Open a new tab. Do it before you read the article someone forwarded you. The Stanford History Education Group tested 45 people, 10 professional fact-checkers, 10 historians, and 25 Stanford undergraduates, and the professionals were the fastest and most accurate. Not because they read more. Because they read less of the page in front of them. Within thirty seconds of landing on an unfamiliar site, the fact-checkers would leave the article behind and search for information about the source itself. This is lateral reading: instead of judging a page by reading it, you read what others say about it. The historians and students stayed put, scrolling vertically, and were consistently fooled. The study, 'Lateral Reading: Reading Less and Learning More When Evaluating Digital Information' (Wineburg & McGrew, 2017), produced a finding that was not subtle. The people paid to verify claims for a living did the opposite of what we assume careful readers do.

The Wineburg and McGrew study gave us the term lateral reading, and it is the core of how you verify anything a stranger forwards you. The habit is simple: open new tabs. When you receive a forwarded note on WhatsApp or Telegram, do not read it first. Do not react to it. Do not forward it. Open a new tab and search for the sender, the website, or the claim. The professional fact-checkers did this immediately because they assumed the page in front of them was not a reliable source of information about itself. A domain age check tells you when a site was created. Lateral reading tells you who created it and why. You are not looking for what the page says. You are looking for what a search engine says about the page. The goal is to see if the source has a reputation for accuracy among people who are not the source itself. If the search results are empty, or dominated by debunking from verification organisations, you have your answer faster than reading a single paragraph of the original.

The Four-Step Verification Sequence

The professional technique goes in a specific order, and the order matters more than any single step. First, stop. Do not share, do not react, do not comment 'is this true?' on a group chat. The failure mode here is verification bypass: amplifying a claim by responding to it while it is still unverified. Second, investigate the source. Open a new tab and search the source's name, the website's domain, or the organisation. The fastest way is to look up the source on Wikipedia. This is not a perfect encyclopedia; it is a starting point. Wikipedia's article on a news organisation, a company, or a person will tell you if there is a history of fabrication, a political bias, or a past scandal. Third, find better coverage. Search for the claim itself, not the page that made it. Add the name of a verification organisation to your search. In Singapore, that means checking the government portal, the independent Black Dot Research, the Straits Times' CheckMate column, or CNA's Sure Anot. Fourth, trace claims to original context. If the note contains an image, do a reverse image search using Google Images, TinEye, or Yandex Images. If it contains a video, search for a key phrase from the transcript or a distinctive frame.

Mike Caulfield's SIFT method is the teachable shorthand for exactly this process. Caulfield published it in 2019, and it has become the standard curriculum for the Civic Online Reasoning program at the Stanford History Education Group. SIFT stands for Stop, Investigate the source, Find better coverage, Trace claims to original context. The 'Stop' step is a pause before sharing or reacting. The 'Investigate' step is the lateral move: use Wikipedia or a search engine to learn about the source before reading it. 'Find better coverage' means search for the claim on a verification site or a reputable news outlet, rather than trusting the page in front of you. 'Trace claims' means locating the original source of an image, video, or quote to see if it has been cropped, doctored, or taken out of context. SIFT turns lateral reading into a checklist. It is the difference between a reflex and a routine. Without SIFT, you are relying on how a note feels. With SIFT, you are relying on what the note is.

Vertical Reading Fails

Lateral reading is not vertical reading. The difference is the direction of your attention. In vertical reading, you stay on the original page, scrolling, clicking links, and reading the 'About' page. Anyone can write an About page. A scam site can have a beautiful 'About Us' page with stock photos and a convincing mission statement. Lateral reading ignores the About page entirely. It leaves the page and consults external sources. Wineburg and McGrew documented the distinction: the factor that separated the fact-checkers from the historians and students was not intelligence or prior knowledge. It was the willingness to abandon the page. The historians and students read vertically, evaluating the internal consistency of the argument, the design of the site, and the tone of the writing. The professionals treated all of that as irrelevant. They knew the best way to judge a source is to see what other sources say about it. A single search result from a reliable news outlet that contradicts the claim is worth more than any amount of polished prose on the original page.

