The Most (and Least) Accurate Polls

As we head into the final stretch of the campaign, we're likely to be inundated by a spate of poll results from all kinds of different organizations, and if, well, everything is any indication, then different pollsters are likely to show very different things about the state of the race in key swing states. There may be no point in trying to make sense of mind-bogglingly contradictory survey results, but in case anyone is interested, I took a look at the polling done in the run-up to the 2010 midterms and constructed a little table on which pollsters came the closest to foretelling the election results in states that are now considered "battleground" territory and which missed the mark.

I measured polling accuracy by looking at the results of the last public poll of each race by each pollster and comparing the margin between the two highest-polling candidates (sometimes the only two candidates) to the eventual winner’s margin of victory over the second-place finisher. If there was more than one poll tied for most or least accurate by this metric, then I compared each candidate’s individual poll position in the survey results with their eventual share of the final vote. This is why PPP is listed as having the most accurate poll of the Florida governor’s race despite the fact that their final pre election poll had Alex Sink (D) leading Rick Scott (R), the eventual victor, 48% to 47%. This was rated "most accurate" notwithstanding Scott’s victory because it underestimated Scott’s eventual share of the vote (48.9%) by only 1.9 percentage points, while overestimating Sink’s vote share (47.7%) by only 0.3 points. The next-most accurate survey of this race was conducted by Rasmussen Reports, which had Scott leading by three points heading into Election Day.

Race	Most Accurate Poll	Least Accurate Poll
Ohio Gov.	Columbus Dispatch*	CNN/Time
Pennsylvania Gov.	Rasmussen Reports	Franklin & Marshall
Florida Gov.	PPP (D)	Suffolk University
Colorado Gov.	CNN/Time	Magellan Strategies (R)
Colorado Sen.	Denver Post/SurveyUSA	McClatchy/Marist
Florida Sen.	Rasmussen Reports	TCPalm.com/Zogby
Iowa Sen.	Des Moines Register	KCCI/Research 2000
Nevada Sen.	Suffolk University	Rasmussen Reports
New Hampshire Sen.	WMUR/UNH	ARG
North Carolina Sen.	PPP (D)	SurveyUSA
Ohio Sen.	PPP (D)*	Suffolk University
Pennsylvania Sen.	Susquehanna	McClatchy/Marist
Wisconsin Sen.	McClatchy/Marist Rasmussen Reports	PPP (D)
Wisconsin Gov.	McClatchy/Marist	Rasmussen Reports Reuters/Ipsos
Nevada Gov.	PPP (D) Suffolk University	CNN/Time
Iowa Gov.	KCCI/Research 2000	Rasmussen Reports
New Mexico Gov.	Daily Kos/PPP (D)	SurveyUSA
Michigan Gov.	EPIC-MRA	Rossman/Team TelCom
New Hampshire Gov.	WMUR/UNH	ARG

*Nailed it! (Poll results conformed exactly to election results.)

Of course, not all these races were as close as the presidential race is this year. The Senate races in Florida, Iowa and New Hampshire, and the gubernatorial elections in Colorado and Michigan, were all blowouts. Take that however you like. Also notice that the above table does not include any data from Virginia, which had neither a U.S. Senate race nor a gubernatorial election in 2010. In the 2009 race for governor there, SurveyUSA was right on the money; its final poll before Election Day showed Bob McDonnell (R) leading his Democratic opponent, Creigh Deeds, by eighteen points 58% to 40%. McDonnell won that election with 59% of the vote to 41% for Deeds. That was three years ago, however, so it may not be worth much.

Finally, there are a handful of states that are assuredly not up for grabs in the presidential election but are experiencing heated U.S. Senate and/or gubernatorial races. For those who are interested, here are my findings on polls of 2010 races in those states.

Race	Most Accurate Poll	Least Accurate Poll
Connecticut Governor	CT Capitol Report/MRG	Suffolk University*
U.S. Senate – Connecticut	PPP (D)	Rasmussen Reports
Maine­ Governor	PineTreePolitics/MECPO	Critical Insights
Massachusetts Governor	Suffolk/7News	Rasmussen Reports
U.S. Senate – Washington	FOX News/POR-Rasmussen	PPP (D)

PPP Uses D+8 Sample to Show Obama Up 1 in Ohio

I had to smile on Saturday when I saw that a Public Policy Polling survey showed Mitt Romney within one point of President Obama in Ohio. It’s not just that a Democratic polling firm says it’s a one-point race in what seems to have become Obama’s one-state firewall, although that is significant; the sample breakdown indicates that their sample consisted of 42% Democrats and only 34% Republicans. This eight-point Democratic advantage is consistent with 2008 exit poll results from the Buckeye State, but given that both parties have lost ground among registered voters, who increasingly identify as "independent", it’s difficult to understand why PPP (or any pollster, for that matter) would survey a greater percentage of political partisans and a smaller percentage of self-described independents when compared to the 2008 electorate.

