5
$\begingroup$

I intend to do an exploratory factor analysis among items that have different types of measurements: for example, some items have bounded scales (1-100%), some items take small and other large values (different units indeed, such as dollars or hours).

I've been reading a lot and haven't found a negative answer, i.e. that it is not possible to do such analysis with such varied types of items; however, every example I came across is based on items that take the same values (e.g. 4-value likert scales from questionnaires).

In sum, does anyone have a clear answer (and reference?) on this matter?

$\endgroup$
1
4
$\begingroup$

EFA is based on the correlation matrix of your items. That means, scale differences between items are not a problem, as the correlation is a scale-free index. You should check your variables for other problems (e.g., ceiling or floor effects in your bounded variables), but different units do not pose a problem.

$\endgroup$
1
$\begingroup$

I do not have a clear answer but it seems that the problem you describe is a small subset of what is usually called “common method variance”. You might find more information on this and some further references in:

  • Podsakoff, P.M., MacKenzie, S.B., Lee, J.-Y., & Podsakoff, N.P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88 (5), 879–903.
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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