Now that it seems the entertaining race for the Republican nomination has reached trash time, where the candidates should be putting in their third-stringers to run out the clock (though Rick Santorum already often seems to be a third-string sub for himself), I thought I’d take a look at what we can gather from state polls.
Dave Leip’s US Election Atlas presents a convenient list of state polling:
Don’t worry too much about the map—it uses a simple average of the last three polls in a state, regardless of how old the polls are or who conducted them, and this leads to some weird results. For example, nobody thinks that Texas is a tossup because of some poll conducted in June of 2011. That said, if you just look at the states where the map has Obama leading, they make up a pretty plausible scenario for a narrow win—the only obvious change I would make is that Iowa is more likely to go Democratic than Ohio (Iowa has been bluer than Ohio for 7 straight elections. There was one poll of Iowa last month that showed Romney leading by 2 points, throwing off the map a bit and also, absurdly, leading NBC to move Iowa to “lean GOP” on its electoral map. As if.
Let’s consider instead the list of recent polls; a good cutoff might be two weeks ago or a bit more, since further back we get into the time when Romney was slumping behind Santorum, and that might skew the general election polling.
There are a couple of polls in safe blue states, and I don’t think we can conclude much from the news that Obama is up 26 in Vermont or 30 in New York. In case you’re wondering whether Massachusetts will be in play since it is Romney’s home state, the answer is no. The home-field advantage is quite variable (I think that it tends to be larger for candidates from small states; Bill Clinton had a big home-state bump, the Bushes not so much), but the average in recent years has been about 6 points. Obama will win MA by double digits unless the national result is a Romney landslide. (In another bizarre move, the NBC map lists Illinois as “Likely Dem”—Likely?). The home-state issue would become a real cause for concern if Rick Santorum were to come back from the dead again, since an extra 6 points would have a big impact in a big and barely-blue state like PA (Obama won it by 3 points more than his national margin in ’08).
There are recent polls in a couple of states that Obama won big in ’08 but were once battlegrounds: he’s up 10 and 14 in New Jersey, numbers that would go along with a small national lead, say 4 or 5 points. In Wisconsin, the Republican’-leaning Rasmussen has Obama up 5 and the Democratic-lieaning PPP has him up 14. Ordinarily WI is a classic battleground state, but Obama won it by 14 in ’08, perhaps getting a boost from the proximity of Chicago. A 10-point lead is compatible again with a modest national Dem edge.
We also have multiple polls in two of the most swingy of swing states; I think that if Obama wins either Ohio or Virginia, he will win the election, and if he loses them both, he will lose. He won them both in ’98 by a bit less than his national 7-point margin. Obama has led in all the recent Ohio polls, including those taken during Romney’s recent surge—it seems likely that the state’s working-class demographics aren’t the best fit for Mitt, who trails by 4 and 8 points in the two most recent surveys. Being 6 points down in Ohio is much worse news for the GOP than being 12 points down in New Jersey.
In Virginia the polls are all over the map, the most recent showing a 1-point Romney lead and a 17-point Obama lead (!). Best guess is that Obama is slightly ahead here too, perhaps by a smaller margin than in Ohio, but we’ll have to wait for some more consistent data.
Finally there have been a couple of polls in states Obama lost in ’08. In Georgia, he is down 7, which is a couple of points worse than he did in ’08; being down 7 in GA is like being up 12 in New Jersey, indicative of a clear but not huge national advantage. More impressive is a poll showing Romney up 6 in Tennessee, a state that McCain won by 15 points. It’s just one poll, but if it is confirmed by others it will be a disaster for Romney; if he wins Tennessee by 6 points, he will lose neighboring states like North Carolina and probably Missouri.
I suppose that Romney will get some sort of bump from the aura of victory that will probably accompany tonight’s results, and whether that lasts will largely depend on the economic news. But if the economy keeps improving, he will have trouble getting traction in some places that he needs to win.
[Update: I somehow forgot to include the recent CNN poll in Georgia showing the race tied; as with the Tennessee poll, if this is not an anomaly it’s terrible news for Romney. OTOH, perha GOPers were getting ready to vote for Newt, they weren’t feeling as positively towards Mitt as they will be in a few months.]