Saturday brought little change to our forecast of the presidential race; President Obama is given a 69.7 percent chance of winning the Electoral College, the same figure as in the update on Friday.
Mr. Obama did get an encouraging number in the Rasmussen Reports tracking poll, which has him leading Mitt Romney by 2 percentage points – a fairly sharp reversal from the same poll on Wednesday, which had Mr. Romney four points ahead.
Could Mr. Romney’s small bounce from his selection of Representative Paul D. Ryan as his running mate already be fading? Possibly – but the model generally will not be swayed much by the movement within a single tracking poll. And Mr. Romney maintained a two-point lead in the Gallup national tracker. We will need to wait for a few more polls before we have a better sense of the momentum that the parties have headed into their conventions.
There was also an interesting poll out on Saturday from the Kaiser Family Foundation and The Washington Post. The poll is not that relevant to our forecast since it is a little out of date – its interviews began in late July and were completed on Aug. 5. But the poll took an especially large sample – more than 3,000 respondents – which allowed them to explore the vote choice among a richer set of partisan subgroups than pollsters typically consider.
The poll found, for instance, that among a group it called D.I.Y. Democrats – adults who identify as Democrats but who tend to be rural and often socially conservative – 21 percent planned to vote for Mr. Romney. However, among a group the poll calls window shoppers – young Republicans with moderate views on social and fiscal issues – 37 percent planned to cast ballots for Mr. Obama.
Another relatively unique element of the poll is that it released horse-race numbers among all adults that it surveyed, something that is rarely done in presidential polls. (Almost all polls in our database listed numbers among registered voters or likely voters instead.) Mr. Obama had a large lead in the poll among this broader universe of adults, beating Mr. Romney 52 percent to 40 percent. But his lead was narrower – seven points – among registered voters.
The Kaiser Family Foundation poll did not publish numbers among likely voters. In general, likely-voter polls tend to find worse numbers for Democrats – typically about 2 points worse than among polls of registered voters.
Our model is designed to adjust for this; the figures you see in our forecasts reflect the projected results among likely voters.
But what about everyone else – what we might think of as unlikely voters? Another poll this week, conducted by Suffolk University and USA Today, tried to get at that question. It surveyed 800 adults who either were not registered to vote, or who were registered but said they were not likely to turn out.
Among these adults, 43 percent said they preferred Mr. Obama, while 17 percent backed Mr. Romney. Since quite a few Americans fit into this category – about 4 in 10 adults will not vote in November – it is easy to see how Mr. Obama could have a double-digit lead when they are added back into the total, like in the Kaiser Family Foundation poll. But those adults will not help Mr. Obama any if they do not show up on Nov. 6.
In addition to Democrats, third-party candidates are hurt by the failure of these “unlikely voters” to show up at the ballot booth. In the Suffolk poll, about 20 percent of unlikely voters said they would prefer to vote for a third-party candidate for president – much larger than in polls of likely voters.
Fledgling political parties like the Libertarians and the Greens thus face something of a Catch-22. Many adults who might otherwise be inclined to support them do not bother to vote, possibly because they do not regard the parties as viable and think voting would be a waste of time. But without doing a little better at the ballot booth, it is hard for these minor parties to demonstrate their viability and gain any momentum.
Our forecast estimates that third-party candidates will collectively account for only about 1 percent of the vote this year, in part because there are some states where only Mr. Obama and Mr. Romney will be on the ballot.
It should come as no surprise that Democrats and third parties generally support efforts to make it easier to vote, while Republicans often seek to erect barriers to voting, like stricter voter identification laws.
My view is that concerns that some Democrats have over voter ID laws is a little overblown – academic studies suggest that photo identification laws do not affect turnout as much as you might think. Still, they could make a difference at the margin. My estimate is that the new voter ID law in Pennsylvania could cost Mr. Obama about 1 percent of the vote there.
Perhaps a broader concern for Democrats is not the small number of adults who will seek to vote on Nov. 6 and will be turned away, but the much larger number who will make no effort to do so.
In 2008, Mr. Obama tapped more deeply into the universe of “unlikely voters” than almost all candidates that preceded him, garnering a significant number of votes from groups like young voters and minorities who have a low propensity to turn out.
But he may not repeat that success this year. In almost every swing state that tracks voter registration by party, the share of voters who are registered as Democrats is down from four years ago.