I have research entitled "Teacher well-being". In my first study, the study participants were very small and removed almost half of the items. However, after discussing with my Professor and reading interesting works sample size has its own contribution either fit of the model or unfit of the model. As result, now I have collected the large data and checked all the items using bifactor, Higherorder factor, correlated factor, and single-factor models. The single factor model shows the poor fit. Is there any source related to this supporting my output or any advice?
I am waiting for your feedback!
Sincerely,
Girum
CFA verifies whether a model is right while EFA tries to describe data. CFA adopts a much stricter approach. A two-factor model always yields a better fit than a one-factor model. You could try a maximum likelihood (ML) extraction in SPSS. For details on ML properties, you might go through chapter three of the following book.
Brown, T. A. (2015). Confirmatory factor analysis for applied research (2nd ed.). Guilford Publications. https://www.guilford.com/books/Confirmatory-Factor-Analysis-for-Applied-Research/Timothy-Brown/9781462515363
Good luck,
Dear Mohialdeen Alotumi,
Thank you very much for your valuable advice and interesting book.
You should evaluate whether the two-factor model is a significantly better fit than the one factor model by comparing the chi-square values for the models.
In addition, there are the various goodness of fit indices, which you can use to examine whether they are "adequate" for your one-factor versus your two-factor model.
Dear professor Morgan,
Really, I thank you for your interesting advice!
- Badges
- Science topic
- Similar topics
- Social Science
- Media
- Journalism