3 ms·
> since “retail investors destroy value when trading” is about as citation-needed as “smoking causes lung cancer” at this point I mean, it could be a confoundi
by Scoundreller 7y ago
> since “retail investors destroy value when trading” is about as citation-needed as “smoking causes lung cancer” at this point
I mean, it could be a confounding variable that’s causing cancer in smokers. We haven’t exactly run randomized control trials in humans to clearly show causation.
The correlation sure is strong though.
- hcknwscommenter 7y agoHuh? Sure, randomized control trials in humans have not been performed. But, there is no debate about whether smoking causes cancer. From cell-based and animal models to retrospective epidemiological studies, the data are clear. Smoking causes cancer. Do we need randomized controlled trials to "prove" cutting someone's head off causes death?
- Scoundreller 7y agoI’d go with cancer is correlated with smoking. But yes, I expect some rigour that avoids confounding variables by design to say “causes”. Great theories have been destroyed before. We should account for what degree other characteristics about smokers causes cancer (e.g. socioeconomics or other factors leading to addictions). You can still blink after your head’s chopped off. Debatable if you’re immediately dead or not.
- woadwarrior01 7y ago> I’d go with cancer is correlated with smoking. Judea Pearl‘s book on causal inference, “The book of why” spends a couple of chapters on the fascinating story of how the causal connection between smoking and cancer was finally established.
- et2o 7y agoThis goes to the definition of causality. Ultimately you can never "prove" causality, even with the best designed experiment. You are allowed to make causal claims for non-RCT data. Lookup Bradford Hill criteria, just to start.
- jtms 7y agotell that to the relatively large amounts of polonium-210 found in cigarette smoke