In other polling news, four national polls out this morning all show Romney leading Obama. He's up by three points in the latest Monmouth University poll, 48% to 45% for Obama. The Rasmussen Tracking Poll and Politico/GWU/Battleground Poll both have him ahead by two, 49%-47%. Gallup's tracking poll continues to give Romney his largest lead, 51 to 45 percent. It would be easy to dismiss that as an outlier, but in addition to being the most experienced and reputable outfit in this field, Gallup also has a pretty good track record in recent elections. On October 20, 2008, they showed Obama leading McCain by seven points among likely voters; he won the election by seven points nationally. On October 24, 2004, they had George W. Bush leading John Kerry, 51% to 46%. President Bush won re-election with 51% of the popular vote; Kerry received 48%. So, despite the naysayers trying to discredit this enduring monument of polling, recent history indicates that their presidential election polls conducted around this time in elections past closely line up with the eventual outcome. 

What should really trouble the Obama campaign and their supporters is the trend in some of these polls (or, in the case of Rasmussen, which hasn't shown the president leading Romney in two weeks, the lack of one). The Governor's two-point lead in the Politico poll represents a three-point swing from last week, when they had the president up by one point, 49% to 48%. Monmouth's poll shows that Romney is up a point and Obama has dropped a point from their last survey, which showed the Republican leading, 47% to 46%. The candidates are tied nationally in the latest NBC News/Wall Street Journal poll at 47 percent, but that's a decline in Obama's poll position from their previous national survey, which had the president ahead with 49% of the vote to 46% for Mitt Romney. Not only does this show the lack of any measurable "bounce" from the second debate for Obama, but it crystallizes his dismal (for an incumbent) poll position, well below 50% with fifteen days to go before Election Day and early voting already underway.

Oh, OH...

If you haven't seen last night's debate between Vice President Biden and Rep. Paul Ryan (R-WI), then, well, I don't really care. Frankly, if you're like me and already know who you're voting for but tuned in hoping to witness an epic clash between a skilled, experienced debater with a propensity toward making gaffes and a policy wonk engaging in his first one-on-one, nationally televised debate, then you probably were disappointed. I don't care to talk about it; instead, I want to focus on the presidential race in that most critical of battlegrounds: Ohio.

Back in June, I made a short video about the 2012 presidential campaign in Ohio and highlighted what I though were each candidates' strengths and weaknesses as they tried to sway voters in the state. Long story short: it was a toss-up then, and it's a toss-up now.

There was a time, though, when it looked like Obama might be running away with Ohio, leaving the Romney campaign trying to fathom a realistic path to victory without what has long been the sine quo non of a Republican electoral majority. From Labor Day through the first presidential debate last Wednesday, every public poll of the presidential race in the Buckeye State showed Obama leading. All the pollsters who had been regularly polling the state had Romney losing ground to the president, who had begun to crack 50% in some surveys.

According to polls released in the last week, however, the race in Ohio is back to a dead heat. Five different pollsters--WeAskAmerica, ARG, SurveyUSA, Gravis Marketing and Rasmussen Reports--published survey results showing the presidential candidates within a single point of each other in the state, a statistical dead heat. Given this consistency among reputable independent polling organizations, a Marist poll of Ohio voters conducted for NBC and the Wall Street Journal that has President Obama leading Governor Romney by six points, 51%-45%, deserves some skepticism. But this even without these five large grains of salt, the Marist poll should be difficult to digest. Look at the crosstabs, and you'll find that the sample of likely voters was 40% Democrats and only 29% Republicans. To put that in perspective, exit polls indicated that, in 2008 (the historical zenith of Democratic turnout in modern times), 39% of voters in Ohio identified as Democrats, 31% as Republicans, and the rest were independents. So the question this poll begs is: Do you really expect the Democrats' turnout edge over Republicans this year to be 22% greater in Ohio than it was in 2008, especially considering there are now 67,000 more registered Republicans than Democrats in the state?

Even with the Marist poll (and a CNN/Opinion Research poll showing Obama up four points in the state), the president's advantage in the RealClearPolitics average of the seven most recent Ohio polls is only 1.3 percentage points. That's down from 5.5 percentage points just eight days ago.

If accurate, then these numbers represent a statistical anamoly: Obama is running ahead of his national poll position in Ohio. As of today, the president trails Mitt Romney in the RCP average of national polls by one percentage point. Yet, in 2008, he won the Buckeye State by a 4.6-point margin, well less than his 7.3-point margin of victory over John McCain in the nationwide popular vote. In fact, in the past nine presidential elections, the Republican candidate has received a greater share of the vote in Ohio than he has nationally. Are we to expect that, with no advantage in party identification and no recent shift toward their side in statewide elections (Republicans pretty much ran the table in 2010.), the Democrats are poised to pull off what would be a historical aberration in Ohio? 

