The European Tribune is a forum for thoughtful dialogue of European and international issues. You are invited to post comments and your own articles.
Please REGISTER to post.
It would be interesting to know exactly how (ie what econometrics software, and how it was tortured) but if the possibilities are many I doubt it would be worth the effort.
But, given the number of parameters he's fitting, my guess would be that he ran a simple model and then expanded the number of model parameters until he got a result he liked, and never bothered to check for misspecification.
At least that's how I would do it if I wanted to do cargo cult science for a belief tank: It produces a superficially plausible result and checking the methodology requires surgical reconstruction. Which means people who don't know better (I plead guilty) tend to take it at face value.
(As an aside, I did know that the author was a liar - I just stopped investigating when I'd found the easiest to detect lie.)
Friends come and go. Enemies accumulate.
by Frank Schnittger - Nov 7 71 comments
by Cat - Nov 3 20 comments
by gmoke - Oct 31 5 comments
by Frank Schnittger - Oct 28 17 comments
by Frank Schnittger - Oct 25 20 comments
by Oui - Nov 161 comment
by Oui - Nov 89 comments
by Frank Schnittger - Nov 771 comments
by gmoke - Nov 6
by Cat - Nov 320 comments
by gmoke - Oct 315 comments
by Frank Schnittger - Oct 2817 comments
by Frank Schnittger - Oct 2520 comments
by Cat - Oct 2127 comments
by Frank Schnittger - Oct 1779 comments