Advertisement

What's Behind the State-National Poll Divergence?

By Sean Trende - October 31, 2012

‹‹Previous Page |1 | 2 |

The Hurricane Sandy-related lull in tracking polls gives us a good opportunity to look at one of the more interesting aspects of polling to develop this month: the divergence between national and state polls. Put simply, the national surveys point to a Romney win, while the state polls collectively point to an Obama win. Both can’t be correct.

The RCP Average currently has Mitt Romney up by 0.8 points nationally. He has held this lead fairly consistently ever since the first presidential debate.

Given what we know about how individual states typically lean with respect to the popular vote, a Republican enjoying a one-point lead nationally should expect a three-to-four-point lead in Florida, a two-to-three-point lead in Ohio, and a tie in Iowa. Instead we see Romney ahead by roughly one point in Florida, and down by two in Ohio and Iowa.

Of course, every cycle sees states shift their partisan leans, sometimes radically. Nevada has swung heavily toward Democrats in a relatively short time, while West Virginia has bolted for Republicans. So it wouldn’t be shocking to see the partisan lean in Florida, Ohio and Iowa shift leftward. Indeed, it may be that Obama’s ad war, ground game, and policy decisions over the past four years (i.e., the auto bailout) have “frozen” these states in place vis-à-vis the rest of the country.

The problem is that shifts in partisan lean are zero-sum games: If one state moves rightward relative to the popular vote, some other state or states has/have to move leftward. (If you doubt me, try constructing a scenario in which all states are to the left of the national vote.) To account for what we see, some have hypothesized that Romney is simply over-performing in the blue states and blowing the roof off in the Southern red states, both of which are untouched by Obama’s ground game.

But this theory immediately runs smack into the face of some inconvenient facts. For starters, Florida, North Carolina, and Virginia collectively represent about one-third of the population of the South, so if the “state poll view” of the races is correct, then it becomes difficult for Romney to run up huge margins in the South. He would have to really be doing well in places like Texas, Mississippi and Georgia.

But this doesn’t square with the reality of the Deep South with Obama atop the ticket. Georgia and South Carolina were single-digit races in 2008, and the latest poll has Romney up only eight in the Peach State -- just a three-point improvement over McCain’s 2008 showing. In places like Mississippi and Alabama, the GOP likely maxed out its potential vote share against Obama in 2008; McCain won 88 percent of Mississippi whites while Obama won 98 percent of the state’s blacks. There simply isn’t much room for the GOP to grow under those circumstances. Texas? McCain won 73 percent of the white vote. So unless the GOP makes big gains among Latinos or turns Austin red, Texas is unlikely to be a vote sink for Romney. There are only so many votes the remaining states, like Tennessee and Kentucky, can provide.

But rather than provide anecdotal evidence, we can reverse-engineer a national poll from the state polls, and compare that to actual national polling. Since the national vote is a collection of state votes, polls of all states should collectively approximate the national vote (since errors should be randomly distributed, they should cancel out). This is done by a simple weighted average. First, I took the states with RCP Averages or, if available, polling from October. This provides data for 31 states. By keeping the poll data post-Oct. 4, our data set is confined to a fairly stable period in the race.

For the 19 remaining states, I assumed Romney would do as well as George W. Bush, who won nationally in 2004 by a little over two points. This is probably a fairly generous estimate for how Romney will do in these states, since I find it highly unlikely that he will match Bush’s 20-point win in Mississippi, due to higher minority turnout.

There were two exceptions to this approach: I downgraded Romney a few points from Bush's showing in Texas, and upgraded Obama a few points from Kerry's Hawaii performance, owing to the “home-state advantages” the former now lacks and the latter now holds.

I took these percentages and multiplied them by the actual number of votes cast in each state in 2008. This produced an expected vote total for Obama and Romney in each state under current polling (or, if not polls were available, under Bush’s 2004 vote shares).


After adding the totals up, the results were plain: If the state polls are right, even assuming Romney performs as well as Bush did in the states without polling, Obama should lead by 1.18 points in the national vote. Given the high collective samples in both the state and national polling, this is almost certainly a statistically significant difference. It’s also a larger margin than all but one of the polls in the national RCP Average presently show.

