|
|
| Message/Author |
|
|
| Kristen posted on Thursday, March 16, 2017 - 12:06 pm
|
|
|
| I want to perform MIMIC modeling using MLR estimation for my dissertation analyses. I am looking at the shared and unique variance between the 10 DSM Personality Disorders and the 4 facets of psychopathy and also examining the residual effects. My data are not normally distributed. Do you happen to know a way I may be able to still use Mplus with non-normal data? |
|
|
|
|
MLR gives robustness to non-normality for standard errors and chi-square. The point estimates are already robust to non-normality using ML and therefore also for MLR.
If you have strong floor or ceiling effects you may choose other models. If your data are better characterized as categorical you can specify that (Categorical=). |
|
|
| Back to top |
|
|
