Poll: Voters are struggling to commit

Around 1 in 5 voters nationwide report themselves as undecided or flirting with third-party candidates, with the exact share depending on the poll and how the question is asked. (AP Photo/M. Spencer Green, File)

Credit: M. Spencer Green

Credit: M. Spencer Green

Around 1 in 5 voters nationwide report themselves as undecided or flirting with third-party candidates, with the exact share depending on the poll and how the question is asked. (AP Photo/M. Spencer Green, File)

WASHINGTON — In this election season of discontent, a lot of voters are having trouble committing.

Around 1 in 5 voters nationwide report themselves as undecided or flirting with third-party candidates, with the exact share depending on the poll and how the question is asked.

That's far higher than in the last several elections, when fewer than 1 in 10 voters were still up in the air at this point, and reflects the distaste that large numbers of voters have for both Hillary Clinton and Donald Trump.

Those who remain uncertain include a couple of groups that may play an outsized role in determining the election's outcome — young voters, many of whom loathe Trump but lack enthusiasm for Clinton, and college-educated suburban Republicans, who often find Trump scary but struggle with the idea of voting for a Democrat.

"I'm just lost," Joanna Gianforcaro, 26, said on a recent afternoon as she sat with her mother at a farmers market in Doylestown, Pa., a swing area in a potentially important battleground state. Both women said they felt barraged by the negativity of the campaign and dismayed by the faults they perceive in both candidates. "I don't find either of them genuine," Gianforcaro said.

About a dozen miles to the southwest, in the suburban town of Lansdale, a longtime Republican expressed dismay about her party's nominee — his temper, his history of bankruptcies — and uncertainty about her decision.

"I think I'll know when I get there," said the woman, who would give only her first name, Evie, and her age, 67. "A lot of people are undecided," she added. "I don't think we'll know till we go to the polls."

The large number of voters in that position probably won't change until at least after the first presidential debate, scheduled for Sept. 26. Their prevalence has important consequences, both for the election itself and for the pre-election polls that voters will be hearing more and more about in the nine weeks between now and Election Day.

Trump causes concern and doubt among a lot of voters, but also inspires seemingly unshakable support from his core supporters. That ardor gives him an edge among voters who are most definite about their choice, according to the University of Southern California Dornsife/Los Angeles Times Daybreak poll, which is structured in a way that allows detailed analysis of the preferences of voters at different levels of certainty.

Clinton, by contrast, wins among voters who are fairly certain but express some doubt either about their choice or whether they will cast a ballot, according to analysis of data collected since the poll began tracking the election in early July.

That suggests a key task for Clinton over the remaining weeks of the campaign is not so much persuading voters to her side, but motivating her soft supporters. The need to do so, in turn, helps explain the Clinton campaign's seemingly insatiable appetite for cash, a lot of which is going to fund a vast — and expensive — get-out-the-vote operation in key states.

As for Trump, his apparent strategy of focusing on his core supporters, emphasizing the issues they care about most, as he did last week with his hard-line speech on immigration, may make considerable sense.

His advisers have suggested that they believe their best chance lies in a relatively low-turnout election in which lack of enthusiasm for Clinton holds down the Democratic vote. They also need to motivate a large number of blue-collar white voters who sat out the last election — a group that strongly supports Trump, but is less certain about voting, the poll data show.

In the most recent poll from the nonpartisan Pew Research Center, about 1 in 4 voters said they were undecided or didn't know when asked to choose between Clinton and Trump. In a four-way contest, 10 percent chose Libertarian candidate Gary Johnson and 4 percent Green Party candidate Jill Stein, while the rest remained unsure, Pew found.

Data from Pew and from SurveyMonkey, which polls thousands of voters each week on their presidential preferences, show younger voters much less certain of their vote than their elders. The SurveyMonkey numbers also show that Republicans, by a small but important margin, are more likely to be undecided than Democrats, reflecting the continued resistance to Trump within parts of the GOP.

A considerably larger group of voters remains at least somewhat uncertain about their choice. Only about 4 in 10 say they are sure of their candidate choice and certain to vote, according to the Daybreak poll, which asks people to estimate on a zero-to-100 scale what their chance is of voting for Clinton, Trump or some other candidate and, separately, the chance that they'll cast a ballot.

Among that absolutely certain group, Trump leads 51 percent to 45 percent, the poll found. But broaden the lens to take in the nearly two-thirds of poll respondents who say they are at least fairly certain of their vote and their likelihood of casting a ballot, and the picture flips, with Clinton holding a 48 percent to 42 percent lead. Overall, the Daybreak poll shows the two locked in a dead heat.

The poll consistently has shown Trump doing better than most other surveys, in part because people who are certain of their vote have more weight in the survey than those who are less sure. If Clinton's support solidifies during the course of the fall campaign, the Daybreak poll's results could move closer to those of other surveys.

The differences between voters who are sure of their choice and those who are less so can bedevil pollsters, adding to the uncertainty of election forecasts. Using a tight definition of which people to count as likely voters will probably yield a higher forecast for Trump, for example. That was true for a CNN/ORC poll released Monday, for example. The poll showed Clinton leading among all registered voters, but trailing by 2 points among likely voters.

The most recent polls show that the race has tightened from early August, when Clinton was enjoying a big bounce after her nominating convention.

At the post-convention peak, Clinton led by about 8 percentage points in polling averages. Now, with both bounces having faded, the race has returned to where it stood at the beginning of the summer: Clinton ahead by between 3 and 5 points in the various polling averages.

State-by-state polling indicates, however, that the Democrat continues to lead in enough states to deliver a victory, although some of those surveys date from early August, when her national lead was greater.

According to the Los Angeles Times' Electoral College map, only four states currently are clear toss-ups — Florida, North Carolina, Ohio and Iowa. To be elected, Trump would probably have to win all four of those and then pick up at least one more state where Clinton currently has a more solid advantage.

A 50-state survey published Monday by The Washington Post showed Trump doing notably worse than the 2012 GOP nominee, Mitt Romney, in some heavily Republican states — most notably Texas, where the survey found him tied with Clinton. Clinton almost certainly won't be winning Texas any more than Trump will carry New York, but Trump's poor showing there is a reminder of his weakness among Republicans nationwide.

That survey and others also showed that while the electoral map looks a lot like the one from four years ago, the demographics don't.

Trump does better among blue-collar white men than Romney did in many parts of the country. Clinton, by contrast, does better than President Barack Obama did among college-educated whites, especially women.

In the CNN/ORC survey, Clinton led among college-educated white voters by 13 points. Trump led among whites without a college degree by 44 points.

That massive educational gap among white voters doesn't necessarily shift many states from one column to the other, but it does mean big shifts within some states.

In Pennsylvania, for example, Trump does better than Romney did in some rural areas and old industrial regions in the western part of the state. Clinton, however, does better than Obama in the suburban counties outside Philadelphia.

In Florida, the biggest swing state, Trump similarly exceeds Romney's margins in many conservative white areas, according to recent polling, but Clinton is running ahead of Obama's mark in Miami, in large part because of Trump's extremely poor showing among Latino voters, as well as Clinton's stronger showing among whites with college degrees.

If trends hold, Clinton will win nationally among college-educated white voters. That would mark the first time in the history of polling that a Democrat has won the group, which used to tilt overwhelmingly Republican.