Exploratory Factor Analysis with dichotomous and continuous items?

General
1

Hey there!

I created a skill performance test with lots of dichotomous items (true and false) and some continuous items (timed performance). And now I want to run an EFA (and later CFA). I can work with SPSS and RStudio but never had to work with mixed type of data before. How should I proceed? Thanks!

2

For EFA, use the principal axis factoring or maximum likelihood methods. SPSS can handle mixed data types, but you might need to ensure that dichotomous items are treated appropriately (e.g., by using tetrachoric correlations).

Regard,
Ultra

1 Like
3

You might want to look into homals (PCA for nominal, ordinal and interval data). Not sure it is able to handle mixed data. See this link: CRAN: Package homals

1 Like
4

Thank you so much! Could you point out how to incorporate tetrachoric correlations into the EFA?

5

Thanks. I will check it out. Hope I can run an EFA with it!