I am trying to measure Social Capital using individuals' membership in different organizations. but there are too many categories such as membership in a political, educational, and recreational organization. the same applies to the trust factor, the survey includes individuals' trust in general, neighborhood, and friends. I am not sure how to combine all these into one dependent variable. anyone can help me with that?
one possible solution is to measure how many organizations a person is a part of, but not sure if it is a plausible solution or not.
Hi Rakib,
I hope you are doing well! My immediate thought is that to reduce the number of variables to one variable for the trust variables (because I am assuming it is some type of Likert scale) is to conduct a principal component analysis (easily done in SPSS). Then, you could see which variables load onto the "trust" component at above .4 after extracting one component. Then, average the observed variable that loaded onto the components to get a composite score. This option might be best for the trust score because that one seems more continuous. However, I think it could be possible to just add up the organization affiliation variables (because into a composite score and get a "total organization affiliation" score) because it is simply a tally of organization involvement.
Hope this helps!
Best,
Bryant
Bryant M. Stone Thank you very much for your insight. It was really helpful. i will try as you directed. Best regards.
- Badges
- Science topic
- Similar topics
- Psychology