Monday, 5 April 2004

Yeay, it works!

I won’t spoil the whole Midwest paper for you, but here’s the pretty path diagram of the LISREL model. Guess the coefficients and standard errors—it’s fun for the whole family!

Actually, the most amazing thing about the paper is that the model works, despite the suboptimal polling data it’s based on—almost all of the manifest variables are dichotomous or trichotomous.