What the Polls are Really Saying

Republican lobbyist Mike Barbera contributes a guest column today, sharing his views about the presidential election. On Friday, we will feature a piece reflecting the Democratic perspective.

By Michael Barbera

It doesn’t seem to make any sense …

The American economy is burdened with sustained unemployment at levels never before seen in American history. Gas prices are through the roof. The budget deficit and national debt are both sky-high. Household incomes are down, and the housing market is still decimated all over the country. And that’s just the news here at home. Overseas, our embassies are attacked, our diplomats assassinated and our soldiers in Afghanistan are slaughtered by our so-called “allies.”

And yet according to the polls, until his poor debate performance in the first debate, Barack Obama was winning the race for the White House. In some polls he was ahead narrowly, in others he was ahead comfortably — but he was always ahead. How can that be right?

The simple fact is, it’s not. The president may or may not be ahead, but he is not winning. In fact, anyone who looks at the numbers closely realizes that the president’s electoral standing is perilous at best.

Why is it that so many talking heads have spent much time telling us that the president was winning? Quite simply, most of them read polls in a lazy and simplistic manner. Most observers only focus on the “horse race” — who is ahead and by how much. And for most of 2012, the president was ahead. So pundits simply regurgitate the numbers and pretend they know something.

At this stage of the race, horse race reporting doesn’t tell us very much. It doesn’t matter if Pres. Barack Obama leads former Massachusetts Gov. Mitt Romney by three points in Ohio or by six in Wisconsin, or whether he is ahead by one point in the Gallup poll or by five in the latest poll by the AP.

Here is what matters: how far is Pres. Obama from getting over 50 percent of the vote? And the answer, based on any fair reading of the evidence, is that he has a lot of work to do.

Look at the Real Clear Politics averages in the key battleground states. Even before Romney’s recent surge, the president was stuck between 47.1 percent and 48.8 percent in Florida, Virginia, Ohio, North Carolina, Iowa and Colorado. Karl Rove reported in the Wall Street Journal the day before the Denver debate that in 91 recent national polls “Mr. Obama was at or above the magic number of 50 percent in just 20. His average was 47.9 percent. Mr. Romney’s was 45.5 percent.”

The polls have now shifted after Romney’s strong debate performance and the president’s weak one in the first debate. The challenger cut into the president’s lead and in some cases erased it altogether.

That brings us to another popular myth about the polls — that they are constantly shifting. In fact, the president’s numbers have remained very steady. With the exception of a slight spike after the Democratic convention, the president is where he has always been — in the mid-to-high 40s. He has a solid base of loyalists, but he has not yet closed the deal with most voters — despite dominating the airwaves. According to the National Journal, as of Oct. 1, the Obama campaign spent a total of $285 million on broadcast and cable television, in addition to radio advertising, while the Romney camp spent just $117 million — and still the president was well-below the magical 50 percent mark. An incumbent stuck below 50 percent in October is an incumbent in trouble.

What has changed throughout the course of the campaign is the level of Romney’s support. This is no surprise. Like it or not, the American people have seen Obama in action. Most voters know whether they like him and whether they think he is doing a good job. Very few voters are truly undecided about him. Until very recently, many voters were very much undecided about Mitt Romney, however. That is why the president spent so much money over the summer on a steady drumbeat of negative advertising. The American people were not convinced by those ads, many of which were powerful. They wanted to decide for themselves about Mitt Romney, and a record 67 million people watched the first debate. Mitt Romney picked a good day to have a good day — and the president picked a bad day to have a bad one. Romney’s strong performance shook up the race, and he is almost certain to win decisively among late-deciding voters.

Many polls are also overstating the president’s support. These polls are using turnout models based on the 2008 election, which means the pollsters are betting that the electorate of 2012 will look a lot like the electorate of 2008. This is likely to be a losing bet.

