Some of my factor loading comes below .5 in CFA. What to do? I am working with primary data. Data size 625. I am working with economic and household empowerment. Only 35% women are working out of total population. Can I proceed with .3 or .4 factor loading? One more question can I perform CFA, having variables with different categories?
Hello Shanjida,
please provide more information, according to https://www.researchgate.net/post/Netiquette_proposal_for_questions_on_structural_equation_modeling
All the best,
Holger
Hello Shanjida,
Holger is correct in noting that, with more information about your specific data set, model, and research aims, folks would be better able to offer focused recommendations.
However, I can offer these general comments:
That an estimated variable-factor loading is below .5 (whether in EFA or CFA) isn't necessarily a problem per se. I suspect that you're trying to follow someone's guideline as to what comprises "acceptable" loading strengths (and there are numerous such guidelines available, many depend on context and research aims). Do remember, these are guidelines and not commandments. Raymond Cattell, who contributed much to the literature on factor analysis, offered up one vignette in which a loading of .15 was considered salient for the model.
If the overall measurement model behaves well, and there are no glaring instances of model-data misfit, then why worry?
Of course, if your intent was to test whether the absolute value of loadings meet/exceed some a priori value, then the sheer values would be of concern, and would call for a different kind of test.
Good luck with your work.
You can use below 0.5....but not more than 2 or 3 values...but in some cases you can also use below 0.5...