I will refer to the
parachute example to confront the idea that RCTs are mandatory to confirm cause-effect. I think an unbiased RCT can reliably distinguish between cause-effect vs coincidence, but it is not the only method, or better, it could never be the method in the parachute example; the same as we can nowadays conclude that insulin improves diabetic ketoacidosis without a RCT, or smoking produces cancer and heart attacks.
