Problem with using ML estimator and FIML

Augustin

Hi everyone,

I'm trying to run the analysis from an article (so run the same analysis) which implies a set of ordered data and use a multigroup SEM. I think that my questions are quite simple : is it impossible to use ML estimator and FIML for missing datas when we have ordered data ? I have a lot of items with scale responses like the answers are "Totally disagree, .. , Totally agree".

Because for your information, I receive this kind of error display when I try : " FIML is not supported for categorical data." or "Error in Model 1 - Estimation failed message: lavaan ERROR: estimator ML for ordered data is not supported yet, use WLSMV instead."

The problem is that in the analysis that I need to replicate, they use these estimators, but with another software than JASP. I saw somewhere that not all of software can use these estimators for ordered/categorical data.

What do you think about this ?

Thanks a lot by advance for you attention.

Augustin

patc3

No it's not possible to use FIML with ordered predictors in JASP (or in the R library lavaan, which JASP uses in the backend). Are you sure the article you're replicating treated them as ordered and not scale?

My best guess reading your post is that you're trying to treat your variables as scale but JASP says they're ordered. This is because you need to change the variable types in the data view (leftmost panel) by clicking on the color icon of the variable:

To be able to use FIML with missing data, your variables need to be of type scale in JASP (the orange ruler), not ordinal (the bar chart icon).

Augustin

Thanks for your answer. I am not sure but I think so. Is it possible to treat the data as scale even if the items of the variable are defined on an ordered scale like Likert's scale ? Because it's that kind of ordered scale that I have. All core items were measured on 7-point scales ranging from 1 = “Strongly disagree” to 7 = “Strongly agree .I mean is it methodologically valid to do it?  

This Monday I will also ask to the person who wrote the article.

Thanks by advance.

Augustin

patc3

Most people treat it as scale.. Actually the opposite (treating a Likert scale as ordered) is quite rare.

My guess is they treated it as scale, in most software a Likert scale is treated as scale by default (R, Mplus), but JASP distinguishes between scale and ordinal a bit more explicitly.

Augustin

Thanks a lot Patrick,

I changed the type of variable. I also sent you an e-mail because unfortunately I face other error messages.

Augustin

patc3

Actually you're not sending me emails, you're responding on the ticket you made on GitHub 🙂 https://github.com/jasp-stats/jasp-issues/issues/2476