The troublesome sign for Romney is that he has yet to break 50% in the state. Obama may have a lower trough, but he also seems to have the higher ceiling.

Still, with the race this close, candidates doubt polls at their own peril, and that's probably why both candidates are treating Ohio like it's the key to victory. Romney and Ryan appeared together at a rally in Lancaster today; it was the second time in three days the former Massachusetts governor had campaigned in the state. Obama hasn't been in Ohio since Monday but has made more visits to the state this year than Romney has.

Romney Rising

Mitt Romney greets audience members at a campaign rally in St.Petersburg, Florida on Friday. (Brian Snyder/Reuters)

When Scott Pelley of CBS asked him how he would turn this campaign around, Mitt Romney confidently told him, "Well, it doesn't need a turnaround."

Whether the Republican nominee for president needed one or not, things certainly have turned around in the last few days. Romney went from trailing Obama by two points, 47% to 49%, in the Rasmussen tracking poll to leading him by the same margin. He's cut the president's lead to two points in the latest Politico/GWU/Battleground survey, and the two are tied at 47% in National Journal's Congressional Connection poll. Last month, Obama led Romney by seven points in their survey.

One of the most interesting findings to note is that, for the first time that I've noticed, in both the Gallup and Rasmussen tracking polls, Obama's poll position is lower than his approval rating. That can either be good news for the president (because it shows he has room to grow) or for Romney (because it shows that even voters who approve of the president's job performance aren't supporting him). 

Of course, the national popular vote won't decide the election, but state polls show the Romney-Ryan campaign surging in crucial territory.

In Florida, President Obama's lead over Mitt Romney is down to just one point in the NBC/Wall Street Journal/Marist survey; last month, they had the president up five points, 49%-44%. Other pollsters have the Republican ticket opening up a lead in the Sunshine State. Romney leads Obama by two points, 49%-47%, according to a survey conducted on Thursday by Rasmussen Reports. Last month, Rasmussen had Obama ahead in the state, 48%-46%. WeAskAmerica finds an even larger shift toward the Republican ticket: their latest poll of Florida voters shows Romney leading Obama by three points, 49%-46%; that mirrors the results of their Florida poll last month, which found the president on top by the same margin. 

In Virginia, the shift is even more stark. A Marist poll conducted for NBC and the Wall Street Journal finds Romney trailing the president by two points, 46%-48%, but that's narrower than the five-point lead Obama enjoyed in their last survey of Virginia voters, conducted less than a month ago. Romney leads Obama in the commonwealth, 49%-48%, according to Rasmussen Reports; that result mirrors the state of the race in their last poll of Virginians, which had Obama up by the same margin. We Ask America shows a six-point swing toward Romney in Old Dominion; he leads the president by three in their latest poll; last month, they had Obama up three points on Romney in Virginia.

Out west, the poll numbers are more muddled. Romney leads the president by three points in Colorado, 49%-46%, according to a Gravis Marketing survey; last month, they had Obama ahead in the state, 50%-46%. A poll conducted by the University of Denver, however, gives the president a four-point lead among Colorado voters, though we don't have an earlier survey they've conducted to compare those results to. In Nevada, where the president has enjoyed a small but consistent lead all year, Gravis Marketing has him ahead by just one point, 49%-48%. The RealClearPolitics average of the latest Nevada polls pegs Obama's lead in the state at 4.6 percentage points. 

Finally, in the all-important swing state of Ohio, it's a one-point race, according to American Research Group, Rasmussen Reports and WeAskAmerica. ARG and WeAskAmerica both have Romney up one point in the Buckeye State, while Rasmussen reported that Romney was down one point there. Last month, Obama led Romney by one point in an ARG survey of Ohio voters, and other surveys gave him as much as a ten-point lead in the state.

It would be easy to credit this marked surge in the Republican ticket's poll numbers to Mitt Romney's rout of Obama in the first presidential debate, but nearly all of the aforementioned polls were conducted, in whole or in part, before Wednesday night's debate. The conclusion is inescapable: the Romney surge began before he cleaned the president's clock last Wednesday in Denver.

What could explain this sudden shift in momentum in a race that, let's be honest, appeared to be slipping away from Romney as recently as last weekend? Perhaps the Libya fiasco is finally taking its toll on the president, but then why hasn't his approval rating dropped along with his poll numbers? Maybe the fundamentals of a race that should have been Romney's to lose all along are finally catching up with Obama, but if that's the case, then what took so long? Maybe those TV ads that feature Romney just sitting down and talking into the camera are resonating with voters in a way that those earlier, business-presentationy commercials with the fancy graphics didn't. Whatever the cause, it now appears to be the Obama campaign that's in desperate need of a turnaround.