But what if my assumptions about the states without polling are incorrect? To double-check this, I turned to Drew Linzer’s “Votamatic” model. It provides estimates for all 50 states. While some of these seem a bit off (I would bet $10,000 of Mitt Romney’s money that he will win Tennessee by more than 10 points), it still gives us a nice uniform data set. The result: When weighted by 2008 voting patterns, these data suggest that Romney should lose the popular vote by 2.5 points -- more than any national poll is presently showing.

Again, this makes sense given what we see in the state polling. If Romney loses big in California, New York and Illinois, and loses or is effectively tied in Florida, Pennsylvania, Ohio, and Michigan, we’ve accounted for seven of the 10 largest states, comprising about 40 percent of the U.S. population. The gains he would have to roll up in places like Texas, North Carolina, and Georgia to offset this simply aren’t possible, given the realities of minority voting patterns in those states.

As a further check, I used Bush ’04 vote totals in the states, rather than 2008, and the numbers didn’t change appreciably.

But what if there is a huge surge in Republican voting? Well, this would be felt across all states, at least somewhat, and should be turning the purple states red. But regardless, I took Linzer’s model, assumed a 20 percent increase in turnout in the red states (those projected to go for Romney by more than 8 percent), and assumed a 6 percent net increase in Romney’s vote share over the baseline projection. The result? Romney wins the popular vote by 0.7 points, still less than the current RCP Average.

You can poke holes in this model, to be sure, but I think the simplest explanation is that the state and national polls really are saying different things, at least for now. In other words, if you are calling for the state polls to be right, you are pretty much necessarily calling for the national polls to be wrong, and vice versa.

How do we resolve this? Which will be correct? My best answer is “I don’t know; it is a source of uncertainty in projecting the election.” I suspect one group of polls will converge upon the other in the next week, and we’ll get a better idea.

After all, there are several good arguments for favoring the state polling: (1) you have more polls -- a much larger collective “n”; (2) you compartmentalize sampling issues -- pollsters focused exclusively on Colorado, for example, seem less likely to overlook downscale Latinos than pollsters with a national focus; and (3) the state pollsters were better in 1996 and 2000, two years that the national pollsters missed (although the truly final national pollsters in 2000 got it right, suggesting that perhaps there was a late shift in the race).

But this is by no means a cut-and-dried case. Among national pollsters, you have a battle-tested group with a long track record performing national polls. Of the 14 pollsters producing national surveys in October, all but three were doing the same in 2004 (although AP used Ipsos as its pollster that year rather than GfK, and I believe a few others may have changed their data-collection companies). Of the 14 pollsters surveying Ohio in October, only four did so in 2004 (five if you count CNN/USAToday/Gallup and CNN/Opinion Research as the same poll).

Pollsters such as ABC/Washington Post, Gallup, Pew, Battleground, and NBC/WSJ are well-funded, well-staffed organizations. It’s not immediately obvious why the Gravises, Purple Strategies and Marists of the world should be trusted as much as them, let alone more. And since virtually none of the present state pollsters were around in 1996 or 2000 (except Rasmussen Reports, which had a terrible year in 2000 and has since overhauled its methodology), it’s even less clear why we should now defer to state poll performance based upon those years.

Finally, remember that in 2008, the national polls were pretty much spot-on; the state polls were off by a couple of points. Although they actually tended to favor McCain that year, it’s just further proof that the larger “n” isn’t a guarantee of greater accuracy.

For now, I think the best thing to do is wait, and to remember that there is probably more uncertainty in this election than partisans on either side would care to admit.

UPDATE: A few people have emailed to ask if we shouldn’t just average the national and state polls. Obviously, we at RCP love averages. But you have to be careful with them. Example: If you say the capital of the United States is Washington, D.C., and I say it is New York, the best course of action is not to compromise and call it Philadelphia. Similarly, we have enough national pollsters consistently saying one thing, and enough state pollsters saying another, that I’m fairly sure the “right” answer won’t be in the middle. It really is a “two men enter, one man leaves” situation. 

‹‹Previous Page |1 | 2 |

Sean Trende is senior elections analyst for RealClearPolitics. He is a co-author of the 2014 Almanac of American Politics and author of The Lost Majority. He can be reached at strende@realclearpolitics.com. Follow him on Twitter @SeanTrende.

Mitt Romney for Mayor
Carl M. Cannon · November 16, 2014
A President Who Is Hearing Things
Richard Benedetto · November 12, 2014
Bret Stephens' Call for Robust U.S. Foreign Policy
Peter Berkowitz · November 16, 2014

Latest On Twitter