Monday, August 15, 2016

On the academic front, some people need to remember that Michael Bellesiles

was not a good example.  And what happened to him.
Fox News reported a few weeks ago that University of Alabama researcher Adam Lankford, author of a media-sensation study on the connection between mass shootings and gun ownership across countries, refuses to share his data and the details of his methodology with skeptics.
Worse yet, Lankford’s editor apparently does not agree that the lack of transparency is a problem. The editor instead fell back on the old standby that the study has been peer reviewed — as if that removes all doubt about its validity. (See Slate’s Daniel Engber on why peer review deserves much less esteem than it receives from the media. Short answer: Peer reviewers are rarely thorough, and they tend to build echo chambers for like-minded researchers.)

So why doesn’t Lankford just release the data? “I am open-minded about sharing data with other scholars for collaborative purposes, and consider those opportunities on a case-by-case basis,” he told Fox News. In other words, he’ll share data only with people unlikely to criticize him.
Which is a good indication that your data is crap, and you know it.

If you're not familiar with Bellesiles, you can start here.  Short version:
Scholar claimed 'there weren't very many guns in early America.'
Didn't want to share data and methodology.
When forced, turned out he'd been lying his ass off and faking data.
Did not end well for him.  Rightly so.

1 comment:

Jerry The Geek said...

Now that you remind me, I recall the Bellesiles kerfluffle. And also Kellerman, et al. I just finished reading John Lott's latest book "The War On Guns". He continues to ignore the patent liars in his genre. I have no idea why he is so patient with them, except that he is focused on truth ... not lies.

Maybe we should be focusing on truth-tellers rather than liars? Lott provides a nobel example.