Skip to main content
ABC News
Aug. 8: Colorado and Virginia Flip Places in Forecast

Mitt Romney got one of his better polls in a long while on Wednesday. The poll, in Colorado, was one of three state surveys conducted by The New York Times in conjunction with Quinnipiac University and CBS News. It gave Mr. Romney a five-point lead there, with 50 percent of the vote, to President Obama’s 45 percent.

This is actually the first Colorado poll of the cycle to show Mr. Romney with a lead of any margin in the state. And a five-point lead, especially in a credible poll with a large sample size, is a big number to put up on the scoreboard.

Of course, Mr. Romney probably does not actually lead by five points in Colorado. Polls are subject to sampling error and other methodological choices, and other recent surveys of the state do not suggest such a favorable outcome for him. Our forecast had Mr. Romney narrowing his deficit in Colorado to 1.3 points from 2.8 points on the strength of the poll, with his winning chances increasing to 43 percent from 34 percent. Still, our forecast is designed to be rather conservative about weighing new data, so that’s a pretty big move.

The problem is that the rest of Wednesday’s data was mediocre for Mr. Romney. In Virginia, the New York Times/Quinnipiac University/CBS News survey had him trailing Mr. Obama by four points. Another survey, from the Republican-leaning firm Rasmussen Reports, put Mr. Obama two points ahead there.

The pair of Virginia polls offset Mr. Romney’s strong showing in Colorado. Mr. Obama’s chances of winning Virginia rose to 67 percent from 59 percent on the surveys, and his projected lead increased to 2.4 points from 1.3 points. Eagle-eyed readers will note that Mr. Romney’s gain in Colorado was slightly larger than Mr. Obama’s in Virginia — but Virginia contains 13 electoral votes to Colorado’s 9, so it basically washes out.

Mr. Obama also got a pair of decent polls in Wisconsin — one putting him ahead by six points, another by five — but they had less impact on the forecast, since the data had already pointed toward Mr. Obama’s holding a lead in Wisconsin by about that margin.

National polls released on Wednesday were slightly unfavorable to Mr. Romney. A week or two ago, we seemed to have a slight shift in national tracking polls toward Mr. Romney. But that seems to have reversed out now, with Mr. Obama’s numbers in daily and weekly trackers a point or so stronger than they were last week. There was also a poll from Reuters/Ipsos, putting Mr. Obama seven points ahead in the national race, although that polling firm has typically shown good numbers for him.

Overall, all the new data didn’t have much of an impact on the forecast. Mr. Obama’s chances of winning the Electoral College rose nominally, to 72.5 percent from 72.0 percent on Tuesday.

What Mr. Romney needs is to have numbers like the Colorado survey more often. Yes, it was probably a bit of an outlier. But then again, some of the polls showing Mr. Obama with impressive numbers last week were probably outliers as well. (I wouldn’t put much money on Mr. Obama’s winning Florida by six points if the election were held tomorrow, for instance.)

Still, even outlying polls are distributed around a central tendency. If Mr. Romney held leads of two or three points in the most important swing states, instead of trailing by about that margin in many of them, we’d be seeing polls putting him up by five points in one of them a lot more often. In fact, we’d see numbers a bit better than that: the occasional poll putting him ahead by six points in Iowa, seven points in Ohio, eight points in Florida and so on.

But back to Colorado and Virginia. Is it really plausible that Virginia is actually the easier of the two states for Mr. Obama to win, as the most recent numbers suggest?

Both are often lumped together under the heading of wealthy, suburban, socially moderate but fiscally conservative states that flipped from red to blue in 2008. But their demographics and political cultures are reasonably different. Virginia has a lot more African-American voters, but fewer Hispanics. Colorado has a decent number of evangelical voters in the Colorado Springs area, but Virginia still has more of them. In Virginia, center-left voters tend to identify as Democrats; in Colorado, they’re more likely to call themselves independent. Virginia, with its connection to Washington, is a bit more of a pro-establishment state, while Colorado, like much of the Mountain West, is more anti-establishment.

There’s also an important economic difference between the states. Virginia’s unemployment rate is well below the national average, at 5.6 percent. Colorado is still suffering through 8.1 percent unemployment. Its jobless rate has risen some in recent months, while Virginia’s has continued to fall.

Mr. Obama won Colorado by a larger margin in 2008 — by 8.9 percentage points, against 6.3 in Virginia — but that isn’t all that large a difference; small enough that some of these differences could reverse their order on the electoral map.

At the very least, it seems misguided to lump Virginia together with states like North Carolina and Indiana that Mr. Romney should have a much easier time winning back.

And it seems naïve to assume that Colorado is part of any kind of electoral firewall for Mr. Obama. If the economic recovery is more sluggish in the West than in other parts of the country, its anti-establishment spirit will kick in, and this won’t be the last poll we see putting him behind there.

Nate Silver founded and was the editor in chief of FiveThirtyEight.