Do you disagree with the idea that Republicans are smarter (except at the top end) than Democrats, or that "liberals" are smarter than "conservatives"?

I don't.

My point was that using a test that heavily relies on ‘learned’ knowledge such as Wordsum may have exaggerated the effect (compared to what one would see if one used a more culture-neutral test such as Raven's progressive matrices) when some of the groups have historically been educated more than others for additional reasons besides IQ (even if said reasons correlate with IQ, so long as the correlation isn't close to 1).