Arnold makes a fair point about the size of the effect of personality on ideology.  Why then am I so impressed?  Because compared to most conventional predictors of ideology, the personality variables do well.  I’ve spent many hours looking at ideology data before.  Variables widely believed to predict ideology (particularly education and income) usually matter quite a bit less than Gerber et al‘s personality measures.

If Arnold’s willing to say, “Nothing predicts ideology well, but if these results hold up, personality is among ideology’s strongest predictors,” we’re basically in agreement.