3 ms·

It's amazing how many people seem to lack even an intuitive understanding of correlation (i.e., "X tends to cause Y"), and think that providing a (usually anecd

It's amazing how many people seem to lack even an intuitive understanding of correlation (i.e., "X tends to cause Y"), and think that providing a (usually anecdotal) counterexample negates the relationship/trend. Alice: Smoking tends to cause cancer. Bob: My aunt Wanda smoked for years and lived to be 89. Alice: Exercise is good for you. Bob: My cousin started jogging and found out he had cancer the next month. Alice: The global climate is warming. Bob: The weather in my hometown was unusually cold yesterday. Alice: Childhood vaccinations prevent many deaths. On the whole, they provide a net benefit. Bob: A child in a nearby town had a severe allergic reaction to a chicken pox vaccine; children would be better off just getting the disease. This kind of innumeracy is distressingly common.

beenBoutIT 8y ago
Logic is not something most people seem to care a whole lot about.
mmt 8y ago
>This kind of innumeracy is distressingly common. I would say it certainly doesn't help that science reporting essentially never gives the numbers that matter, nor explains what they mean. > Covering 33 randomized clinical trials with 1,877 participants That's an average of under 57 participants per clinical trial, which seems a bit sparse to me. Still, that's not my point so much as that those are the only numbers about the study in the article. Where is the mention of statistical significance or uncertainty? It's also difficult to teach, since I don't think we're naturally geared for intuiting about numbers bigger than we might encounter in nature. Also, anecdotes about people we know personally are more likely to resonate with us than mere statistics about faceless strangers (not that it explains why a few hundred Anthrax vaccine deaths of strangers would have been more objectionable than the routine tens of thosands of annual road deaths of strangers).

beenBoutIT 8y ago

Logic is not something most people seem to care a whole lot about.

mmt 8y ago

>This kind of innumeracy is distressingly common. I would say it certainly doesn't help that science reporting essentially never gives the numbers that matter, nor explains what they mean. > Covering 33 randomized clinical trials with 1,877 participants That's an average of under 57 participants per clinical trial, which seems a bit sparse to me. Still, that's not my point so much as that those are the only numbers about the study in the article. Where is the mention of statistical significance or uncertainty? It's also difficult to teach, since I don't think we're naturally geared for intuiting about numbers bigger than we might encounter in nature. Also, anecdotes about people we know personally are more likely to resonate with us than mere statistics about faceless strangers (not that it explains why a few hundred Anthrax vaccine deaths of strangers would have been more objectionable than the routine tens of thosands of annual road deaths of strangers).