By LAURIE KELLMAN, The Associated Press
WASHINGTON (AP) — The Associated Press will debut a new survey of the nation's electorate for Tuesday's midterms that aims to more accurately capture the story of how Americans voted and why.
The launch of AP VoteCast is the largest change in the way final information about voters' choices is collected by media organizations since CBS News first conducted an exit poll in 1972. The AP's leaders say their decision to break away from the traditional exit polls is rooted in long-standing concerns about the accuracy of those surveys.
"These results will be scrutinized and analyzed by academicians in the months after the midterms," Gary Pruitt, CEO of The Associated Press, said of the seven-figure project. "And so, the tale will be told."
VoteCast is conducted by NORC at the University of Chicago for AP and Fox News. Both AP and Fox were members of the National Election Poll that has for decades conducted the exit poll.
The remaining members of the pool -- CNN, ABC, CBS and NBC -- will again rely on an exit poll conducted by Edison Research. Several other major news organizations, such as The Los Angeles Times will use VoteCast.
The most significant difference between the two is in how voters are surveyed.
Under AP's system, postcards are mailed to a random sample of registered voters in 25 states, inviting them to take a survey either online or by phone. At the same time, a random-sample survey of registered voters nationwide is conducted using a panel designed to be representative of the U.S. population. Finally, VoteCast surveys self-identified registered voters in all 50 states using opt-in online panels.
Those who take the survey are asked if they have, in fact, voted or are sure to.
AP believes its methodology will more accurately reflect the makeup of an electorate that increasingly does not vote on Election Day. In 2016, more than 40 percent of Americans voted early, absentee or by mail. It will also capture the opinions of registered voters who choose not to cast ballots.
Joe Lenski, co-founder of Edison Research, said the strength of its exit polling hinges on the fact that it is conducted with people as they are leaving polling places.
"Without doing that, there's no way to confirm that people actually voted," he said.
However, news organizations have long held concerns about the accuracy of exit polls, which have in recent years tended to overstate Democrats. In 2016, for example, they incorrectly suggested early on election night that Hillary Clinton would defeat Donald Trump.
AP Executive Editor Sally Buzbee said that as the actual vote started coming in and it became clear Trump was headed toward victory, she directed the newsroom's race call team to disregard results from the exit polls. AP accurately called the race for Trump before any other major news outlet.
"It was a very sobering moment that told me that we needed to try to see if we could use an alternative going forward," said Buzbee, who was AP's Washington bureau chief at the time.
After the election, the AP and NORC spent more than a year designing a new, nonpartisan method for sampling a bigger slice of the American public.
Working in close collaboration with Fox News, the product's first customer, AP tested this approach in the 2017 gubernatorial elections in New Jersey and Virginia, and in the Alabama special election for U.S. Senate. At poll close, the surveys accurately forecast the winners and their final vote shares.
On Nov. 6, VoteCast will be used in all 47 states that have statewide elections, including all races for Senate and governor and states with only one House district, such as Alaska. AP says about 122,000 Americans will be surveyed by phone and online.
Edison says its survey will sample 100,000 people.
The rubber meets the road when the results of the traditional exit polls and VoteCast are made public and compared. News outlets will make expensive post-Nov. 6 decisions about whom to trust ahead of the 2020 election.
David Lauter, Washington Bureau Chief for The Los Angeles Times, said his news organization opted for a four-figure contract for VoteCast's results only in California in 2018, and the Times' writers will base their analyses on those. But he said that since the exit poll consortium's results will be public, he'll also be comparing their accuracy with VoteCast.
"Our thinking was, this a different way of doing it. The old method clearly has flaws. Let's give this a try see how it works," Lauter said.
For AP's complete coverage of the U.S. midterm elections: http://apne.ws/APPolitics