|
![](http://www.statmodel.com/discussion/icons/csc/1px.gif) |
Internal rescoring of ordinal data |
![Previous](http://www.statmodel.com/discussion/icons/csc/navgen_arrow_left.gif) ![](http://www.statmodel.com/discussion/icons/csc/1px.gif)
|
Message/Author |
|
Jan Ivanouw posted on Wednesday, April 26, 2017 - 4:05 am
|
|
|
As I understand it, Mplus internally rescores for example an ordinal variable with 1-5 scoring into 0-4 (1=0, 2=1, 3=2, 4=3, 5=4). However, if for instance I have 3 variables u1-u3, and no person in my data set indicates the value 1 for variable u2, then u2 is rescored 2=0, 3=1, 4=2, 5=3. This gives errors when using the set u1-u3 in models. Is there a way to secure that even if u2 has no person giving answer 1, u2 still gets internally rescored: 2=1, 3=2, 4=3, 5=4 ? Best Jan Ivanouw |
|
|
See the CATEGORICAL option in the user's guide. There are different settings that can be used for maximum likelihood estimation. For weighted least squares, you may need to collapse categories. |
|
Back to top |
|
|