3 ms·
Citing the conclusion of a detailed comment on pubpeer (https://pubpeer.com/publications/AB8B5472EB1E3301B3581AA7EDB750 https://pubpeer.com/publications/AB8B547
by karencarits 4y ago
Citing the conclusion of a detailed comment on pubpeer (https://pubpeer.com/publications/AB8B5472EB1E3301B3581AA7EDB750 https://pubpeer.com/publications/AB8B5472EB1E3301B3581AA7EDB...):
> a highly-selective analysis of highly selected participants shows that a small minority of them seem able to respond to simple questions while they sleep. the stats here are a mess.
- giantg2 4y agoI don't think the stats were the point. The conclusion isn't about quantifying it. It's essentially a binary existence statement and "we found this phenomenon that needs more research".
- radu_floricica 4y agoLike giantg2 said. There are (at least) two types of studies: this is purely exploratory, proof-of-concept type. The reason "a group of british scientists" studies got such a bad reputation is because they were largely aimed at finding out new stuff, but got presented by the press as the proving things. For discovery (or university science play) you're good with samples of 10-20 people. For proving stuff you need meta-analysis of several studies, ideally including RTCs and a couple of large longitudinal ones. shrug science is complicated. Some stuff about how to do it properly has been leaking into the water supply, but overall public opinion is pretty much detached from the details and still thinks a study is a study.