Timeline for ML vs WLSMV: which is better for categorical data and why?
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 3, 2016 at 1:49 | comment | added | Aleksandr Blekh | @AdamO: Frankly, I'm not too sure at this time. Sorry, I have significantly deviated from SEM since March 2014. :-) | |
| Nov 2, 2016 at 19:40 | comment | added | AdamO | To say that this estimation procedure does not require normally distributed errors would make sense when modeling categorical/ordinal outcomes as continuous variables, which would make sense if the sole interest was estimating a mean difference. If data are binary, their mean is a proportion and the resulting model estimates proportion differences. However, other probability models do not require normal error assumptions, like a logit or quasilogit model. Are you saying that SEM will not use a mean variance relationship to improve estimation with categorical outcomes? | |
| Mar 18, 2014 at 23:16 | comment | added | Aleksandr Blekh | You may also find the following paper helpful: "Structural equation modeling in practice: A review and recommended two-step approach" (Anderson & Gerbin, 1988). While it doesn't refer to WLSMV, it contains discussion on various estimators. | |
| Mar 18, 2014 at 23:07 | vote | accept | Xander | ||
| Mar 18, 2014 at 23:07 | history | edited | Aleksandr Blekh | CC BY-SA 3.0 |
added 26 characters in body
|
| Mar 18, 2014 at 23:01 | history | answered | Aleksandr Blekh | CC BY-SA 3.0 |