From Traditional To Online Fact-Checking

July 8, 2013 • Ethics and Quality • by

Fact-checking is an old journalistic practice now being brought into considerably greater prominence due to the possibilities offered by the Internet.

Fact checking is an important job, that began to be done systematically at some US magazines – like Time and the New Yorker in the 1920s. It is also done ex-post in some newspapers, through a column for “mistakes and corrections”, as in the Guardian (corrections & clarifications), and in the Wall Street Journal (corrections & amplifications). But behind the issue of fact-checking (especially political fact-checking), as developed on the Internet through blogs and independent platforms, there’s much more going on. When you deal with concepts like that of “truth”, things tend to get blurred and the way you define the role of the fact-checker is actually intimately connected to your core attitude towards journalism.

At first, it may seem easy to grasp. “Every man has a right to be wrong in his opinions, but no man has a right to be wrong in his facts” is a familiar quotation, attributed to the publisher and philanthropist Bernard Mannes Bruch. Fair enough. Some may see Mario Balotelli as a great player or just as a spoilt young man, but if I say that he scored 13 goals in 2011-2012 in the League, with Manchester City, that’s not an opinion, that’s a fact (as long as my source is reliable). This is easy because numbers are involved. But take politics, the field where current online fact-checking focuses the most, and you’ll often find that matters are not so straightforward. “Why spoil a good story with the truth?” is a more accurate description of the philosophy behind a politician’s words than the one attributed to Mannes Bruch.

As the language of politics tends to become closer and closer to the language of advertising, what is a journalist’s main obligation? To report both sides of an issue, according to the notorious “he said, she said” fashion even when one of the two positions is apparently inaccurate, or should the journalist correct the falsehoods, whether they have been said in good faith or not.

Many news outlets stick to the former idea, calling it “impartiality”. Some prefer the latter interpretation, and take advantage of the new capabilities provided by the Internet to fact-check lies and mistakes. What exactly can the Web do to enhance fact-checking? The answer lies in five key points: real time checking, gamification, open data, crowd sourcing and semantic analysis of content. The wealth of data now available on the Internet, often coming from official institutions (the World Bank, the IMF, etc.) and provided in an “open” format, where facts can be searched and cross referenced,  allows journalists to discover interesting patterns without spending too much time and money in costly old-style investigations.

By gamification we mean the use of entertainment and “funny” formats in order to help the audience digest even complex news – tests, quizzes, newsgames, smartphone apps that use ad-hoc meters to rate the politicians’ claims are some examples of the use of gamification. Twitter is the most widely used medium for real time fact-checking: in recent years we have seen it in action during a number of presidential debates in various countries. It usually works this way: while the candidates are speaking on TV, a team of editors check and verify their statements and tweet the result. Usually a first version of the tweet is limited to saying whether a certain claim is true or false, and a deeper analysis will follow later on the team’s website. Crowd-sourcing uses “the wisdom of the crowd” to help the journalist investigate the veracity of a certain declaration. Real time semantic analysis of a text contained in a video interview could be considered the last frontier of fact checking.

Since 2003 several different Internet projects, most but not all of them based in the USA, have tried to combine some or all of these elements to evaluate controversial claims. The pioneer of them all was, a no-profit, non-partisan platform founded by the Annenberg Public Policy Center of the University of Pennsylvania, supported financially by the heritage of the late Walter Annenberg (and lately, also through donations).

The team of editors and reporters is free to investigate without the constraints of deadlines or of newspaper’s political leaning. A recent investigation concerned the correlation between gun ownership and murders in the USA (Gun Rhetoric vs Gun Facts, 2012). It took into account data provided by the United Nations Office on Drugs and Crime and by other sources, including the different outlook of several scholars on the matter. spelled out the difficulties in establishing any definitive causal relationship between gun ownerships and firearm homicides. Another important task performed by this verification platform, through the sister website, is to debunk the rhetoric in video ads of political candidates.

