So I’m familiar with the Many Labs project that was designed to test replicability in psychological research because replicability is largely nonexistent in these studies. But I’ve read a lot of statistical criticism of the Many Labs studies too. Can someone explain it to me like I’m 5 what the replicability crisis is, what Many Labs did, what their conclusions were and what the criticism centers on?
In: Other
The simple version is that a large amount (more than half) of psychology reaearch is probiably bunk. Even some major highly respected and cited ideas are based on 1 expiriment by 1 group (for example social priming).
some people tries to replicate these expiriments and failed, but journal’s ignored them since those arent new research.
then it finially came to light when a group called many labs attempted to replicate 100 major papers, and failed 2/3rds of the time.
This kicked off the ongoing replication crisis as people try to replicate all the old expiriments and try to figure out what is actually true, and what was 1 time flukes.
The controversy over many labs, is that for some of these studies that failed, many labs didnt always have the same statistical power as the original, so it is possible many labs had a fluke failiur instead of the ov being a fluke success.
The solution to both is the same though, we need more researchers replicating eachother’s expirements, especialy in human centric fields like psychology and medicine, so we can sort out what actualy works and what doesnt
here is a good video about it (and other things) https://youtu.be/42QuXLucH3Q
Latest Answers