RUTHFULLY YOURS

Since the start of the Democratic convention, almost all the published polls in the national presidential race between Barack Obama and Mitt Romney have shown a lead for Obama. Polls of the competitive states that will decide the winner of the Electoral College have also, by and large, favored the president.

Some conservatives have argued that many of the polls are badly distorted by oversampling Democrats. They make the argument that the oversampling is deliberate, since polls which show Obama ahead will discourage Republican voters from showing up at the polls either during early voting periods in the states which allow it or on Election Day in November. In essence, these conservatives are arguing that the pollsters, like the mainstream media, are in the bag for Obama, and doing what they can to enhance his chances of winning.

The truth of the matter probably lies somewhere between the conspiracy theorists’ vision of the real state of the race (that Romney would actually be ahead if a more reasonable sampling of Democrats and Republicans occurred), and the current polls that show Romney about 4-5 points behind.

The number of Democrats versus Republicans who are interviewed in national and state surveys is highly correlated with the outcome of a poll. The reason for this is that about nine in 10 Democrats will likely vote for Obama, and about nine in 10 Republicans are likely to vote for Romney.

The real battle for votes is among independents who may lean to one side or the other, but are not as settled in their preferences, and in turning out those from a party who are committed to the party candidate. In some of the recent polls, Romney has been ahead among independents, but trailed overall because the survey sample contained many more Democrats than Republicans.

When pollsters randomly call phone numbers, whether on landlines or cellphones, the individual who is polled is always asked for party identification. A key consideration in evaluating the poll’s reasonableness is whether the pollster adjusts the results to conform to a particular partisan breakdown (say 40% Democrats, 35% Republican, and 25% independent), or just publishes the raw results of the surveys. In 2008, the electorate consisted of 7% more Democrats than Republicans. But in 2004, when George W. Bush was re-elected, the two sides were approximately even. If a pollster is adjusting the results of his sample to fit a particular breakdown between the parties, is he using the 2008 election as his model, which will favor Obama, or the 2004 model, which will favor Romney?

One well respected polling firm, from Quinnipiac University in Connecticut, is partnering this year with The New York Times and CBS News. Last week they published polls for Florida and Ohio, two states that Mitt Romney almost has to win to be elected.

In Ohio, the poll showed Obama up 10%, and in Florida, 9%. On the face of it, these results are absurd. In 2008, a great year for Democrats, Obama won the national popular vote by just over 7 percent. But he won Florida by less than 3%, and Ohio by less than 5%. In other words, Florida and Ohio were less Democratic than the nation as a whole (Florida by 4%, Ohio by 2%). For Obama to be ahead by 9% in Florida would suggest he was ahead nationally by 13%, if Florida continued to be 4% less Democratic than the nation in 2012 as it was in 2008.

Similarly, a 10% margin for Obama in Ohio would suggest he was ahead by 12% nationally, assuming Ohio continued to be 2% less Democratic than the nation as a whole. The problem with these state polls is that no one is intimating that Obama is ahead by anywhere near that much nationally. On Saturday, the four national tracking polls have Obama ahead by 2%, 5%, 6% and almost 7%, an average of 5%.

On the day these two state polls were , The New York Times ran a front-page story about the polls arguing that Romney was in huge trouble in both Ohio and Florida and his chances to win in November were increasingly remote. That would be true if Romney actually was behind by 10 and 9 points in the two states.

As suggested above, this is extremely unlikely. A look into the details of the two state polls revealed that many more Democrats were interviewed than Republicans, and the results were not weighted to conform to any expected distribution of voters between the two parties. In essence, more Democrats answered the phone and responded to the pollster, and the results of those calls accounted for the survey numbers that were released. If the pollster had weighted his sample results, he might have shown a 3-5 point lead for Obama in each state, a result that was still favorable to the president, but not nearly as awful for Romney as the numbers that were released.

The New York Times story was picked up by all the major broadcast networks and local news media in both states, who hammered Romney and his running mate Paul Ryan, as well as campaign advisers, with questions as to whether their campaign was now on life support and all but dead. Remember that this one snapshot poll was conducted before any of the three presidential debates or the vice presidential debate, and six weeks before the election.

It is clear that Mitt Romney has had a tough few weeks. His campaign was knocked off stride when a tape was released from a private fundraising event with Jewish donors in Florida from May, in which Romney seemed to write off 47% of voters because they did not pay income taxes or were dependent on government.

While attacking the expansion of welfare is a popular campaign theme, labeling half the voters as dependents is not going to secure new voters and will likely lose some he already had. Predictably, Romney’s numbers dropped a few points in pretty much every survey after the tape was played and replayed hundreds of times on cable TV and radio and started appearing in Obama campaign ads and in his speeches. It is certain it will be brought up in the first debate between the candidates on Wednesday night, Oct. 3. The race seems to have stabilized a bit this weekend, and Romney’s numbers are a bit better in a few polls than a week ago, including in competitive states such as Nevada and Virginia.

Surely, Obama is the favorite to win at this point. But the debates present a unique opportunity for the challenger to show he is up to the job, and make his case before a large audience of voters without the filter of a media screen whose members are behaving just as in 2008, as cheerleaders in Obama’s corner.

If Romney can make his case about the failure of the president’s economic policies — the lost jobs, trillion-dollar deficits for four years running, the downgrading of America’s debt rating, the staggeringly high unemployment numbers three years into a supposed recovery, and the tepid economic growth numbers — he can remind voters of why they have been uneasy about giving Obama four more years.

For much of the year, the media and the Obama campaign have done everything possible to focus voter attention on everything but the economy: Romney’s years at Bain Capital, the tax rate he pays, his comments before the London Olympics, and most recently the 47% gaffe. The theme is simple: Do you really want the alternative to Obama — an out-of-touch rich white guy who pays a low tax rate, hides his money overseas and outsourced jobs to China?

The first debate is the most critical for Romney to change the narrative, introduce himself to voters, and bring Obama’s many failures back into the picture. If Romney is unable to accomplish this, it may be a steep uphill climb the rest of the way.