In 2007 two new protagonists appeared: the Fact Checker column of the Washington Post and a for-profit project of the Tampa Bay Times, the website PolitiFact. Both use a mix of rigorous in-depth analysis and gamification. The Post uses a scale made of ‘Pinocchios’ (the more Pinocchios you earn, the more you are likely to have indulged in a blatant lie) to rate the truthfulness of statements by politicians of both sides of the American spectrum. Sometimes, albeit not often, they turn out to be completely true. So, when Republican Dave Camp complained that “It takes the average American taxpayer 13 hours to comply with the tax code, gathering receipts, reading the rules and filling out the forms the IRS requires… The tax code forces Americans to spend over $168 billion to comply and 6 billion hours”, he chose his words carefully, quoting data from the Internal Revenue Service and earning a prized Geppetto checkmark.

On the other side of the spectrum, Representative Michelle Bachmann’s claim that 70 percent of food stamps go to ‘bureaucrats’ – instead of going to those in need – made in a speech at the Conservative Political Action Conference on March 16, 2013 – earned her 4 Pinocchios, with fact-checker Glenn Kessler adding that “there really aren’t enough Pinochios for such misleading use of statistics in a major speech.” Here Bachmann seems to have misinterpreted a figure contained in the book, “The Poverty of Welfare: Helping Others in Civil Society”, and then blithely applied the very same figure to the Food Stamp Program (a program in which recipients receive a plastic Electronic Benefits Transfer card to buy food at stores) not realising that official data provided by the Agriculture Department show that less than 6 percent of the program is spent on administrative costs.

PolitiFact, which won a Pulitzer for its coverage of the 2008 presidential debate, has a similar approach: a meter that goes from true to mostly true, half true, mostly false, false and “pants on fire”. They call it Truth-O-Meter. The Truth-O-Meter has a section dedicated to the politician that “flipped” from one side to the other, i.e. partially or completely changed their view on a certain subject. So we learn that Mitt Romney has always been consistent on his views on gay marriage, while Obama, on the very same subject swung like a pendulum – for it, against it, for it again.

The staff of PolitiFact also devised the Obameter, to track the promises made by Obama and verify if they have been fulfilled, or not, or if they are in the works. Another noteworthy feature of the website is the “Lie of the year”. Started in 2009, the “award” usually identifies a statement made either by a Democratic or a Republican representative on a topic that’s particularly controversial. In 2012, according to PolitiFact, the lie of the year was Mitt Romney’s claim that Jeep would start producing Jeeps in China at the cost of American jobs.

In August 2010 the Guardian introduced the online Pledge Tracker, which aims to keep tabs on the more than 400 promises made by the Coalition Government in its first weeks in power. More recently, the Guardian set up its Reality Check series, now very occasional, based more on collecting the voices of different experts on a certain subject and less focused on speaking with its own voice. Other experiments in online fact-checking include the Pagella Politica  website in Italy, inspired by PolitiFact, and Le Veritomètre in France.

In the UK, an interesting case is that of Full, whose motto is “promoting accuracy in public debate” and which aims to differentiate itself from the sometimes OTT style of its American counterparts. This is clearly stated in their own editorial guidelines: “We believe in playing the ball not the man and leaving it for our readers to make judgements about people and their motives. There is enough cynicism about politics and journalism without us contributing”. is a not-for-profit company, supported by three independent charitable trusts: the Joseph Rowntree Charitable Trust, the Nuffield Foundation and the Esmee Fairbairn Foundation, and from individual supporters’ donations. FullFact aims at exercising a kind of soft-power towards the institutions, achieving change without directly attacking single individuals (playing the ball, not the man, that is) this has led to some successes, one of the most notable being getting the then-Prime Minister Gordon Brown to drop a claim for which No. 10 was unable to provide backing, i.e. that the government had helped 300,000 small businesses
through the recession.

