Sunday, June 01, 2003

The value of replication 

I've added the blog of John Lott to the blogroll. As anyone who's followed the gun law story even cursorily will tell you, Lott's two books, More Guns, Less Crime and The Bias Against Guns contain important information in the debate on whether Minnesota's new shall-issue law will help reduce crime in Minnesota. What John has done has been to recreate the data from More Guns (the original dataset being lost in a hard disk crash) and made it avaiable for anyone who wants to replicate the study. (That information is available on the academic side of his site -- warning: You have got to have someone able to run STATA to use the datasets.)

Several economics journals keep datasets to verify all of their empirical work, and a whole journal, the Quarterly Journal of Business and Economics, seeks "replicative works and articles that synthesize the literature in business, economics, and social sciences, especially topics that bridge several disciplines." I really wonder why this doesn't happen at more places. Much of the flap over Lott's work is easily resolved by going to the data. At least with him, you can get it.

You can also see on his site yet another example of how the Gray Lady fails the test of objectivity in their reporting -- and even an example of how Fox has slanted coverage of gun issues (scroll through John's blog). I assume Ally Mitch will take after those.