The 2008 election featured the following: a historically unpopular Republican president, an economic meltdown one month before the election, an incredibly motivated Democratic base, a demoralized Republican base, record turnout of young voters, blacks and Latinos, and a sizable number of Republican crossover votes for Barack Obama. Does that sound like the electorate we have today? Yet many seemingly reputable pollsters construct their polling around 2008 models. I believe the technical polling term for this is “wishful thinking.”

If we look at the polls with a trained eye — and not simply repeat the horse race numbers fed to us by lazy pundits — it is clear that Obama is not winning. And if he isn’t winning now, he’s not likely to be ahead on Nov. 6.
_______________________

Michael Barbera is a lobbyist and consultant with the American Continental Group.

Advertisement

The VP Debate and a Changing Map

The vice presidential candidates took center stage for one night last evening, and at least the in the CNN poll (conducted by Opinion Research Council of 381 debate watchers via post-event telephone interviews), results suggested that Paul Ryan fared better than Vice President Joe Biden.

A 48-44 percent pro-Ryan division relating to winner perception was recorded, though a slight over-sampling of Republicans was present in the CNN respondent universe. Ryan scored more favorably on two critical perception points, being more likable (53-43 percent) and better expressing himself (50-41 percent). Of those sampled, 28 percent said they are now more likely to vote for Mitt Romney because of this debate, versus 21 percent who made similar comments in relation to Pres. Barack Obama.

Thus, the debate ended with neither candidate scoring a knock-out blow or inflicting serious damage upon their own campaign effort. It is likely the after-effects from this forum will be short-lived, which likely helps Romney because he has the current momentum.

Last night, Real Clear Politics also released their version of the new electoral map. According to them, based upon all polling data to which they have access, the president has a safe/likely/lean count of 201 electoral votes from 16 states and the District of Columbia, while Romney’s commensurate total is 181 from 22 states. Curiously, RCP lists Missouri in the Undecided/Uncommitted column. Based upon complete voting history since 2000, inclusive, and the fact that Romney has led in virtually every Show Me State poll, it is more than plausible to add their 10 Electoral Votes to the Romney column. If so, the challenger would pull within 10 votes of the president, leaving 146 votes in 11 states as uncommitted.

Among the states moving into the Undecided/Uncommitted category are two places with a clear Democratic history in presidential elections, Michigan and Pennsylvania. Once the votes are completely counted, it is more than likely that both of these domains will remain in the president’s column but the fact that he is having to fight so hard to keep them is indicative of his weakening national standing.

The remaining nine states: Florida, North Carolina, Virginia, Ohio, New Hampshire, Wisconsin, Iowa, Colorado and Nevada are the ones we have been analyzing for some time. It is clear that some combination of these states will make the final electoral determination in early November.

Assuming the moving of Missouri to Romney and Michigan and Pennsylvania to Obama prove correct, then the Republican would need 79 Electoral Votes to win the Presidency and Obama 33. Of the Undecided/Uncommitted states on the current board, the Republican’s easiest path to victory would include Florida, North Carolina, Virginia, Wisconsin, Iowa, and Nevada. This would give Romney a bare 270-268 victory and send him to the White House.

While originally it was thought that a challenger would have to carry Florida, Ohio, and Virginia to defeat Obama, it is now possible to win with taking Florida and only one of the latter two. With more states firmly in play, the Romney campaign now has several options to possibly cobble together a winning coalition of states in order to reach the magic 270 electoral vote plateau.

Getting through last night’s VP debate with both sides intact, means the stakes for the remaining two debates grows even higher. Pressure will continue to mount upon both candidates and what remains true is that the final defining events for this particular presidential election have yet to occur.

The Shifting October Winds

Poll-watching

October is generally determination month for hot-race candidates, and now that it has begun prepare for some major swings in several political venues.

Recently, with Election Day coming in less than five weeks, we have seen polling that detects significant change in several Senate campaigns, all showing a shift away from the initial leader. With the Senate up for grabs, each race becomes critical and could itself determine whether Republicans or Democrats control the body in the new Congress. Today, we isolate four such campaigns.

