robust estimators in SEM module
Hi,
I want to use the SEM module in JASP (specifically, the mediation analysis) and since it's based on lavaan, I was wondering whether the option of estimation (e.g. ML) with robust standard errors is available or is actually the same as the ML option?
If not, is it possible to use the bootstrap option to account for non-normal data?
Thank you for your work and all the best,
Katharina
Comments
I'll ask our expert!
E.J.
Thank you!
Hi Katharina,
yes, it's possible to get robust standard errors, under options -> confidence intervals / method -> robust. There the bootstrap option is also available!
Erik-Jan
Dear all,
Can you please confirm that if I use the settings (as in the photo attached) I am accounting for the fact that my data is non-normal?
Because I must use a robust estimator. With this, the ML turns into MLR, for instance?
Many thanks in advance!