Obama’s Ad Blitz Failing in Iowa and Out West, Working Elsewhere

President Obama has spent more money on ads per electoral vote in Iowa than any other state, but it doesn’t seem to be paying off.

The latest poll results out of the Hawkeye State show Mitt Romney leading the president, 47% to 44%. That’s within the margin of error, but if you’re an incumbent trailing your opponent by three points, then it also has to be worrisome.

Following the Republican National Convention, the Romney-Ryan campaign hit Iowa with this ad:

As you can see/hear, the ad is not very good. I think this, however, is a better ad:

Just imagine what effect more ads like this one would have on the polls. 

The Desert West Deserts Obama

In 2008, President Obama won Colorado by nine percentage points and Nevada by twelve points. Polls have shown the presidential race this year to be a dead heat in both states. Rasmussen Reports shows Romney with a two-point lead in Colorado, 47%-45%, while the president leads him by the same margin in Nevada. (The president has a one-point lead in Colorado, according to a Quinnipiac poll, and a CNN/Opinion Research survey puts his lead at three points in the Silver State, 49%-46%,.) Nevada is a particularly interesting case because, while polls have shown the president with a consistent lead in the state for months, he can't seem to break 50%. In any other state, this would portend inescapably bad news for the incumbent in what is essentially a two-way race, but Nevada is unique, in that it allows voters to cast their ballot for "None of these candidates." Thus, Obama's sub-50% approval rating need not be drag on his campaign if he can convince enough voters that Romney is an unacceptable alternative.

There's a lot of stupid people out east.

The money the president's campaign has been pouring into other swing states has yielded more appreciable dividends, however; he's opened up a small but obvious lead over Romney in Ohio and Virginia, and Michigan--which once looked like a swing state--now appears to be out of play. The Obama campaign has not been spending a lot of money in Michigan, but outside groups and labor unions have been covering for them, and the polls indicate a marked shift toward Democrats. The Detroit News poll has Obama up leading Romney by 14 points in Republican's the native state; last month, that lead was six points. EPIC-MRA also shows the president widening his lead over Mitt Romney in Michigan; late last month, their poll showed Obama ahead by just three points, 49%-46%, but their latest poll of likely voters in Michigan gives the incumbent a ten-point lead, 47%-37%. (Note that both candidates have lost support, though Romney has lost a lot more than Obama.) Republicans aren't convinced Michigan is a lost cause, however; Restore Our Future, the conservative super PAC that boosted Romney in the primaries, will be running ads in several media markets throughout the state through October 2nd. 

I'll cover Ohio and Virginia in a separate post, but I wanted to also mention that the presidential race remains a dead heat in Florida, which has expectedly been the largest recipient of campaign spending this election cycle, and of course, Obama's campaign has outspent Romney's, while pro-Romney/anti-Obama groups have slightly outspent pro-Obama/anti-Romney groups on the airwaves. With less than seven weeks to go before election Day and early voting already underway, the question becomes whether Romney, Ryan and their allies can overpower the president's machine (which I'll be posting about next week), especially with voters in some swing states feeling overwhelmed by all the attention focused on them. As November 6th draws closer, the effectiveness of campaign ads becomes more and more limited. Let's hope the Romney campaign makes the most of the next few weeks.

Pols & Polls: Scott Walker, John Kasich, Rick Scott are "No longer the least popular people in politics" - The Washington Post

Yesterday, Aaron Blake reported for the Washington Post on a cluster of recent polls from what he called "the nine crucial swing states The Fix has identified as key to the 2012 election". Here's the headline: Walker, Kasich and Scott--all first-term Republican governors--have seen their poll numbers improve after  voters in Wisconsin, Ohio and Florida, respectively, turned on them for administering the harsh medicine necessary to improve their states' business climates and get their fiscal houses in order. Kasich and Scott are each at 41% job approval--not stellar, but a significant rebound from where they were a year ago--and Walker, who survived a recall election earlier this month, is enjoying a 51% approval rating, according to Marquette University.

A little over a year ago, I blogged about how several newly elected GOP govs, including Walker, Kasich and Scott, were having a rough go of it with a seemingly large chunk of the voters who had just elected them. I pointed out that Indiana Gov. Mitch Daniels had also drawn the ire of voters during his first couple years in office, but once the effects of his policies were felt, his poll numbers turned around; he was handily reelected in 2008 and is now one of the most popular governors in the country.

Blake may have buried the lead, however, as his post also lists The Fix's latest "Line" of gubernatorial races (a countdown of the five governorships most likely to change parties this year). Every race on the Line is in a state with a Democratic governor, meaning there is virtually no chance of Democrats winning the governor's mansion in a state with a Republican gov this year, but Republicans are poised to flip a handful of governorships in states where Democrats are retiring.