Sunday, 24 May 2015

and that's why...

you want the data of articles posted for replication purposes... Have a look at

where two grad students found out through Sherlock-Holmes-style sleuthing that a study published in Science was actually (probably) made up 100%. The claim of the paper was that exposure to a gay canvasser changed people's minds about gay rights enormously and lastingly; but it seems that the data were actually taken from a national survey and doctored. More detail here.

Over the years (...) I have pushed for a data replication (and posting) policy at every journal where I was an editor. At Explorations, we are finally doing it for over a year now. I think that I failed at the European Review of Economic History, where we always agreed it was a good idea but then never got round to it. It seems that the Journal of Economic History is going to start requiring replication files in the near future (the policy is on the books for almost a year). Now, it's the Economic History Review that is also still missing from the list (actual implementation score 1:3; policy score 2:2). [update 2-6-2015: one of the editors wrote to me saying that, well, you know, the Review is a journal of economic and social history, and hence, it may not be appropriate to have a one-size fits all policy, such as one that may allow actual replication, but they remain open to discuss this in the future... (in case you missed it - this is sort of polite British parlance for forget it, we think people should not post their data or share anything - imagine someone else might actually want to use it! or that it was actually forged, and cannot be replicated! better to keep it all hush-hush, like in the good old days.)]

If you ask me, in a field that prides itself on its datawork, it is nothing short of amazing that this has been taking so long (relative to the general interest journals in economics), and that half the field is not even going to do the basics. Next time you ask yourself why economic history isn't getting as much respect as you would like... think about this kind of thing, and be embarrassed.

No comments:

Post a Comment