Is it proper to use dependent t-test to measure difference across time for scores IF the participants are different people? Although they are different people, the demographic characteristics that's relevant to the scores being compared is very constant.
Interesting. I've posted this question to couple of statistics forums and it all seem to not get much traction. Perhaps this is a harder question to answer than I initially thought. Perhaps subjective in a way.
How can you do this (and why would you want to)? The point of dependent t test is you are comparing the same thing at two points in time (normally) so it is not valid to use an independent t test. If your observations are different why do you need to use paired t test?
I am guessing you are matching something at point one to something similar at point 2.