Second order confirmatory factor analysis in Stata (Jan 2023; see links in video description)

In this video, I provide a conceptual introduction to second-order CFA (comparing it to the first order CFA model); I move to testing both first and second order CFA models using Stata and then comparing them in terms of their fit. I demonstrate the use of the chi-square difference test and information criteria for judging relative fit between the models.
A copy of the data can be downloaded here:
drive.google.c...
File containing path diagram for first-order factor model:
drive.google.c...
File containing path diagram for second-order factor model:
drive.google.c...
Do file used in demonstration of lrtest in video:
drive.google.c...
A supplemental powerpoint can be downloaded here:
drive.google.c...
Be sure to check out my online sites at:
www.drmcstats....
sites.google.c...

[@Phoebe87100]

Thank you very much. You have saved my life :D

[@adamssam6380]

I am a big fan of your video, Dr. Crowon and I have been recommending your video to my friends!! I sincerely appreviate your video. I am curious whether you could upload EFA, CFA, SEM videos using Mplus? I am a Mplus user and I need to learn how to perform those analyses in Mplus :)

[@gemechuhotw.7978]

I appreciate your video and have been recommending to my classmates and students. I'm very interested if you upload same video for large Likert scale data.

[@hadjijalotjot7439]

Hello Mike. Thanks for your videos. It has been very helpful especially for me who is new to CFA, SEM. BTW, I am running a model similar to this...several manifest variables then four first order latent variables then one 2nd order latent variable. I've been trying to fit the model for days now and I always have non-convergence problem. Any tips on this?

[@origingaming4248]

I am currently working on my dissertation, where in I am required to conduct cfa but my second order latent variables have their own observed variable ( questions directly related to the variable were asked , i dont know what kind of model is that.

[@will74lsn]

I have just tried the same with one of my data sets that has the exact same structure (second-order 2-factor model with 3 + 3 + 3 + 3 indicators) -> I get "convergence not achieved" "Warning: The LR test of model vs. saturated is not reported because the fitted model is not full rank. There appears to be 2 more fitted parameters than the data can support" 😞

[@ashenjoseph4114]

Hi, I am going through the same problem with my model. Did you find solution for what you were going through?

[@will74lsn]

@@ashenjoseph4114 no

[@will74lsn]

is the syntax available?