1
$\begingroup$

I would like get the factor scores of R/lavaan, achieved by a CFA model (single factor, 4 continuous indicators), by hand. Which is the formula (equation) as function of the observed indicators. In addition, I would like also get them by ordinal Likert indicators (again single factor). How I can get the continuous values that R/lavaan assigns to the Likert categories after the fit?

Many thanks in advance.

Best,

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ Do you have a particular type of factor score in mind? $\endgroup$
    – Preston Botter
    Oct 13, 2021 at 23:38
  • $\begingroup$ No, a generic latent varibale score, implemented in lavaan package. $\endgroup$
    – Davide Guido
    Oct 14, 2021 at 12:47

1 Answer 1

Reset to default
2
$\begingroup$

Welcome to CrossValidated Davide,

As Preston harkens to, there are different types of factor scores, and so you need to be specific about the kind you are wishing to compute--there isn't necessarily a generic factor score.

The seminal work in this area (IMO) is the paper, "Regression Among Factor Scores", by Skrondal and Laake. The authors discuss there being two approaches to calculating factor scores: the Regression method and the Bartlett method (this, I suspect, was the clarification Preston was after). Importantly, Skrondal and Laake demonstrate that each method is better/worse suited for when the factor score is to be used in a particular role (either as an explanatory or response variable) in some sort of substantive model.

Anyways, you'll find the formula for each in their paper.

References

Skrondal, A., & Laake, P. (2001). Regression among factor scores. Psychometrika, 66(4), 563-575.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.