R - SEM - Hierarchical Confirmatory Factor Analysis Class Assignment

preview_player
Показать описание
Recorded: Summer 2015
Lecturer: Dr. Erin M. Buchanan

Packages needed: lavaan, semPlot
Class assignment for structural equation modeling. Topics covered include hierarchical or second order confirmatory factor analysis (CFA), bi-factor models, fit indices, and loadings.

Used in the following courses: Structural Equation Modeling
Рекомендации по теме
Комментарии
Автор

Hello Erin, I just wanted to say THANK YOU for sharing your knowledge. I was looking for this model for a month and finally I could undertood it. Best regards.

gonzalomarchant
Автор

this was very helpful, thanks! I have a question: in the example in the book (9.1.1), he said "I need constraints for the bi-factor model (lines 4-6). Because the LVs are un-correlated with each other, the Gf, Gsm, and Gs LVs are empirically under identified"

line 4:6
gf =~ a*matrix + a*picture.concepts
gsm =~ b*digit + b*letter
gs =~ c*coding + c*symbol

can you explain what he's doing there? and why you didn't include those parameters in your model specification? thanks!

donnyz
Автор

Hey, Erin! I was doing a CFA and the same thing happened to me. The higher-order model had the exact same fit as the correlated model. As in your example, the factors I estimated had high correlations prior to the higher-order model solution. I searched your video specifically looking for an answer to this. Have you worked out what's happening in these cases?

GabrielRodrigues-qgbn
Автор

Dr. Buchanan, I was wondering if you have a video that can explain identification issues. I have 6 DSM diagnoses that I want to load onto 2 lower order factors and 1 bifactor but I keep getting issues w/ identification. Thank you so much for sharing your excellent lectures with us!

perpetualconsideration
Автор

Hello! What if my bifactor.fit still generate the warning despite the fact of standarization? I've created the two CFA in jamovi to compare first and second model but in R I have the specific warning. Because the model is not good enough? How I can check it? :)

krystianmacheta
Автор

This video is very good. Thank you. However I am still not sure how to interpret my models. Interpretation of the indices part was very fast. I didn't get what are your cut offs for index values also didn't get where exactly you look at in summaries.

komakino
Автор

It is really helpful, thank you so much. Furthermore, i couldnt get how can built third order cfa in R. Do you suggest any resources? For instance, after second order model, i also want to add one additional layer which depression and anxiety for one factor, stress and other fourth one (imaginary) is for another factor. should i combine first level questions under these third order equations? I am really confused.

nevcihant
Автор

Hi, did you posted any video with hierarchical categorical data? I am analyzing interview in likert scale.

sofiaguardado
Автор

Hi, what book is being used, please?

julianeventurelli
Автор

Hello,
I am analyzing a high order model with only two factors. I cannot run the analysis, because of the number of factor. I was wonderinf if you know which type of constrains are necessary to modify the model?
I will appreciate any information.
Diana Alvarez

dianaalvarez