3 ms·

With this kind of argument, nothing ever happens! >Sickness levels in people working for the NHS is pretty specific and is very likely not to generalise. "The

by aredox 1y ago

With this kind of argument, nothing ever happens! >Sickness levels in people working for the NHS is pretty specific and is very likely not to generalise. "The canary in the coalmine just fainted, but this is pretty specific to canaries and very likely not to generalise to us human coal miners." Your fundamental mistake is that you think "correlation does not imply causation" when the correct sentence is "correlation does not prove causation" - but it does certainly imply it. That is the job of statistics and experiments to prove it beyond doubt, but it always starts with a correlation that the researchers then dig into.

csnerd9 1y ago
Perhaps unfortunately, math and computer science have redefined "imply" and the "=>" implication operator to mean "certainly follows", which does means "prove". The OP is probably using those fields terminology, rather than casual commonplace meaning of imply, which is as you were arguing, merely "suggests". Also, "beyond (a reasonable) doubt" is a legal standard, not a scientific one. Science almost always maintain some doubt :)
- graemep 1y ago
  Yes, exactly how I was using it, because it is a widely used phrase in discussing statistics, and especially when teaching statistics. https://en.wikipedia.org/wiki/Correlation_does_not_imply_causation https://en.wikipedia.org/wiki/Correlation_does_not_imply_cau... It is also of particular importance when discussing time series statistics because spurious correlations are extremely common.
  - aredox 1y ago
    Oh, sorry, French is my first language and I am more used to "imply" in sentences like "what do you imply?" In French, that saying is much more direct: "correlation is not causation".
AnimalMuppet 1y ago
https://xkcd.com/552/ https://xkcd.com/552/ (The alt text: Correlation doesn't imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing 'look over there'.)

by aredox 1y ago

csnerd9 1y ago

Perhaps unfortunately, math and computer science have redefined "imply" and the "=>" implication operator to mean "certainly follows", which does means "prove". The OP is probably using those fields terminology, rather than casual commonplace meaning of imply, which is as you were arguing, merely "suggests". Also, "beyond (a reasonable) doubt" is a legal standard, not a scientific one. Science almost always maintain some doubt :)

graemep 1y ago
Yes, exactly how I was using it, because it is a widely used phrase in discussing statistics, and especially when teaching statistics. https://en.wikipedia.org/wiki/Correlation_does_not_imply_causation https://en.wikipedia.org/wiki/Correlation_does_not_imply_cau... It is also of particular importance when discussing time series statistics because spurious correlations are extremely common.
- aredox 1y ago
  Oh, sorry, French is my first language and I am more used to "imply" in sentences like "what do you imply?" In French, that saying is much more direct: "correlation is not causation".

AnimalMuppet 1y ago

https://xkcd.com/552/ https://xkcd.com/552/ (The alt text: Correlation doesn't imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing 'look over there'.)