It ain’t over. (But the fat lady is warming up.)

September 24, 2012



Listen to the broadcast of Decision 2012 with Paul Gleiser, Monday, September 24, 2012.

If you believe the polls, Mitt Romney has lost the election.

Here are the troubling numbers.

• In the Real Clear Politics Average of national polls, Mitt Romney is down by 3.7 points.
• In Ohio, perennially the must-have state for Republican presidential candidates, Romney is down by 4.1.
• In Florida, the swing state that cost both Al Gore and John Kerry the presidency, Romney is down by 1.5.
• In Virginia, Iowa, Colorado, Wisconsin and Nevada, Romney trails by an average of 3.7.
• Of the twelve identified swing states, Obama leads in all but Missouri and North Carolina, giving him a 131 to 25 vote Electoral College advantage.

Thus, according to the polls and the reporting thereon, the race is over.

But two questions must be asked. First, do the polls accurately reflect the sentiments of the electorate at large? And second, do today’s poll numbers reflect what will actually happen on November 6 or will something intervene to change them between now and then?

According to Dick Morris, (and he is not alone, there are others), to the first question — Do the polls accurately reflect the electorate at large? –the answer is no. Morris cites the fact that, in keeping with past practice, almost all of the major polling organizations are using some variant of the 2008 voter turnout as the model for weighting respondents to 2012 polls. The only exception to this is Rasmussen (and Rasmussen has Romney tied or holding a slight lead).

Weighting is important.

It is not feasible to create a polling sample that accurately reflects the actual distribution of age, party preference and demographic characteristics across the entire population. Some demographic groups are relatively easy to contact by telephone while some others are not.

White elderly people, for example, are far more likely to cooperate with a polling organization call than a young African-American. Thus, the responses to polls by elderly white people are given less weight and responses by young African-Americans are given more weight. How much weight, more or less, is determined by each group’s proportionate likelihood to vote.

For most of the current polls, the model for determining that likelihood is being drawn from the actual data of the 2008 election.

However, anecdotal evidence and the actual data from 2008 suggest that using 2008 data exclusively as the model for sample weighting in 2012 is problematic.

As nearly everyone agrees, the 2008 campaign was historic. The confluence of having the first-ever African-American on a presidential ballot, the fatigue arising from eight years of the Bush presidency, the freshness of an Obama campaign unsullied by an actual record and the September 2008 financial meltdown; all served to significantly skew voter turnout.

Two examples. Black voter turnout was up by 27 percent in 2008 as compared to the averages of past elections. Turnout by college students nearly doubled.

The anecdotal evidence, supported by the fact that Obama’s campaign rally crowds are down sharply from 2008, suggests that these groups, disillusioned to one degree or another by high unemployment and poor post-graduation prospects, will not turn out in the kinds of numbers they did in 2008.

As to question two — Will something intervene to change the polls? — the answer is, ‘probably. Two things actually.

The first is Mitt Romney’s TV ads. By this time in 2008, John McCain was essentially broke and he was off the air in most of the key states. Such does not afflict the Romney campaign.

The poll numbers you see from the 12 swing states come after the Obama campaign has been on the air for weeks carpet-bombing Romney with negative ads. The fact that the average Romney deficit is yet less than four points perhaps represents the worst return-on-investment in television advertising history.

Unlike the McCain campaign of late September 2008, Romney has plenty of money in the bank and the money continues to roll in. Romney has just now placed his television buys in the swing states ahead of early voting that begins as soon as September 27. McCain, to his detriment, had essentially no money to spend on early voting. If only the votes cast on election day had been counted, McCain would have won Florida, Colorado, North Carolina and Iowa.

If Romney’s television ads, which are just now breaking, address voters’ economic fears and offer an alternative vision to the bleakness of the country’s current economic situation, it is likely that many voters, in the privacy of casting their votes, will move in Romney’s favor.

The second potential intervening factor is the debates. Obama has real reason to be fearful here. He is no doubt aware that the 1980 race between Jimmy Carter and Ronald Reagan was essentially tied or showing Carter with a lead until the debates. Carter’s record in 1980 was every bit as shaky as Obama’s is in 2012. Unemployment topped 10 percent, interest rates were sky-high, the economy was in the mud and the United States was being embarrassed by radicals holding U.S. embassy personnel hostage in Iran. Reagan was effective in the debates in putting Carter on defense with respect to his record.

Therein lies opportunity for Romney. Romney has more than 20 debates under his belt from the spring and summer primaries. Obama hasn’t been in a debate in four years . If Romney can get on offense against Obama’s record, it’s advantage Romney. The result will very likely change the polling data in Romney’s favor.

With all of this said, Romney supporters are nevertheless right to be concerned. Obama is an unusually weak incumbent, with practically nothing from his record or current events breaking his way, and yet he leads in the polls.

But if concern is warranted, despair is not.

Much still stands in the way of an Obama victory. Something the Obama campaign no doubt understands even if his cheerleaders in the media do not.