It's practically impossible to to do a controlled study on any of these, so we'll likely never know for sure. If you were tasked with creating a study to determ

It's practically impossible to to do a controlled study on any of these, so we'll likely never know for sure. If you were tasked with creating a study to determine, for any reasonable definition of "certainty", whether small amounts of alcohol increase or decrease one's lifespan, you'd never be able to come up with anything that passed any sort of moral or practical sniff test. So what you are left with is trying to tease causation out of relatively loose correlation, which is and always will be so problematic as to be near useless. The results will always just support the researchers' pre-conceived notions. Smoking is only so obvious because the rates of lung cancer there are an order of magnitude more. 80-90% of lung cancer is in smokers, while only 12% of Americans smoke. No other health issue is so clear, people who are moderate drinkers only have very slightly elevated risks of almost everything, so it could be confounded.