3 ms·
You don't make them. You let people volunteer to feed poo to their babies and then randomly select half of them to actually feed their baby a placebo.
by AndrewDucker 6y ago
You don't make them. You let people volunteer to feed poo to their babies and then randomly select half of them to actually feed their baby a placebo.
- lordnacho 6y agoIt's the letting them volunteer that's problematic. People who volunteer might be systematically biased toward/against whatever thing you were trying to measure. Then giving half of them a placebo would not reveal the true effect. Suppose there's a hidden gene that causes people to be able to fly like superman if they're given their mother's poop. The gene also causes people to not volunteer in poop feeding studies, because the experimenters unfortunately placed the adverts on the ground. You find volunteers, you give half of them the poop and the other half the placebo. You find that none of them can fly, so you conclude poop doesn't cause flight. If you make this probabilistic it gets less obvious to pick apart statistically, but same problem.
- __s 6y agoThe point of double blind isn't to get around your crazy scenario, but to be able to say you evaluated the effect of one factor on your self selected group Which generally ends up being that lots of studies are about isolating the effect of one factor on young white educated men, but you can at least say you've isolated the factor as opposed to just comparing young white educated men to the general population
- watwut 6y agoThe volunteer thing is pretty much standard in medical studies due to ethical considerations. You dont get to make experiments on people without proper consent etc done.