4 ms·
Right except this (alleged) fraud has been going on for a little over a decade. So error correction is always nice (and better late than never) but it also mean
by DonsDiscountGas 3y ago
Right except this (alleged) fraud has been going on for a little over a decade. So error correction is always nice (and better late than never) but it also means that any particular finding can be very unreliable. If I'm deciding on a drug to take (or a medical procedure) I need to decide pretty soon; if I'm getting bad information it doesn't help me if in 10 years the field says "we're sorry, our bad".
- Yoric 3y agoThe problem you mention is very real but do you know of any criteria that actually work better than "trust the science"?
- Nevermark 3y ago> […] better than "trust the science"? Probably change that to a gradient: “Trust science more, when it’s been tested by multiple groups, it’s been fruitfully built on, etc.” I think it is fair not to overly trust any single paper. Also worth considering the different qualities of evidence in different papers. Mathematics has the highest form of evidence, being easier for readers to vet and reproduce themselves. Human behavior the lowest given how impossible it is to isolate all conditions in a human head, and the lack of any reliable overall models of a human mind.
- Yoric 3y agoFair enough.
- dcow 3y agoMaybe: only trust science that is statistically significant and can be replicated?
- Yoric 3y agoI'd say that it's not science until it is replicated, but you are right, this is not always what people mean by "science".
- ta8645 3y agoThe point is to not treat it like a religion. Scientific conclusions are not blessed by god as divine and absolute truth. As much as we wish the world was black and white, it's much more shades of gray, and we should each do our part to make sure that religious zealots do not use "the science" as a political cudgel.