|
![](http://www.statmodel.com/discussion/icons/csc/1px.gif) |
Use only one of the factors from a model |
![Previous](http://www.statmodel.com/discussion/icons/csc/navgen_arrow_left.gif) ![](http://www.statmodel.com/discussion/icons/csc/1px.gif)
|
Message/Author |
|
sunyforever posted on Thursday, November 07, 2013 - 8:56 pm
|
|
|
Dear Prof Muthen, Say if I have 4 indicators, x1, x2, x3, x4, which fit well in a CFA model: f1 by x1 x2; f2 by x3 x4; But I am only interested in one of the two factors, say, f1. I want to use the score of f1 in a regression model as an IV and leave f2 alone. Do you think this is a valid approach to use factor score? |
|
|
A factor with two indicators is not identified without borrowing information from other parts of the model. I would think the factor scores for f1 would have very low factor determinacy and would not see using them as good practice. |
|
Back to top |
|
|