Want to be notified when a new article is posted? Entere your email here.

Sunday, November 18, 2007

Econometric Haiku

From econometrician Keisuke Hirano.
T-stat looks too good.
Use robust standard errors--
significance gone.
HT: Marginal Revolution

This hit home, since I just had to figure out how to do clustered standard errors for a paper. Lukily, my significance didn't go away.

No comments: