Respond to prefers-color-scheme dark in app web views

Follow section front redirects in Spectrum

Enable NLP source in PCS for load testing

Switch off the News Alerts Survey Widget when we have enough responses

Poll to determine if new LUF updates are available

Allows the winning A/B test variant to be turned on without Spectrum deployment

Show infinite scrolling sections in Assembler MultiView

Show option for NLP article summaries in Action Bar

A general feature flag for Recipe Features that aren't yet released

Feature flag for media team features that aren't yet released

Use new Orca video player for preroll plugin in Spectrum

Show the snow tracker (instead of the heat tracker) on Spectrum

Biden honors distinguished Americans; Trump back in N.Y. court

Hope Hicks witnessed nearly every Trump scandal. Now she must testify.

State laws measure if a woman is close enough to death for an abortion

Trump Media auditor charged by SEC with ‘massive fraud’

Google’s empire is massive. A judge will soon rule if it’s a monopoly.

An I bond is now at 4.28 percent. Are they still a good deal?

Use these 6 points to say no to a college you and your kid can’t afford

How this new rule protects retirement savers from costly advice

The New York Times, alone in its outrage over access to Biden

Wait, what is Stephen A. Smith talking about? Find out with our quiz.

Stephen A. Smith would like even more of your attention

Israel’s offensive is destroying Gaza’s ability to grow its own food

Riot police and over 2,000 arrests: A look at 2 weeks of campus protests

‘It’s been abused’ Rep. McCaul on motion to vacate threat

What’s the best bus to New York? We tested 5 for comfort and value.

You asked: Do I need a passport for my cruise?

Adaptive equipment is making national parks more inclusive

Ask Amy: Should I be the best man if I think the bride is abusive?

Carolyn Hax: Husband keeps pushing modest spouse to wear sexier clothes

Miss Manners: Host eats luxurious breakfast, leaves cereal for guest

Teen grabs wheel when driver passes out and school bus veers into traffic

Cat climbed into Amazon return box, found alive 630 miles away

Brave Little Hunter is free: Baby orca escapes lagoon for open seas

Journalism in India is under assault

The dangerous new call for regime change in Beijing

Ukrainian men abroad voice anger over pressure to return home to fight 

On Press Freedom Day, little to celebrate for Hong Kong’s journalists

Young Georgians want to be part of Europe. Their government is in the way.

Taxes are inevitable. Paying to prepare them doesn’t have to be.

The Capitals, a team in transition, know what they want this offseason

Is Pastis as good as our food critic thought? He sent spies to find out.

Nationals can’t solve Nate Eovaldi, drop series finale to Rangers

What are the chances you will vote in the 2016 general election? Did you vote in the 2012 presidential election?

These are the questions pollsters often use to identify “likely voters,” perhaps the most ubiquitous and least-understood phrase in election news — and for two big reasons:

Pollsters employ widely differing methods for identifying likely voters (and many keep their methods under wraps)

Research on the accuracy of likely voter identification is relatively rare, since checking whether respondents actually vote can be expensive.

Enter a <a href="http://www.pewresearch.org/2016/01/07/can-likely-voter-models-be-improved/">major new study from the Pew Research Center</a>, testing which methods work best for picking likely voters and how this impacted on election survey accuracy in 2014.

Strikingly, the study found likely voter "screens" developed a half-century ago are still effective at filtering out non-voters and improving polls' representation of actual voters. But Pew also found knowing voters’ actual history from official records improves accuracy even more and complex machine-learning algorithms to assess people's probability of voting also improved accuracy.

Before we get into the nitty-gritty, here is a brief primer on likely voter models:

Most election polls start with a sample of the overall adult population or registered voters. But not everyone actually votes in a given election, and voters and non-voters can differ quite a bit; typically, actual voters are more Republican. Nobody "knows" who will vote, but pollsters try to identify who is likely to vote by asking questions that have been correlated with voting in the past — like those above — and filtering out respondents who are less engaged in the election.

The Pew study conducted a "validation" of likely voter methods. Basically, they asked respondents to a September 2014 survey who they supported in the 2014 congressional elections while using their traditional likely voter questions. But then they checked whether those respondents actually voted by matching their names, addresses and demographic information to official state-level voting records aggregated by the company TargetSmart.

This allowed Pew to compare the effectiveness of asking about likelihood to vote to identify future voters with relying on official records of past behavior — a method routinely used by campaign polls but less so by media surveys.

