|
|
 |
| Constrains to categorical variables |
 
|
| Message/Author |
|
|
| yvette xie posted on Saturday, June 06, 2015 - 8:42 pm
|
|
|
Dear Professor,
I have a factor which only contains one categorical observed variable. In order to solve the identification issue, I know I can add constrains to the residual of that variable according to the reliability (e.g. 0.8) and variance if the variable is continuous. Can I use the same method to deal with categorical observed variable?
Thank you in advance! |
|
|
|
|
The residual variance of a categorical variable is not a model parameter. To put a factor behind a categorical variable, say:
f BY u@1;
You might as well use u in the model instead of f because f is equal to u. |
|
|
| Back to top |
|
|
