Hacker News Reader

Comment by schrototo | Hacker News Reader

The TL;DR version: "This journal does not publish replication studies, whether successful or unsuccessful"

Which, to me, seems to be an admission that the Journal of Personality and Social Psychology isn't actually a scientific journal. An initial experiment/study is only a part of the process. If you get extraordinary results out of an experiment, but no one else sees the same results with the same experiment, well, something was up with your run of it. It is vital to the body of scientific knowledge that re-runs of published experiments are performed, recorded, and published. Otherwise, there's no way to know whether the original result was real or a fluke (or fraudulent).

When dealing with failed replicability it is important not to conflate a poorly designed experiment with a poorly written methods section. Often there is tacit or socially negotiated knowledge required to properly replicate an experiment. This doesn't mean the science is bad, it means science is more social than some would like to admit. For some interesting reading on this check out Harry Collins's paper "The Seven Sexes: A Study in the Sociology of a Phenomenon, or the Replication of Experiments in Physics"[1]

[1] http://soc.sagepub.com/content/9/2/205.short

The truth of that depends on the accuracy of the title of:

Violence is a curvilinear function of temperature in Dallas: A replication. Rotton, James; Cohn, Ellen G. Journal of Personality and Social Psychology, Vol 78(6), Jun 2000, 1074-1081. doi: 10.1037/0022-3514.78.6.1074

http://psycnet.apa.org/psycinfo/2000-07798-005

I spluttered incoherently after reading that line. Obligatory xkcd: http://xkcd.com/882/

Reply on news.ycombinator.com