Will Moy, editor-in-Chief of FullFact described the situation to me as follows. “Our approach is connected to the fact that we see our fundamental mission as promoting well-founded trust in public debate, whereas other factchecking organisations often seem to come more from the ‘why is this lying bastard lying to me’ take on journalism. That is, they seem to see their role as a consumer/voter advocate guarding citizens from misleading claims. We see our role a bit more broadly, as helping citizens, policy-makers and others with important decisions to make get trustworthy information and avoid spreading unreliable information”.

Almost all of the online fact-checking tools are either no-profit, independent enterprises, or are embodied in the normal coverage of an online newspaper. The one exception is PolitiFact which is in fact much criticized (especially by those who earn the “pants on fire” medal) by some for being too aggressive in order to get the attention needed to have the advertising to repay its costs, by others, in contrast, for allegedly being “too balanced” in order to avoid accusations of bias.

But a question, naturally arises: do these fact-checking platforms have any impact on the actual deeds or words of politicians; or on the way newspaper shape their coverage? The answer to the first question, in general, might seem to be “no”. Politician still quote misleading figures or release completely false statements. But things are slightly more complex. Michael Dobbs, the inventor of the Washington Post fact-checker column, and its first editor, says in his essay “The rise of political fact-checking” that politicians’ assistants started to call him before running a speech, to make sure they would not fall into contradiction. And, on the other hand, many factcheckers argue that stopping politician from lying is not a good metric of their platforms’ success. As PolitiFact’s Bill Adair wrote: “To argue that fact-checking is a failure because politicians keep lying is like saying that investigative reporting is worthless because politicians are still corrupt. Yes, they are still corrupt. But we do investigative reporting and fact-checking to give people the information they need to make wise choices”.

A poll of 1,522 adults conducted in September by Social Science Research Solutions for the University of Pennsylvania’s Annenberg Public Policy Center, showed that persons reporting use of fact-checking information on the Internet answered 55.5 percent of a set of the “political knowledge” questions correctly, compared to 45.5 among non-users. The way the media cover politics (especially American politics) has also changed in recent years due to the work of fact-checkers. Many headlines in the 2012 American elections referred to the falsehoods in debates as a key part of the news, often quoting the problem in headlines such as “Mr. Ryan’s Misleading Speech” (Washington Post) or Deficit Vow Lacks Specifics (AP). And politicians started to use factchecking as a weapon against their adversaries or a tool to promote their own proposals. On the other hand Neil Newhouse, the Romney campaign’s pollster, claimed during the Party national’s convention in Florida “we’re not going to let our campaign be dictated by fact-checkers”.

In the end the question about the impact of intense fact-checking is closely related to another, much debated question: does journalism have a real impact on society? Fact-checking tools are more likely to be used by newspapers than by broadcast media: texts are easier to crunch and analyse than words spoken in video, though this may change, and sooner than might be expected. The Washington Post has developed a prototype, called Truth Teller, of software able to automatically extract relevant quotes and figures in a speech and compare them to a series of official databases, saying on-the-fly whether the speaker is telling the truth or not. It is still very primitive, but is a first step towards automatic truth telling. Does all this mean that we’ll no longer need reporters to investigate and confront official versions? Not so. For all its power online fact-checking still has and will probably always have, some limits: yes, it works well when numbers are involved, but this is rarely enough. Figures must be contextualized and put into perspective; some
politicians may choose to mention only numbers that are favourable to them, or to make us of ambiguous statements, that can be read both ways. Being only human, fact-checkers themselves can be biased. Human beings have a strong inclinations to believe in what suits them, even if it is clearly unrealistic. After all, who needs hard earned truths when lies are so beautiful, and they come for free?

Photo Credit: International Journalism Festival / CC Flickr

The Reuters Institute for the Study of Journalism is a partner in the European Journalism Observatory network.

 This article was first published in the Oxford Magazine.

Print Friendly, PDF & Email

Send this to a friend