I am using Stata to perform a GSEM on 30 binary and ordinal variables. I did the EFA, the Cronbach Alpha is below 0.80 but above 0.70 fro the first 3 factors. I did some reverse coding on a few items to achieve unidimensionality,. All the binary variable (0-1) were placed under constructs of factors with very low reliability. What would be the next steps to CFA and GSEM, taking into account that my variables are ordinal and binary?
I recommend that you use free access programs such as JASP. If you are validating an instrument I recommend that you make an exploratory analysis and then confirm it by means of structural equations. Considering your dichotomous items we recommend guttman coefficient, with factorial loads greater than .50 with a good index of goodness of fit of the theoretical model that you propose.
https://www.theanalysisfactor.com/principal-component-analysis-for-ordinal-scale-items/
https://www.stata.com/statalist/archive/2012-08/msg01035.html
See this link
https://www.theanalysisfactor.com/principal-component-analysis-for-ordinal-scale-items/
- Badges
- Science topic
- Similar topics
- Philosophy
- Moral Philosophy
- Free Will