Dear all,
I am currently looking for a paper which discusses this issue: I am comparing a mediation model between three groups. The outcome measure is ordinal, all other measures interval scaled measures. Is it advisable to use WLSMV in this case?
Thank you very much in advance!
Julian
Take a look at this paper I wrote about mediation models. It will give you some ideas for such analyses. The references are comprehensive, so if the text itself does not elaborate sufficiently, you'll find some references with greater detail
Article Using mediation analysis to identify causal mechanisms in di...
Thank you very much! I will take a lot at it during the course of the day.
Besides that suggestion you may have a look at the Mplus website: http://www.statmodel.com
On this website they have several articles explain the different types of estimator and when to use which one of them. Further, they have a section about mixed model, which might be applicable at your case. The users' guide is very helpful as well, especially if you don't want to get stuck with mathematical formulas.
Groups size are always a big discussion. That will depend on the type of model you will use. Further, minimal sample size for this type of analysis is around 200 subjects. You have to take into account that the size of your sample will influence the outcomes of your analysis, in particularly some goodness of fit.
Best,
Dear Dario, thank you for your helpful suggestion. Meanwhile, I re-analyzed the data with WLSM as the dependend outcome variable is defined as an ordinal one. This als led to changes regading the fit indices, however, they are still satisfactory. Especially when considering the suboptimal measurment properties of some scales that were used. Sample size is not an issue with the current data set, but thank you for your suggestion.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling