanqidata2.mnl
as a detailed code might be useful to others find attached a kind of a tutorial on the subject, hoping it may help
Best.
Some modifications made, nothing meaningful.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R