developing a valid scale for data from multiple countries

#1
Hi,
I am an undergrad Sociology student. For my final thesis I am trying to asses the level of political and social trust using the European Social Survey, which contains data from different European countries.

Using EFA I want create two scales, based on the items from the questionnare, one for political and one for social trust. Besides measuring the trust in the whole population, I want to also compare the individual countries.

How do I make sure that my scales are valid for all countries? I have read about Multi Group Factor Analysis, but this method seems a bit too hard for me right now. May I just run an EFA for all countries separately and adjust the scales so that the factor structure is the same? If not what are methods can you suggest? I would prefer a method that I can implement using SPSS
 

staassis

Active Member
#2
How do I make sure that my scales are valid for all countries?
Define "valid". By definition, EFA will offer one of the best ways to explain variability of all those political and sociological measurements with just a few factors. Statistics is not concerned with whether the factors make sense from the point of view of existing social sciences literature.