Combining samples for EFA


I have conducted an experiment in two different countries. The experiment included a survey. The same survey was used. Is it valid to combine the two sets to perform exploratory factor analysis, or there will be a scale effect (e.g. as in the case of logit choice models).

Thank you


New Member
Hey mazendanaf.

Do you think there is any kind of correlation involved between the different data sets?

I ask this because if you some kind of correlation (especially intra-class correlation) then you will need to adjust for this when combining data from different sources.

If you assume independence and your analyses/models also assume independence then you can just combine them in the standard way: otherwise you will need a way to factor in the correlation aspects.