is there a best practice for combining multiple factor scores into a singel index

[Christoph Dietrich]

Hi everyone

I am uncertain how to proceed in this case:

Based on theoretical assumptions I have run four separate factor analysis and each resulted in a separate standardized factor score. However, I now want to combine these scores into one index that I can use as dependent variable in further analysis.

To do so:

Would it be advisable to run another factor analysis over the four standardized factor scores? (Resulting just one pre-defined factor)
Or would it be better to create an additive Index with my four standardized factor scores?

Like I said before the four standardized factor scores are based on theoretical assumptions and are pooled in a way so that higher values indicate a stronger manifestation of the theoretical assumptions.

Best and thanks a lot
CD

[Nick Cox]

The point of your first round of factor analyses seems to have been to get careful composites of bundles of variables.

If you now put them together the risk is that you get a mush of "variables that interest me" that won't be easier to interpret and you would be throwing careful work away. You'd be treating from theory-guided work to the opposite if you don't have a clear idea of why there should be a single response and how it is to be estimated.

Just filling the silence with an awkward opinion.