Confirmatory factor analysis using Stata (March 2020)
ฝัง
- เผยแพร่เมื่อ 7 ก.ย. 2024
- This video provides a basic introduction to confirmatory factor analysis using the drawing program in Stata. I demonstrate how to draw out the model using the drawing icons, etc. I demonstrate options for estimation, and how to obtain standardized results. Finally, I demonstrate how to interpret various results. You can download a copy of the data here: drive.google.c...
Download the Powerpoint referenced in the video here: drive.google.c...
You can download a copy of the file used when drawing out the model here: drive.google.c...
This was very helpful. Thank you so much! I have watched many other videos of yours and found those straightforward and easy to follow.
Thank you for the insightful video. Is Cronbach's alpha not used anymore in SEM contexts to assess whether latent indicators make sense?
i have more than 100 of your vds download. Thanks
amazing video sir! You have helped me and my fiancee to prepare for exams :)
Very good explained, let me know...why you made covariance line between error terms, pls tell the reason n base on which we draw it between the errors?
Thank you very much. I am wondering how to get factor scores to use them for further analysis?
Please suggest how to calculate CR AVE and HTMT in STATA
Video very helpful. I however seek for guidance on why for my data i get initial values not feasible
Great video sir! But What if I have multiple types of dat (continuous and categorical) within a dataset?
Thank you very much for this overview. Is it also possible to perform a bootstrap during a CFA? Thanks in advance for your response!
Pls tell me....if we have many itmes under one factor or more factors, how we can make a CFA model?
Great video. It is really helpful and clear. I want to ask why did you use double headed arrow (finding covariance) for e3 and dep1 only? Why did you not use for the pair of other observed variables?
Hi there. Usually an error covariance in a model such as this is added based on some rationale; for example, you might do this based on the assumption that part of the correlation between two items is explained by a common method factor during measurement. Nevertheless, I included the covariance in this example simply to show viewers how to obtain it if desired. Cheers!
Super
Thanks for the video. Can you explain why you set some of the factor loadings to 1 ?
Hi there. One condition for your model to be identified is that the latent variables (and this will also go for the error variances) in your model to have a measurement scale. If one or more don't, then when you run your analysis you will find your parameters cannot be estimated (and you also get a message back typically that your model is 'not identified' or 'unidentified'; not sure the verbiage in Stata). There are two ways of establishing a measurement scale for latent variables (and error terms): fix the variance of the factor to 1 , or fix one of your factor loadings from your factor to an indicator to 1. If you do the latter, which is the more conventional approach, you are setting the measurement scale of the factor in reference to the indicator with the path coefficient fixed to 1. Again, all latent variables require a measurement scale, so that's why you see multiple paths in the model with 1's. The choice of which indicators should serve as reference indicators is up to you - although it is generally recommended that you choose to use indicators that are more reliable. But oftentimes programs and researchers end up picking the first item in a set to serve as an indicator of a factor (pretty arbitrary, right???). At any rate, you now know why the paths from the errors are pointed to the indicators and fixed to 1, as they are all setting the measurement scale for the errors. Hope this helps!
Great video. I am getting a 404 error when I try to access the powerpoint - can you repost the url? Thanks very much!
how to use SEM if my data is panel dataset? thank you
Would you explain it with spss?
Hi there. Unfortunately, SPSS does not allow you to perform confirmatory factor analysis. You can use the SPSS add-on AMOS to do it though (or use some other SEM program). I have a few older videos on using AMOS if you want to see those demonstrations:
th-cam.com/video/lleBuEEQXM4/w-d-xo.html
th-cam.com/video/sjzoOLOhNUk/w-d-xo.html
th-cam.com/video/OdBIJrGov_0/w-d-xo.html
th-cam.com/video/rOeFJKrGhlM/w-d-xo.html
Best wishes!
Can you please share stata code
Hi there, there is a link under the video description to a text document containing the code. I just refreshed the links as well.
I am really thankful to you for such a good lecture. It is very helpful in my PhD research. I might need some more help kindly send me your email for more inquiries.
Again thanks.
how can I estimate T values for SEM ?