2022-11-10 09:56 UTC
@tomstafford one of the first thing every student/researcher is taught is the importance of random sampling and randomization... and then they go ignore all that and just pretend it's all good (for instance, you can't directly compute p-values in observational studies as the null distribution is not known)
Replies (1)
-
@tomstafford@mastodon.online 2022-11-10 11:24
@mzloteanu I'm not saying people don't underweight the importance of selection bias, but for many research questions non-convenience samples are hard to collect, and you have to start somewhere