Lateral Reading On A Phone

On a phone, the process is identical but the habit has to be more aggressive. The screen is small, so the temptation to scroll vertically is strong. You do lateral reading when you tap and hold a link to open it in a new tab, then switch between tabs using the tab switcher icon. The physical act of switching tabs is the lateral read. Do not read the new tab; just switch to it, check the domain, and switch back. Do it fast. The Wineburg and McGrew study found that professional fact-checkers left the original site within 30 seconds. On a phone, investigate the source in under a minute. If you cannot find anything about the source in that time, that is itself a finding. A legitimate news organisation, a government body, or a known company will have a digital footprint. The absence of a digital footprint is a red flag. The effort required is low: a few taps, a few keystrokes, and a glance at the search results. You are not uploading anything, just searching.

Lateral reading replaces an older method still taught in some schools: the CRAAP test. CRAAP stands for Currency, Relevance, Authority, Accuracy, and Purpose. It was developed in 2004, in a different internet. The problem with CRAAP is that it asks you to evaluate the source by reading it, which is exactly what the professional fact-checkers do not do. A well-crafted fake news site can pass the CRAAP test: it has a recent date, a credible-looking design, and a stated purpose. Lateral reading is faster and more reliable because it outsources the evaluation. You do not have to decide if the source is credible. You find out what other credible sources say about it. The CRAAP test puts the burden on your judgment, which is precisely the thing the study showed to be unreliable. Lateral reading is the professional technique because it does not trust your instincts. It trusts your search skills.

Where To Verify In Singapore

The SIFT method is not just for checking news articles from strangers. Use it for the forwarded note from your aunt about a new scam in Singapore, the viral photo of a crowd at a protest, and the video of a politician saying something absurd. The verification reflex, opening a new tab before believing, is what separates a person who is merely online from a person who is media literate. The Stanford History Education Group frames this as three questions of Civic Online Reasoning: Who's behind the information? What's the evidence? What do other sources say? Lateral reading answers all three. You find out who is behind the information by searching for them. You find the evidence by tracing the claim to its original context. You find out what other sources say by searching for the claim on a verification organisation's website. In Singapore, the government portal (on the gov.sg domain and the @govsg_factually Telegram channel) publishes clarifications on false claims circulating locally. Black Dot Research, an independent organisation and signatory to the International Fact-Checking Network (IFCN) Code of Principles, publishes local assessments. The Straits Times' CheckMate column and CNA's Sure Anot do the same in the mainstream media.

A Scam Alert In Practice

You receive a forwarded note on WhatsApp claiming the Singapore Police Force has issued a new scam alert. The note contains a logo and a warning. First, stop. Second, open a new tab and search 'Singapore Police Force scam alert'. If the claim is real, the official website or the official social media account will show it. If it is fake, a verification organisation has probably already debunked it. The claimed real-world enforcement decisions by the Personal Data Protection Commission (PDPC) are public; check the official website. Look for actual endorsement from a government website, not a screenshot. A screenshot can be faked. An official URL cannot. If the note tells you to forward the warning to everyone you know, that is a classic sign of a hoax. The real police force does not ask you to forward warnings; they ask you to check their website. A claim that is on the government portal is likely to have been there for hours. A claim that is not there is likely false.

The One-Glance Domain Check

When you are tired, in a hurry, or in a group chat where everyone is sharing the same story, the normal route, opening new tabs, checking Wikipedia, searching for the claim, is still the right route. But there is a faster check that fits in a single glance. Look at the domain of the link. If it is a long string of random letters and numbers, or a variation of a real domain like 'facebok.com' instead of 'facebook.com', stop. The officialdom halo is a powerful psychological trick: a note that uses a government logo, official-sounding language, or a URL that looks like .gov.sg will get your trust without you checking. The check is simple: does the domain actually end in .gov.sg? Not '.gov.sg.evilsite.com', but '.gov.sg' before the slash. A source with 'gov.sg' in the middle of the URL is not a government site. The note claims verification by the Singapore Police Force or the Ministry of Health; checking the official website or social media account shows no such advisory was issued. That is the entire check in one glance.

