You Are Here: Home » Archives for December 2017

Month: December 2017

A big year for fact-checking, but not for new U.S. fact-checkers

An "Almost Abe" rating from Indy Fact Check in Nevada. (The Nevada Independent)

A big year for fact-checking, but not for new U.S. fact-checkers

Published December 13, 2017 by Mark Stencel

All the talk about political lies and misinformation since last year’s election has been good for the fact-checking business in the United States — but it has not meant an increase in fact-checkers. In fact, the number has dropped, much as we’ve come to expect during odd-numbered years in the United States.

We’re still editing and adding to our global list of fact-checkers for the annual census we’ll publish in January. Check back with us then for the final tally. But the trend line in the United States already is following a pattern we’ve seen before in the year after a presidential election: At the start of 2017, there were 51 active U.S. fact checkers, 35 of which were locally oriented and 16 of which were nationally focused. Now there are 44, of which 28 are local and 16 are mainly national.

This count includes some political fact-checkers that are mainly seasonal players. These news organizations have consistently fact-checked politicians’ statements through political campaigns, but then do little if any work verifying during the electoral “offseason.” And not all the U.S. fact-checkers in our database focus exclusively — or even at all — on politics. Sites such as Gossip Cop, Snopes.com and Climate Feedback are in the mix, too.

The story is different elsewhere in the world, where we have seen continuing growth in the number of fact-checking ventures, especially in countries that held elections and weathered national political scandals. Again, our global census isn’t done yet, but so far we’ve counted 137 active fact-checking projects around the world — up from 114 at the start of the year. And we expect more to come — offsetting the number of international fact-checkers that closed down in other countries after the preceding year’s elections.

Still, the number of U.S. fact-checkers accounts for about a third of the projects that appear in the Reporters’ Lab’s database, even after this year’s drop.

So why do so many U.S. fact-checkers close up shop after elections? PolitiFact founder Bill Adair, who now runs the Reporters’ Lab and Duke’s DeWitt Wallace Center for Media & Democracy, asked that question in a New York Times op-ed on the eve of last year’s election. He attributed the retraction in part to the fact-checkers’ traditional focus on claims made in political ads, which was how the movement began in the early 1990s. Also, newsroom staffing and budgets often shrink after the votes are counted. That’s too bad, because, as Bill noted, “politicians don’t stop lying on Election Day.”

A handful of U.S. newcomers began fact-checking in 2017. One was Indy Fact Check. It’s a project of The Nevada Independent, a nonprofit news site based in Las Vegas. The Independent got its feet wet in January with a look at the accuracy of Gov. Brian Sandoval’s 2017 State of the State address before launching a regular fact-checking series in June.

To rate the claims it reports on, Indy Fact Check uses a sliding, true-to-false scale illustrated with cartoon versions of Abraham Lincoln. The facial expression on “Honest Abe” changes with each rating, which run from “Honest as Abe” and “Almost Abe” on the true side to “Hardly Abe” and “All Hat, no Abe” on the false side.

One of Indy Fact Check’s regular contributors is Riley Snyder, who previously was the reporter at PolitiFact Nevada at KTNV-TV (13 Action News). KTNV was one of several local news outlets owned by Scripps TV Station Group that briefly served as PolitiFact state affiliates before closing down the partnership — after the 2016 election, of course. So in Nevada at least, one site closes and another opens.

Another new player in the U.S. fact-checking market this year was The Weekly Standard. This conservative publication based in Washington has a dedicated fact-checker, Holmes Lybrand, who does not contribute to the political commentary and reporting for which the Standard is generally known. With this structural separation, it recently became a verified signatory of the International Fact-Checking Network’s code of principles. The Standard is owned by Clarity Media Group, a division of the Anschutz Entertainment Group that also publishes the Washington Examiner and Red Alert Politics.

By January, we may have a few more additions to add to our 2017 tally, but that won’t change the bottom line. This was a year of retraction in the U.S. That’s similar to the pattern our database shows after the last presidential election, in 2013, when PunditFact was the only new U.S. fact-checker.

But the numbers began to grow again a year later, during the midterm election in 2014, and continued from there. Because of the large number of candidates and the early start of the 2016 presidential debate and primary process, a number of new fact-checkers launched in 2015. So we’ll be watching for similar patterns in the United States over the next two years.

Student researcher Riley Griffin contributed to this report.

Comments closed

The wide world of fact-checking apps

Published December 6, 2017 by Bill McCarthy

It is no secret that news consumers are finding it increasingly difficult to separate fact from fiction, especially when it comes to politics.

