Tips for fact-checking
Whether you are a reporter, an activist, a business leader, a health worker or a regular citizen how can you know when public figures tell the truth and when they distort it? How can you decide what claims are fair and who to trust? To produce the reports for this site, we have used our own experience as journalists and the help and assistance of specialist experts in a range of fields, to draw up this checklist:
1. Ask where is the evidence?
Whenever anyone in public life makes a claim, big or small, the first question you should ask – once you have got past whether the claim is plausible and worth investigating – is ‘Where’s the evidence?’
Many now agree that, in the build up to the war in Iraq in 2003, (shown left), too little attention was given to the clear lack of hard evidence that Iraq possessed the weapons that was claimed.
There and elsewhere, there is often a good reason for an official to refuse to reveal the evidence behind a claim they make. They may need, as journalists do, to protect their source. But if sources need protection, we still need evidence. And often the reason officials refuse to provide it is that their evidence is weak or partial or contradictory.
So first ask for the evidence, and if none is forthcoming, you know there is, or may be, a problem with the claim.
2. Is the evidence verifiable?
The next step, if evidence is provided, is to see whether it can be verified. One of the key tests made before the results of any new trial are accepted by the scientific community is to see whether the trial can be repeated by other researchers with the same or similar findings. As Thomas Huxley, a prominent 19th century biologist, put it: “The man of science has learned to believe in justification not by faith but by verification”.
It should be the same in public debate. When a public figure, in any field, makes a claim they want believed, they should be asked to provide verifiable evidence. If they can’t, can you take what they say on trust?
3. Is the evidence sound?
As much as we have looked, there is yet no one, single checklist of tests that covers all the different types of evidence you might have to assess before you decide it is sound. Listed below are the main questions we ask.
Could they know what they claim to know?
If the evidence is based on an eye-witness account, could the person know what they claim to know. Where they there? Is it credible to believe they would have access to this sort of information? Is the information first- or second-hand, something they had heard and believed? Is it something that could be known?
If there is data, when was it gathered?
It is a favoured trick of public figures to present information collected many years previously, as if it were from today, and make no mention of the dates. But data ages and, unlike most wines, this is not good. To understand the data, you need to know when it was gathered and what the picture looked like before and after. It is a favoured trick of people presenting data to choose the start and end-dates not for how well they reflect the true performance, over time, but to make the numbers look good, starting at the bottom of a regular cycle and ending at the top.
Was the sample large enough? Was it comprehensive?
An opinion poll that samples the views of a few dozen – or even a few hundred – people is unlikely to be representative of the views of a population of millions. Most polling organisations suggest that a well-chosen sample of around 1,000 people is the minimum required to produce accurate results. But it is surprising how often surveys of a few hundred, or few dozen, people are quoted by public figures and reported in the media as representing wider views.
And remember, even large scale surveys – that do not look in all the right places – can give an inaccurate picture. This is something known by some as the ‘black swan problem’ – a reference to the erroneous assumption made in Europe for centuries that all swans are white because the hundreds of thousands of swans seen there over the years, all were. And indeed, it was only when 17th century Dutch explorer Willem de Vlamingh returned to Europe after discovering black swans in western Australia, that people began to know better. The previous ‘sample’ had been large, but had not been extensive enough.
How was the data collected?
Sample size is not all that matters. When taking an opinion poll, the pollsters have to include people of all the relevant social groups – both genders, all ages, from the different races, social-economic groups and different regions – and in the right proportions, if it is to be thought representative of society as a whole.
How was the study done? Similar surveys done door-to-door can produce different results from those done on the telephone because of how the interviewee responds to someone face-to-face and on the phone.
And studies that rely on the respondents filling in forms tend to show more errors, particularly among respondents with low literacy skills, than person-to-person interviews. If the claim is based on a survey like this, could that be a factor?
Meanwhile, what people taking part in a study or a trial know, or think they know, about it, will also affect the outcome. This is the so-called ‘placebo effect’ and explains why medical trials often are, or should be, ‘blinded’ – so that those being studied do not know the nature of the treatment they have or have not been given.
What about the wider picture?
Once you know how the data was collected, assess the way it was presented. Did the person tell the truth, the whole truth and nothing but the truth? Public figures from all walks of life like to select what they tell you, and what they don’t, cherry-picking the juiciest evidence, favourable to their side in an argument, and leaving the less tasty morsels in the bowl.
Is the data presented in context, and would mean the same, if other, unmentioned factors, are taken into consideration.
When a politician says, for example, that he or she put “record sums” into the public health system, and does not mention inflation, the claim may be true, in itself, but misleading if inflation means “real terms” spending is falling. So make sure to look at other factors that make up the wider picture.
And always remember to keep numbers in proportion. Spending $50million on a health project may sound like a lot, for a small community. But divide it among a population, and note that the programme is set to run over 10 years and it seems a lot less generous than it seemed at first.
4. Data sources, experts and the crowd
The fact that the person making a claim cannot, or does not, offer evidence to back up their statement, makes it harder to check but does not prove it wrong. To check it, you can turn to credible data sources, acknowledged experts and the crowd.
