Isn't this a classic ethical problem in psychological research? You can measure the correlation all you want, but to prove causation you have to be able to con

Isn't this a classic ethical problem in psychological research? You can measure the correlation all you want, but to prove causation you have to be able to control the source (creteris paribus). It's hard to get a study approved that intents to expose subjects to something you know is harmful.(?) With sigarettes, researcher could just prove the material harm, and conclude causation, but proving purely psychological harm seems hard to prove