AI Images And Reverse Search

Lateral reading also protects you from the AI realism trap. A photorealistic image of a politician shaking hands with a foreign leader, or a video of a celebrity saying something offensive, might be generated by an AI tool. The visual tells change with every model update, so there is no durable way to spot an AI-generated image by looking at it. The only reliable check is the reverse image search. Upload the image to Google Images, TinEye, or Yandex Images. If the image has appeared before in a different context, or if the search results show it being used to illustrate a different story, you have your answer. The effort is low but the payoff is high. A reverse image search traces the claim to its original context. It is the SIFT method's 'Trace' step applied to an image instead of a sentence. The digital footprint of an image, where it has appeared before, how long it has been circulating, who else has used it, is a chain of evidence that you can examine without leaving your messaging app.

Law And Platform Rules Are Not Your Check

The habit of lateral reading is not about being a professional fact-checker. It is about being a competent citizen of the internet. The Singapore government's Online Criminal Harms Act enables authorities to issue directions to remove or block access to online content used for criminal activities, including scams. The Online Safety Code for Social Media Services, issued by the Infocomm Media Development Authority (IMDA) on 18 July 2023, requires designated social media services, Facebook, Instagram, TikTok, X (Twitter), YouTube, and HardwareZone, to minimise exposure to harmful content. But these are legal mechanisms, not personal ones. A POFMA direction is not the same as a verification article, though both address falsehoods. The law can take down a page; it cannot stop you from believing a false claim. The only thing that stops you from believing a false claim is the verification reflex. Platforms remove content under their terms of service; the police act under the law. Neither is your personal fact-checker. That is your job.

Avoid The Single-Source Trap

The permanent risk in lateral reading is the single-source settlement. You find one source that confirms your belief, and you stop. The professional fact-checkers did not do this. They opened multiple tabs, checked multiple sources, and looked for corroboration. If you have opened one new tab and found a page that agrees with you, you are not done. Check whether that source is independent. A source that copies and pastes from the same press release is not independent. A verification organisation that publishes its methodology and funding, as required by the International Fact-Checking Network code, is a better source than a blog that does not. The source transparency of a check is crucial: does it name the original source and show the evidence chain? A verdict without visible evidence is unverifiable. The methodology matters too: does the organisation publish its process? If it does not, it is less credible. A claim that was true last year may be false this year.

When The Search Comes Up Empty

You will not always have a strong signal on your first search. A newly established website with no digital footprint, a claim too obscure to have been checked, or a screenshot too low-resolution for a reverse image search to match. In these cases, the correct move is to do nothing. Do not share the note. Do not say 'I'm not sure, but...' and then share it anyway. The failure mode is single-source settlement: accepting a claim as true after finding one source that confirms it, without checking whether that source is independent. The effort of lateral reading is low, but it is not zero. The cost of a mistake is high. If you cannot verify the claim, the most powerful thing you can do is stay silent. The absence of a check is not a fact; it is just an absence. The claim will survive your failure to share it. The internet will not collapse because you did not forward a note. The only thing that collapses is your credibility.

Two Tools Worth The Extra Tap

Two verification tools are worth the extra tap. The first is metadata analysis. ExifTool and FotoForensics let you examine the hidden data in an image file: the camera model, the date, and the location. A 'forwarded' image stripped of its metadata may have been through a screenshot, which is a sign that someone is trying to hide something. The second is the search itself. A search for a key phrase from the note, wrapped in quotation marks, will show you where that exact sentence has appeared before. If it has appeared on a dozen different websites, all in the same wording, it is likely a piece of viral spam. The forwarding label on WhatsApp, 'forwarded as received', is a structural feature that will persist in some form. It is there because the platform wants you to know that the note is not from the person who sent it to you. Use that label as a trigger. When you see it, your verification reflex should kick in automatically.

Why The Method Works

Lateral reading is not a guarantee. You can follow the SIFT method perfectly and still be wrong. A fake news site can have a long digital footprint, a well-known fact-checker can make a mistake, and a genuine news article can be completely false. The method does not promise certainty; it promises a higher probability of being right. The difference between a professional fact-checker and a normal reader is not that the professional is never wrong. It is that the professional knows when they are wrong. They have a system for checking, and they use it. The Wineburg and McGrew study showed that the fact-checkers were not smarter than the historians. They were just more humble. They assumed they could be fooled, so they did not trust their reading. They left the page, opened new tabs, and checked. The most dangerous page on the internet is not the one that is obviously false. It is the one that is almost true, well-written, and well-sourced. That page can only be caught by a method that leaves the page.