Numerous sources of data useful in fact-checking claims that people make exist. Depending on the sort of claim you are checking, you may seek information from government papers and official statistics, company records, scientific studies and health research databanks, through to school records, development charity accounts, religious orders’ papers and others besides.
We set out tips and examples of data sources we find useful on our Resources for fact-checking page. As with all sources of information, it is important to know all you can about the organisation that gathered and holds the data before you use it.
Depending on the topic – if the claim made is on medical matters, or require detailed knowledge of a major company’s accounts, or a fine point of law – it may be more suitable to check a claim by talking to a number of recognised experts.
When doing this, the most important thing is to know and declare any interest the expert may have in the matter that may cause, or be seen to cause, a bias in their analysis, one way or another.
Sometimes, the people you speak to may seek anonymity. This weakens your report but, if the information they provide is independently verifiable, may be acceptable. Univerifiable information from an anonymous source, who will only talk “off the record” should not be used.
Again depending on the topic, the best source for information to check a particular claim may not be a set of papers or a particular expert, but the knowledge to be found in the wider community; crowdsourcing as it is known.
If an official claims on election morning that all polling stations received their ballot papers on time, or an environmental group claims a factory is polluting a neighbourhood, the best placed people to confirm or undermine what they say may be people in the wider community.
When sourcing information from the crowd, you need to be cautious about a number of things. To start with, it is important you guard the security of your sources. In many cases, information sent by SMS, email and other means can be intercepted and in some countries people who supply ‘sensitive’ information to media sites may suffer for it, so it is important to set up ways to communicate that are as secure as possible.
At the same time, you need to know who your sources are and whether the information they supply is reliable. Seek to verify the identity of anyone who sends you information. Information sent in anonymously should be treated with necessary scepticism. Be wary of mass mailings by groups with agendas to push, and of using anecdotal evidence as if it were representative.
5. Spotting fakes
What you want to verify may not, of course, be a spoken or written claim but material – photos, videos, blogs or other content – sent to you or published online. In the digital age, photographs, video footage, text documents, websites and Twitter and other social media feeds can all be falsified. How to spot what is genuine and what is fake? These are our tests.
Do the words or images ring true?
First things first, before even you start to look for evidence, the most important thing to do when sent material is to engage your brain. Do the images or words ring true? Is the language or sentiment expressed the way the person would talk? Is it the sort of thing they might really have said?
Colleagues understand when people are taken in by clever hoaxes. But if it is obvious, after the event, that the person quoted would never have been likely to say the thing that was quoted, and you did not check, you can look foolish.
So first, think. And then, if in doubt, check with the person or organisation quoted or shown to verify.
- Is there a telling detail out of place?
Hoaxers are often let down by the details. Be sceptical always. The quote used in this cartoon is not wrong, but something should make you realise it was probably not the 19th century US president who said it.
Look at the phrase used and ask if that could have been said at the time. Look at the photo or video and ask whether it abides by the laws of light and shade. Are there things you can see in the background that should be there that aren’t or shouldn’t be there and are. Does the weather shown reflect the weather you would expect in that place, at that time of year? Are the views, plants, cars, buildings the sort you would expect to see?
If the details are out of place, it may be a hoax.
Has it – or something similar – appeared elsewhere before?
Unlike lightning, hoaxers often do strike twice If you are suspicious about an image or text, check online to see whether it – or something similar – has appeared elsewhere before.
Run a search on Twitter referring to the material with the hashtag ‘fake’ and see if others on Twitter have spotted something too.
If it is text you think might have been used before, drop it into Google search.
If it is a photo, or video free-frame in PNG format, drop it into a website such as www.tineye.com which allows you to check photos or videos to see whether they might have appeared online previously. If the same image, or one very similar, has been published previously in different circumstances, what you have been sent may be a fake.
- Has the person filed material elsewhere
Remenber people often use the same username on various platforms, so if you are searching for similar material from one person, put their username into different platforms such as Google Search, Facebook, Flickr, Twitter, YouTube, 123people.com, blogsearchgoogle.com, Technorati.com.
Check the person who sent it is where they say they are
If you have doubts about the source of some information, and have the numerical address – the IP code – of the computer it came from, you can check the country the computer is located in it you enter that into this address: www.domaintools.com/reverse-ip/
6. Be persistent
Fact-checking takes time and persistence. When someone tries to fob you off, refusing you access to data you are entitled to, or failing to provide evidence that backs up their claim, keep pushing.
Verifying public debate is not easy. The devil is, often, in the detail. To find it you need stamina and persistence.
7. Be open & accept you’ll have critics
Finally, be open in the way you write up any fact-checking reports, providing links to the evidence you use. And be honest, if you make a mistake, admit it. Even so, you need to accept that you won’t convince everyone.
Most people show some reluctance to accept evidence that goes against what they believe. And there are some that no amount of careful argument and linking to evidence will convince. It is a phenomenon known by scientists as the “persistence of discredited beliefs”, and describes a state where, according to the psychologists Craig Anderson and Lee Ross said: “Beliefs can survive potent logical or empirical challenges. They can survive and even be bolstered by evidence that most uncommitted observers would agree logically demands some weakening of such beliefs. They can even survive the total destruction of their original evidential bases.”
Some people, you just can’t convince.