Probit regression - I'm not sure if I have a dependant variable

My problem seems simple but I am a bit lost.
Can I model a probit/logit regression over a survey in order to estimate the probability of beeing classified as D,if I am working with two surveys that are different (different questionnaire) and they are using different samples (persons)?

The doubt arises because I know that I can work over a probit if I were working with the same person and two surveys. I would be defining an X and Y variable. But if I am not treating the same people with the two survey to model the probit regression, then I would playing with nothing.

I hope I made myself clear.
Thanks for your time and interest.