Trace Every Claim To Its Origin

The SIFT method's 'Trace' step is where you find the original context. This is about the claim, not the source. If the claim is a quote, find the original interview. If it is a statistic, find the original report. If it is an image, find the original photograph. The goal is to see if the claim has been altered or misrepresented. A politician's quote taken out of context can make them look like they said the opposite of what they meant. A statistic from a study that was later retracted is not a true statistic. A photograph from a different event, used to illustrate a current story, is a lie. The reverse image search is the tool for the third case. The quote's origin is the tool for the first. The primary source is the goal for the second. You are looking for the primary source, not a secondary account of it. A secondary account is someone else's version of the truth, with their own biases. The primary source is the raw material. Corroboration between multiple primary sources is the bedrock of the check.

Trust The Network, Not The Text

When you open new tabs and check the source, you are doing the same thing a journalist does when they call someone to confirm a story. You are not trusting the text; you are trusting the network. The strength of the network is the strength of your verification. A claim confirmed by the government portal and by an independent organisation is stronger than a claim that is only on a blog. A claim denied by the official source and confirmed by no one is a claim you should not share. The digital footprint of a claim is a trail you can follow. The trail may go cold, but the absence of a trail is itself a finding. The forwarder's trust in the person who sent the note is not evidence. It is the single-source settlement. It is the reason scams work. A scam works because it uses a trusted name, a bank, a government agency, a delivery company, to lend credibility to a false note. The officialdom halo is the scammer's best friend. The verification reflex is your only defense.

Reverse Image Search Workarounds

You will hit a specific failure when you start doing lateral reading on a phone, particularly with a forwarded image. You open Google Images, tap the camera icon, and try to upload the image from your camera roll. It fails. The file is too large, or the format is unsupported, or you are in a messaging app and the image is not saved to your phone. The failure is not the search itself; it is getting the image into the search engine. The workaround: take a screenshot of the image, which strips the metadata and reduces the size, then search for the screenshot. Screenshotting a low-resolution image makes it even lower resolution, but it is better than nothing. A second workaround: describe the image in the search bar. Type 'crowd protest [city name]' and see what comes up. A third workaround: use a different search engine. Yandex Images is often better at finding the original source of a photo than Google Images, especially for photos from non-Western sources. Do not give up at the first failed attempt.

Build The Reflex

Lateral reading has a real cost, and it is a cost in attention. It is easier to stay on the page, read the story, and believe it. It is harder to stop, open a new tab, and do the work. The professional fact-checkers in the Stanford study were not faster because they were smarter; they were faster because they had a better system. The system is the SIFT method. It is not a magic bullet. It will not catch every lie, and it will take more time than doing nothing. But the time you spend lateral reading is time you are not spending being wrong. You are not just verifying a claim; you are building a skill. Every time you leave the page, you train your verification reflex. Every time you check the source, you build a habit. Every time you trace a claim, you get better at spotting the patterns of disinformation. The next time you see a forwarded note with a government logo and a sense of urgency, you will not feel panic. You will feel calm. You know what to do.

The Honest Limit

No method is foolproof. Lateral reading reduces the risk of being fooled; it does not eliminate it. A determined and well-resourced disinformation campaign can create fake news articles, fake social media profiles, and fake verification websites. They can make a source look credible to a search engine. They can game Wikipedia. They can even hack a real journalist's account and post a false story. The SIFT method will not protect you from a sufficiently sophisticated actor. What it will do is protect you from the 99% of misinformation that is not sophisticated. The amateurish scam, the out-of-context quote, the doctored image, the fake government warning, these are the things lateral reading catches. The modern disinformation ecosystem is mostly noise. The signal is rare. Your job is to find it. You will not always succeed. But you will be wrong less often, and you will be able to explain why. And that is the whole point.

How Professional Fact Checkers Use Lateral Reading to Leave the Page and Judge the Page