Dear researchers,
Can we perform factor analysis (EFA and CFA) for scale of 25 items with yes/ no response? If yes which software can be used ? If no , what is the alternatives to reduce the number of items?
Hello Anonymous,
Yes, people frequently perform such analyses. Doing so isn't ideal, as the preferred variable type for factoring is continuous, not discrete, scores. If you are comfortable conceptualizing the yes/no response as indicative of a value from what would truly be a continuum, you could/should use tetrachoric (special case of polychoric) correlations instead of ordinary Pearson correlations as the basis for the factoring.
A number of software packages can handle this, but one that is pretty easy to use is the free program, FACTOR: http://psico.fcep.urv.es/utilitats/factor/Documentation.html
Good luck with your work.
Dear Anonymous
You may use SPSS Amos or more preferably SmartPLS 3 software for this purpose.
You may approach this issue with Correspondence Analysis that deals with Burt’s contingency tables and binary disjoint tables. It’s a factorial method suited to encoded variables.
- Badges
- Science topic
- Similar topics
- Developmental Psychology
- Attachment