I am using SEM for my studies. My questionnaire involve several adapted questionnaire and one newly established questionnaire. I am sure to run EFA after Pilot study. Do I still run CFA using SPSS or is it already within the SMART PLS I will utilise later?
Hi Gira,
the most important thing ist to consider wether a factor model (which the CFA applies) is at least theoretically plausible. "Scales" are often heterogeneous sets of measures that "tap" the "domain" of a (multidimensional) construct (note my marks), not reflective indicators of ONE and the same underyling dimensions/latent variable.
In this thread, you will find more:
https://www.researchgate.net/post/Dear_scholars_how_to_design_a_Structural_Equation_Modeling_surveys
In addition, as you started with some conceptional ideas, I would skip the EFA and go direct to CFA (if this is reasonable).
Best,
Holger
I'd start with a parallel analysis to identify the number of factors to extract. Then I'd do an exploratory factor analysis (EFA) unless I had very strong theoretical preconceptions. I'd finally do a confirmatory factor analysis which "cheated" based on the EFA results. (I have more confidence in the consistencies of respondents' responses than in the talents of most would-be theoreticians including myself.)
- Badges
- Science topic
- Similar topics
- Correction