Ok. The election went well and we're all happy. Still I see something perplexing regarding Ohio.
Ohio finished with a 1.9 point Obama advantage. Great. But when I look at how the predictors fared with Ohio, I see something concerning.
538 predicted Obama by 3.6; Huffpo, Obama by 3.4; and Real Clear Politics, Obama by 2.9. They were off a little. So big deal. Right?
This is the ONLY state, of the eleven Real Clear Politics toss-ups, where RCP showed an Obama bias. For the other 10 states, Real Clear Politics had nearly a 3 point Romney bias. With a one point Obama bias for Ohio, this is a four point bias differential.
538 had a 1.3 bias average for Romney for the other 10 states, but had a 1.7 bias for Obama in Ohio: a 3 point bias differential. He had two other states (of these 11) with an Obama bias: Pennsylvania .7 and North Carolina .5
Pollster had a 1.8 bias average Romney for the other 10 states, but had a 1.5 bias for Obama: a 3.3 bias differential. They had two other states (of these 11) with an Obama bias: Pennsylvania .5 and North Carolina .6
It could be simply coincidence that the polling in Ohio was so far off, and so uniquely. But if the polls and the results don't match, the polls are not the only possibility for the mismatch.
I don't have the answer but if I lived in Ohio I'd be asking a lot of questions.