Pew used the results among verified voters to judge which likely voter models performed best. Before the election, Pew's overall registered voter sample showed Democrats with a four percentage point lead in the generic congressional ballot, 42-38. But among verified 2014 voters, Republicans held a three point edge, 44-41. The most effective likely voter model, then, is one that resulted in a vote margin with Republicans leading by three percentage points.*

Which likely voter models worked best?

The chart below outlines how all the likely voter approaches performed. We'll walk through the specifics, but one of the clearest takeaways is the benefit of official historical voting records. As Pew researcher Scott Keeter explained in an interview, "If you have verified past votes you can significantly improve the accuracy of predictions — at least in a low-turnout election."

Take Pew's traditional Perry-Gallup scale method for instance, which includes respondents with only the top 60 percent of scores on a seven point index. It showed an even 47-47 race between Democrats and Republicans. The traditional model was effective in producing a more accurate estimate than registered voters overall (42 Dem-38 Rep).

But when past voting indicators from the voter file were added to the model, the margin shifted to a one point Republican advantage — R+1 — slightly closer to the R+3 benchmark. The impact was larger under a more complex logistic regression approach, where the addition of vote history as a predictor flipped a two-point Democratic advantage (D+2) to a two-point GOP edge (R+2), coming within one point of the verified voter margin.

The study also showed the promise of some more complex statistical methods to predict likelihood to vote, such as "random forest," a machine-learning method that determines combinations of attributes that best predict whether someone will vote.

Likely voter models using these methods performed at least as well as the traditional Perry-Gallup scale and produced a slightly better estimate. The voter list firm TargetSmart's proprietary measure of voter likelihood developed by Clarity Campaign Labs also resulted in a strong likely voter model, resulting in a Republican lead of four points (R+4) compared with the benchmark of three points. Logistic regression models using likely voter scales performed worst but were greatly improved when combined with records of historical voting.

The report offered some clues as to why likely voter models using official records for voting tended to provide a more accurate picture of the future electorate. The study confirmed a longstanding challenge of asking about vote likelihood: Many people who say they will vote don't actually show up, and some who appear unlikely to vote eventually do cast ballots. The table below shows how respondents ranked on a 7 point scale of likely voting used by Gallup and Pew; 83 percent of those who scored the highest on the scale (most likely to vote) actually voted, but there was no record of voting for 17 percent of them. On the opposite end, more than 1 in 5 of verified 2014 voters scored a 4 or lower on the likely voter scale.

While all Gallup's likely voter survey questions were correlated with voting, they were not as powerful predictors as actual records of past voting behavior.

What this means for poll watchers in the 2016 election

The fast-approaching 2016 primary contests offer a big test for pollsters' likely voter models — turnout will be even lower than the 2014 midterm general election, and sorting voters from non-voters will be difficult in an already error-prone polling situation. Add to that mix Donald Trump, who is motivating non-traditional voters who it's not clear will actually turn out to vote, and you get the picture.

Most relevant to caucus and primary polls, the Pew findings indicate that polls based on voter registration lists — which often include voters previous history of participation — have a leg-up in identifying likely voters, compared with polls that only ask whether respondents will vote or their previous voting history (such as based on Random Digit Dialing of samples of adults). Better likely voter identification may help mitigate other challenges, such as the inability to contact large portions of the electorate whose names cannot be matched to a telephone number.

For both the primary and general election, the results illustrate the large impact of pollsters' decisions on identifying likely voters. Even this deep dive is based on just a single election, so it's unclear how well the same models would perform in a higher-turnout scenario; Pew's Keeter says he hopes to conduct a similar assessment this year to find out.

* Pew's September margin of Republicans plus 3 among validated voters is smaller than their final advantage in the election; a post-election survey of the same respondents found Republicans' advantage grew to nearly the identical actual vote margin.

Vanderbilt University, BA in political science

University of Maryland, MS in survey methodology

Scott Clement is the polling director for The Washington Post, conducting national and local polls about politics, elections and social issues. He began his career with the ABC News Polling Unit and came to The Post in 2011 after conducting surveys with the Pew Research Center's Religion and Public Life Project.

Scott Clement

The Washington Post

Biden administration

The Fix

The 202s

Polling

Democracy in America

Election 2024

Politics

prism://prism.query/site,/politics/the-fix?limit=20

The Washington Post's political analysis blog.

prism://prism.query/site-articles-only,/politics&limit=20

Post Politics from The Washington Post is the source for political news headlines, in-depth politics coverage and political opinion, plus breaking news on the biden administration and White House, Congress, the Supreme Court, elections and more.

prism://prism.query/site,/politics/the-fix&limit=6