You are currently viewing the printable version of this article, to return to the normal page, please click here.

YOUNG: Skewed polls not telling the full Romney story

Republican could be significantly ahead in the race

- - Wednesday, October 24, 2012

The latest skewed presidential poll recalls Reagan's admonition: "There you go again." Once more, Republicans are being under-polled, and a close race is actually even closer than it already appears. The result could be that instead of an "October Surprise," this year's election might offer a November one -- on election night.

The latest ABC News/ Washington Post poll (923 likely voters, MOE +/-3.2%) released Oct. 15, showed a continued close race, with President Obama leading Gov. Romney 49 percent to 46 percent. The problem is it also shows a very skewed party sampling of its respondents.

While Democrats made up 35 percent and Independents made up 33 percent of respondents, Republicans made up just 26 percent. If that breakdown actually reflected what we have seen in recent elections, there would be no issue. It is not even close, however, especially when it comes to Republicans.

According to national exit polls (Edison Media Research/Mitofsky Int'l) from the two Obama-influenced elections -- 2008 and 2010 -- Democrats averaged 37.5 percent, Independents 28.5 percent, and Republicans 34 percent. Compared to those averages, the ABC News/ Washington Post poll actually underrepresents Democrats too -- by 2.5 percentage points -- but it is Republicans who are the real losers here, by an enormous 8 percentage points.

Why such a skewing happened can be open to debate. What cannot be open to debate is that this kind of skewing has a real effect on the poll's outcome -- and even more importantly, on its accuracy and ultimately, the electorate.

Again, looking at the exit polls from the two "Obama elections," partisans broke heavily for their party's candidate. On average, 91.5 percent of Democratic and Republican voters supported Democrat and Republican candidates, respectively. Just 7 percent of each went for the opposing party.

The real fluidity in both elections came from Independents and their shifting support of the two parties -- going 51 percent for Democrats in 2008 and 56 percent for Republicans in 2010.

Interesting, yes, but what does that mean for the poll in question? A lot.

If both Republicans and Democrats are undercounted in the sample and continue to break heavily for their party, then both are losing some support-rich respondents. Democrats lose 1.8 percentage points; however, Republicans lose a whopping 7.3 percentage points. If you think that's inconsequential consider: That 5.5 percentage point difference is only slightly less than the margin (6.3 percent) by which Mr. Obama beat John McCain four years ago.

Of course, you have to factor out the support each would get from the poll's over-represented Independents (and how much support each party will get from this fluid faction is certainly one of this election's biggest questions), but the point remains: This election is indeed close, but it's nominal leader appears to be Mr. Romney.

That conclusion is not just based on an under-representing of Republicans either. For Mr. Obama to just have a 3 percentage point lead in the poll, despite its heavy under-inclusion of Republicans, means Independents must be going for Mr. Romney as well. Nor should this be surprising based on the "Obama elections" exit polls -- Independents went heavily Republican in 2010 and far less so for Democrats in 2008, so a Republican advantage would not be surprising.

Granted, a larger pool of Republicans and the presumption that Independents are breaking for Mr. Romney likely does not mean that the poll's adjusted results would be outside the poll's margin of error. As any statistician will tell you, that means the race is still statistically tied. However, as any political analyst will tell you: Nominal leads still matter a great deal.

Americans love competition -- that's why we call our elections "races" -- and that means we want to know who is winning. Inevitably, people follow the leader first. For that reason, the money and the media coverage follow the leader, too.

Leading can beget winning. People want to feel their vote "mattered." The best way to feel your vote mattered is to have voted for the winner -- that's why so few third party candidates have a real electoral impact. It's also the reason the media refuses to call winners until the polls have closed. The perception of winning can mean winning.

Undoubtedly, Mr. Romney's debate performance has changed the way the electorate perceives this campaign. His closing the gap in the polls creates further momentum. Lurking beneath the polls' surface, however, may be the biggest bump of them all: the fact that he has always been closer than he appeared and now may well be leading.

J.T. Young served in the Treasury Department, the Office of Management and Budget and as a congressional staff member.