In Connecticut, Quinnipiac University released its latest poll (Sept. 28-Oct. 2; 1,696 likely Connecticut voters, +/- 2.0 percent error factor) that finds Democrat Chris Murphy now trailing Republican Linda McMahon 47-48 percent. This is a surprising result and much different from the 48-42 percent Murphy lead that Public Policy Polling found just a week ago (801 likely voters, +/- 3.5 percent error factor). It is important to note that both polling firms have surveyed this race repeatedly. Having the opportunity to study another reputable firm’s results in order to bring a fresh perspective might provide us a better directional indicator.

Shifting to another hot race, we turn to Massachusetts and a recent poll by Opinion Dynamics for the consulting firm Mass Insight Global Partnerships. This rather flawed poll – because the sample size is less than 350 respondents statewide and the survey period is a long five days – posts Democrat Elizabeth Warren over GOP Sen. Scott Brown by a 48-44 percent clip. This same polling firm gave Brown a large 52-42 percent lead back in January. Last week, Rasmussen Reports also released their poll showing the candidates to be in a statistical tie at 48-48 percent (Sept. 24; 500 likely voters, 4.5% +/- error factor). Conversely, that very day, the Boston Globe publicized their data projecting Warren to be holding a five-point advantage over the freshman senator, 43-38 percent (502 likely voters, 4.4% +/- error factor). With continuous polling producing conflicting results it is clear this race is going to come down to the final hours.

Wisconsin continues to provide us with a close and hard-fought Senate race. Rep. Tammy Baldwin (D-WI-2) received a measurable bump in the wake of her speech at the Democratic Convention followed by an extensive early media advertising blitz, and polling indicates that the tide has turned in her favor. At least in the short term, Rep. Baldwin is now the clear front-runner over former four-term governor Tommy Thompson (R), after he consistently posted an advantage before and after the mid-August primary. Two recent polls show Baldwin leading: the first by Marquette University Law School showing a four-point edge, 48-44 percent (894 likely voters), and Real Clear Politics revealing a five-point advantage, 49-44 percent. Republicans, at one point, thought this seat would be a clear pickup opportunity, however, it is currently trending more Democratic. This polling trend could just as quickly snap back toward Thompson once he responds to the current line of attack being put forth against him.

Joining this see-saw Senate campaign group is the open seat race in Arizona. For months, Rep. Jeff Flake (R-AZ-6) had held a significant lead in all polling against Democratic nominee Richard Carmona, the former US Surgeon General in the George W. Bush administration. After consistently gaining ground on Flake after the Aug. 28 primary, Public Policy Polling, in their Oct. 1-3 poll of 595 likely Arizona voters, puts Carmona into the lead for the first time, albeit by a scant 45-43 percent margin.

It is clear, differing from what we have witnessed in the past four election cycles, that no tsunami wave is emerging for either party this year. As we turn the corner into the homestretch of campaign 2012, it is important to monitor all momentum changes as a barometer for predicting final outcomes.

Polls In Sync

We often bring polling inconsistencies to light and analyze why a particular survey may be more accurate than another. Right now, however, four diverse pollsters all are showing virtually the same result in the presidential race. The conclusion: The presidential campaign, in mid-May, between President Obama and Mitt Romney, is a virtual tie.

Our first comparison is between the two polling firms that track the national race on a daily basis: Gallup and Rasmussen Reports. The latest published Gallup numbers (April 15-May 6; 3,050 registered voters grouped in five-day tracking segments) show Obama and Romney with the exact same level of support, 45-45 percent. Rasmussen, which calls 500 people every night and then groups them in three-day tracking segments, arrives at virtually the same conclusion as Gallup, but their data has Romney grabbing the smallest of leads, 46-45 percent.

