3 ms·
All this article says is that Haidt's analysis is just correlation, and then cites a single metastudy from 2019, and doesn't go on to explain what Haidt did wro
by dadrian 3y ago
All this article says is that Haidt's analysis is just correlation, and then cites a single metastudy from 2019, and doesn't go on to explain what Haidt did wrong. This is despite extensive writing from Haidt that deconstructs that specific metastudy.
- pama 3y agoOnly correlation can never explain something. A bomb will tear a building into bricks, but finding bricks every time after a building explodes does not help identify the cause. Suppose it ended up eventually that the causal source was a novel virus epidemic, say N1H1, instead of social media or the post-2008 economic mess: it wouldn’t help much if people stopped using their social media. I don’t support social media usage for other reasons that are easier to quantify and assign causality, including the waste of time, but I understand the severe limitations of correlations and the differences to causality. Medical research has suffered greatly from errors in causality occasionally dismissing partial self-therapies due to observed correlations, whereas the careful delineation of causation is one of the key ingredients of what made physics such a mature science over the centuries. It is of course hard to try causal experiments for understanding the root of epidemics, and in this case any attempts for causality in therapies are also challenging.