You're absolutely right. Consent makes all the difference in the world. I mis-understood your last statement and I apologize for I misrepresenting your idea.

If you're talking about research done with consent, what you're suggesting is already widely done, and not just in prisons. Prisons, group homes, and the like are often studied because some of the variables are already controlled. So, I wrongly assumed that you were suggesting something different.

For example, there are almost 5,000 results on Pub Med related to prison populations:
https://vsearch.nlm.nih.gov/vivisimo/cgi-bin/query-meta?query=prison&v%3Aproject=nlm-main-website