|
|
 |
| Multiple Regression with Imputed Data |
 
|
| Message/Author |
|
|
|
|
I am running a multigroup analysis using MLR estimation. When I received a negative F value using the SB chi-squared when unconstraining one of my paths, I utilized the Strictly Positive SB test to calculate c(10). The path that I tested needed to be freed, so now the model that includes the unconstrained path becomes the new nested model as I test additional paths for invariance. When moving forward with this model, do I use the scaling correction (c10) or return to the original scaling correction (c1) from this model to be used as the new c(0)?
Thank you! |
|
|
|
|
| I am sorry, I attempted to change the title of this post but it did not save. This is not a question about imputed data. |
|
|
|
|
| You should not use C10. Use C1 as usual. The C10 value is specific to the pair of models that you tested and it would not apply to another pair of models. |
|
|
| Back to top |
|
|