Next comes TIPP, the regular polling partner of Investor’s Business Daily and the Christian Science Monitor. TIPP is the acronym used for the polling section of the TechnoMetrica Market Intelligence corporation, which claims the mantra of “America’s most accurate pollster” based upon their presidential track record in 2004 and 2008 as compiled by Real Clear Politics. According to their traditional polling interview method conducted during the May 9-16 period of 778 registered voters, the President has a slight 43-40 percent lead, certainly consistent in margin of error terms with the other pollsters just cited.

Finally, Zogby Analytics, which has in the past been criticized for being among the least accurate of pollsters, and in this case polling for the Washington Times, also returns current data that agree with the rest. According to their traditional live interview survey May 11-12 of 800 likely voters, Romney has the one-point lead, 44-43 percent. Zogby also went one step further. They asked this particular sampling segment a second ballot test question that included Libertarian candidate Gary Johnson, the former New Mexico Republican governor, who will likely be on the general election ballot in many states. When adding Johnson to the mix, the result slightly changed the Romney-Obama split; the returned result showed a consistent 44-43-2 percent division, but this time in the President’s favor.

What this tells us – because the results from four such different pollsters each with their own methodological approach pertaining to sample selection and interview philosophy – is that at the present time, the national race deserves an unequivocal toss-up rating.

Doing the Delegate Math: Exactly Who’s Right?

In tracking the delegate count for the Republican presidential nomination, it is clear that no “official” tabulation exists. In fact, virtually all major media organizations and political websites have different totals for Mitt Romney, Rick Santorum, Newt Gingrich and Ron Paul, and with wide divergence.

Below are the latest published delegate counts from key media and political outlets. Notice that only ABC News and the New York Times agree, which probably means they are using the same source.

   • ABC News: Romney 105; Santorum 71; Gingrich 29; Paul 18

   • CBS News: Romney 111; Santorum 44; Gingrich 30; Paul 15

   • CNN: Romney 127; Gingrich 38; Santorum 37; Paul 27

   • Fox News: Romney 107; Santorum 45; Gingrich 32; Paul 9

   • The Green Papers: Romney 107; Santorum 43; Gingrich 43; Paul 35

   • NBC News: Romney 84; Gingrich 29; Santorum 14; Paul 11

   • New York Times: Romney 105; Santorum 71; Gingrich 29; Paul 18

   • Real Clear Politics: Romney 99; Santorum 47; Gingrich 32; Paul 20

To recap, the eight entities don’t even show a consistent order of candidates – CNN and NBC have Gingrich in second place and the others project Santorum in the runner-up position. Romney ranges from a low of 84 pledged delegates (NBC) to a high of 127 (CNN). Santorum’s spread is from 14 (NBC) to 71 (ABC/NYT). Former Speaker Gingrich appears to be most consistent, tallying in a range from 29 (ABC/NYT) to 43 (The Green Papers). Finally, Rep. Paul runs the gamut from 9 (Fox) to 35 (The Green Papers).

Why the differences? First, even some states where voters have already participated – Iowa, Nevada, Missouri, Maine, Colorado and Minnesota – won’t actually apportion their delegates until convention events later in the year. Some of the aforementioned trackers are estimating what these states will eventually do based upon the public votes already cast.

Secondly, the media and political sources either are, or are not, projecting unpledged delegate votes. Many of the unpledged delegates themselves aren’t even chosen yet.

Third, at least two states’ delegations, Florida and Arizona, will likely be challenged at the Republican National Committee Convention. Both are apportioning their delegates under a winner-take-all option in defiance of RNC rules. Only states that vote after Super Tuesday (March 6) are entitled to use the winner-take-all format. Therefore, the Florida count (50 delegates for Romney) could change. The same with Arizona (29 delegates), which votes Feb. 28.

Considering that no official delegate count actually exists, it clearly means we will continue to see a very fluid situation that could lead to a surprising conclusion. Keep in mind, regardless of the apportioned delegate count’s accuracy, or lack of it, only 11 percent of the 2,286 delegates have been assigned. The mathematics continue to show that this race is still very much undecided.