Conceptual example of confirmatory factor analysis

[@fzj1280]

Awesome video! You save my Phd study. Thank you Mikko!

[@mronkko]

Glad to hear that!

[@NathalieTamayo]

Your videos are super useful. They focus on the practical aspect of the topics. Thank you for sharing this information

[@mronkko]

You are welcome!

[@edgarchang2511]

Great video series

[@mronkko]

You are welcome. I assume you mean the factor analysis playlist?

[@edgarchang2511]

@@mronkko Yes professor

[@cracrul]

Professor, thanks for the explanation. 1.- Allowing two errors of different constructs to be correlated would not indicate carelessness in the wording of the items? 2.- I have frequently seen correlated errors within the same construct, I think it is easier to justify. 3.-What happens if the correlation between the constructs is very high (>.85)? 4.- Is there any difference using polychoric or Pearson correlation in CFA?

[@mronkko]

1) Items can correlate for many reason and carelessness could be one 2) Yes. 3) See Rönkkö, M., & Cho, E. (2022). An updated guideline for assessing discriminant validity. Organizational Research Methods, 25(1), 6-14. doi.org/10.1177/1094428120968614 4) I answered this in another comment.

[@cracrul]

@@mronkko Thank you Professor. In his experience as an author and reviewer of articles that I recommend: 1.- In the case of CFA, is it necessary to evaluate and report the inferential multivariate normality assumption (Marida coefficient or any other)? 2.- In Social Sciences, non-compliance with multivariate normality and the use of ordinal scales (Likert with 5 options) are very frequent and it is customary, for CFA, to use the WLSMV estimation method (Weighted Least Square Mean and Variance Adjusted) by Muthén, du Toit, and Spisic, 1997. What do you think, use ML or WLSMV. 3.-Regarding the sample size, Kline and other authors indicate that there is no equation to calculate the sample size in CFA and SEM. If it is true, what recommendations can you give me to justify my sample size? Thank you!