Sure, they can visit journalism’s traditional truth-seeking outlets — such as PolitiFact or FactCheck.org — if they are looking for the whole story. But what if they want a quicker fix? What if they want to know, with the click of a button, if the article they are reading may include fabricated content? Well, there may now be an app for that — in fact, many apps.

The wave of falsehoods that dominated the 2016 election cycle has inspired several enterprising companies and individuals to create mobile applications and web browser extensions to promote fact-checking and detect stories with falsehoods.

In a recent analysis for the Reporters’ Lab, I identified at least 45 fact-checking and falsehood-detecting apps and browser extensions available for download on the Apple or Android app stores, the Google Chrome web store and Firefox. Many share similar design characteristics and functionality.

Several of the best apps and extensions simply make fact-checks more accessible. These apps, including Settle It! Politifact’s Argument Ender, let users view and filter through fact-checks aggregated from online fact-checking sites. (Disclosure: Bill Adair, director of the Reporters’ Lab, contributed to the creation of this app.) Some, like The Washington Post’s RealDonaldContext, are specifically tailored to fact-check President Donald Trump’s tweets.

A few extensions — such as FakerFact or NewsCracker — evaluate credibility online by generating algorithmic scores to predict whether particular web pages are likely true or false. I found both extensions questionable because it is not clear which inputs are driving their algorithms. But they show nonetheless that fully automated fact-checking may not be so far away — even if FakerFact and NewsCracker are themselves lacking in transparency and value.

Other extensions enable users to crowdsource fact-checks. Users of these community-oriented platforms can flag and provide fact-checks online for other users to see. Where these extensions fail, however, is in training their users to fact-check. My analysis noted that several users have submitted fact-checks for opinion statements — and several others have disputed statements on a hyper-partisan basis.

Many of the existing apps and extensions are designed to spot, detect or block false stories. Some alert readers to any potential “bias” associated with a website, while others flag websites that may contain falsehoods, conspiracy theories, clickbait, satire and more. Some even provide security checks for spear phishing and malware. One drawback to these apps and extensions, however, is that their assessments are subjective — because all such apps and extensions are discretionary, none can honestly claim to be the end-all arbiter of truth or political bias.

In summary, some of the identified apps and extensions — like FactPopUp, our own Reporters’ Lab app that provides automated fact-checks to users watching the live stream of a political event — show signs of being on the cutting edge of fact-checking. The future is certainly bright. But not all of the market’s apps and extensions are highly effective in their current form.

Fact-checking and falsehood detection apps and extensions should be considered supplements to — not replacements of — human brain power. Given that caveat, below are three of what I found to be the most refined options. They are ready for action as news-reading supplements.

GlennKessler

Glenn Kessler

GlennKessler, available for free download on Apple’s app store, is an aggregation of fact-checks from Glenn Kessler of The Washington Post’s Fact Checker. Kessler’s son, Hugo, created the app when he was 16 years old.

Users of GlennKessler can view fact-checked claims and filter them according to the number of “Pinocchios” they received or the political party of the speakers. The app also includes videos related to fact-checking and interviews with Glenn Kessler, as well as a game where users can test their fact-checking knowledge. As an added feature, users can learn about and email questions directly to Kessler himself.

Official Media Bias Fact Check Icon

The Official Media Bias Fact Check Icon, a free extension for Chrome browsers, purports to provide “bias” ratings for more than 2,000 media sources online. While browsing the internet, users are presented with a color-coded icon denoting each website’s “bias.”

A related extension, the Official Media Bias Fact Check Extension, highlights “bias” within Facebook’s news feed. Users can ask the extension to eliminate sources fitting a particular “bias” rating from appearing in their feed. Unfortunately, this “collapse” feature brings with it the possibility that users will abuse the extension to reinforce existing filter bubbles within an increasingly fragmented social media landscape.

It is important to remember as well that Media Bias Fact Check claims to find “bias” according to its own labeling methodology. This is a complicated assessment, so users should take the ratings with a grain of salt. As committed as a site may be to the truth, there can truly be no definitive rating for something so sensitive as political bias.

ZenMate SafeSearch and Fake News Detector

ZenMate

ZenMate SafeSearch and Fake News Detector, a free extension for Chrome browsers from the Berlin-based startup ZenMate, signals whether a website is “good” or “suspect.” Users see ratings not only of a website’s credibility, but also of its security and ownership. The extension does not work for articles appearing on social media.

Per the extension’s description, ZenMate SafeSearch “aggregates and enriches various databases and feeds” in order to assess the credibility of various webpages. I found this low level of transparency alarming. As with Media Bias Fact Check’s extensions, users should be wary that ZenMate’s ratings are by nature subjective. The concept of “bias” is likely more complicated for an algorithm to score.

Comments closed

Web Annotation presents a new way to break down the news

Published December 1, 2017 by Mark Stencel

When NPR used a new web annotation tool to live-annotate and fact-check the first presidential debate in 2016, the site brought in record traffic, with 7,413,000 pageviews from 6,011,000 users; 22 percent of visitors stayed all the way to the end, something worth noting in an era when quick, shareable information dominates the news.

“We think that people were coming for the transcript and then getting the annotation, which we were fine with,” joked David Eads during an interview with the Duke Reporters’ Lab. Eads was part of the NPR visuals team throughout the annotation project, and now works as a news applications developer at ProPublica Illinois. He held that skimming through long documents and transcripts, as opposed to reading excerpts or summaries of what people said, is the new way that people like to get their news and information online, hence why NPR’s monster of a debate transcript was so popular.

Whether or not that’s true, the success that some new web annotation tools have recently had when used by major newspaper companies is something to which we should pay attention. NPR’s live transcript was one example. Another is Genius, which started out as a website called Rap Genius that allows users to annotate the lyrics to rap songs, and has since expanded to include an annotation tool that’s being utilized consistently by The Washington Post. Every speech, statement or debate transcript published in an article on the Post gets annotated by a handful of journalists via the Genius sidebar, and this has clearly been working. According to Poynter, “engaged time on posts annotated using Genius are generally between three and four times better than a normal article.”

It makes sense that consumers would want to read original statements and primary sources, given current skepticism of the media and allegations being thrown around by politicians and social media bots. Web annotation provides journalists with a tool that allows them to be present while readers go through these documents, not to push an agenda or argument, but to provide expert context, analysis and background for their audience.

Web annotation may be limited in what it allows journalists to actually do; The Washington Post mostly only uses it to make it easier for journalists to comment on speeches and statements. But other organizations have gotten creative with their own annotation tools, like The New York Times, which annotated the U.S. Constitution, and FiveThirtyEight, which wrote annotated “Perfect Presidential Stump Speeches” for both Republicans and Democrats. And many of them, including Vox and The Atlantic, are utilizing web annotation in different styles and formats for the same purpose as The Washington Post: a tool their journalists can use to break down speeches and debate transcripts into something more digestible, whole, and transparent for readers.

The first few attempts at all-powerful web annotation programs similar to Genius were epic failures. Third Voice was a browser plug-in created by a team of Singaporean engineers in 1998 that allowed users to annotate anything on any website. hough it showed promise in fostering intellectual conversation on tense topics online, it couldn’t shake its reputation of being a destructive method of Internet graffiti. A few years later in 2009, Google came up with Google Sidewiki that essentially did the same thing and encountered similar problems, as well as complications with advertisements and user communities. It lasted two years.

It’s no surprise that when you give the greater online community proverbial markers with which to annotate the entire internet, things go badly. It essentially becomes another YouTube comment section in which people can say and do whatever they want — except they can do it directly on top of a paragraph in a news article, which horrified many website owners.

But when the power to annotate is given to an expert (such as David Victor, co-chair of the Brookings Initiative on Energy and Climate and global policy professor at the University of California San Diego), in a specific setting (such as Trump’s Paris Climate Agreement speech, posted on Vox), the resulting article is an archive of knowledge and information that allows readers to know not only exactly what was said, but also what it really meant and why it matters.

An old example of annotation that I did find interesting to look back on was the first-ever instance of something being annotated: Martin Gardner’s Annotated Alice. Gardner was apparently frustrated by how much of Lewis Carroll’s clever genius was being missed in Alice’s Adventures in Wonderland, as many of the witty jokes Carroll made were strict references to Oxford and to the Liddell family (whose daughter Alice inspired the novel). In the introduction of The Annotated Alice, Gardner remarks that “in the case of Alice we are dealing with a very curious, complicated kind of nonsense, written for British readers of another century, and we need to know a great many things that are not a part of the text if we wish to capture its full wit and flavor.”

If you ask me, “a curious, complicated kind of nonsense” also sounds like an appropriate way to describe most of today’s political discourse. American citizens don’t know how to tell what is true and what is false, what is being exaggerated or made up, and what agendas these claims are trying to serve. And with social media setting such a complicated, manipulated stage for information to spread, it’s becoming increasingly harder to objectively assess the news.

Web annotation isn’t the answer to all of these problems, but it’s a nice way to start breaking down the primary sources, original statements and speech and debate transcripts. When used carefully, as NPR and The Washington Post have proved is possible, it allows journalists to present information in a transparent, natural way. The role of a journalist is, after all, to present complicated, sometimes nonsensical happenings in a way that an audience can understand, and if there’s a way to make that education easier and more entertaining, it should be welcomed.

Comments closed