You have eleven seconds and a forwarded message that says 'URGENT: Your SingPass will be suspended unless you click this link.' Your thumb is already moving toward the share button because your aunt sent it. Your aunt would never send you something false. This is the moment source evaluation stops being an academic exercise and becomes a matter of whether you just amplified a phishing attempt to forty-three people. Most of us were taught to evaluate sources the slow way: write down the acronym, check the date, ask who published it, decide if it looks trustworthy. That approach fails exactly when you need it most. You do not have eleven seconds; you have the time it takes for your finger to twitch. The researchers who studied professional fact-checkers found that they do not do this at all. They open a new tab, leave the page, and check who is behind the claim before they read a single sentence of it. That difference, between staying on the page and leaving it, is the entire argument between two competing frameworks that most people have never heard of, even though they have been arguing with each other for the better part of a decade.

CRAAP test origin: Sarah Blakeslee, California State University, Chico, 2004
CRAAP acronym: Currency, Relevance, Authority, Accuracy, Purpose
SIFT method origin: Mike Caulfield, 2017
SIFT acronym: Stop, Investigate the source, Find better coverage, Trace claims to original context
Fact-checker lateral reading time: Average 6 minutes per source, vs 4 minutes for historians and 2 minutes for Stanford undergraduates (Wineburg & McGrew, 2017)
Fact-checker accuracy: Significantly higher than historians and undergraduates across all source evaluation tasks (Wineburg & McGrew, 2017)
Singapore S.U.R.E. framework: Source, Understand, Research, Evaluate; National Library Board programme

The CRAAP Test: A Checklist for Vertical Reading

What The Acronym Asks You To Do

The CRAAP test was invented in 2004 by Sarah Blakeslee at California State University, Chico, as a checklist for students who needed permission to slow down. Its five letters stand for Currency, Relevance, Authority, Accuracy, and Purpose. Each one is a question you ask about the document in front of you. Is it current? Does it relate to your topic? Who wrote it and what are their credentials? Is the information supported by evidence? Why does this source exist, to inform or to persuade? The test forces a kind of vertical reading: you stay on the page and interrogate it top to bottom, checking the URL, the date, the author bio, the citations, and the tone before you let it into your paper. Its strength is that it teaches a deep source analysis that builds a habit of scepticism.

Where The Checklist Breaks Down

Its weakness is that it is slow. It also assumes the source will honestly answer your questions, which is exactly what a well-crafted fake news site will not do. Researchers have criticised CRAAP for encouraging you to spend time on the source itself rather than leaving it to consult external information. Mike Caulfield first made this critique in 2017. The Stanford study by Wineburg and McGrew confirmed the problem: professional fact-checkers did not use CRAAP at all. They were right not to, because vertical reading fails when the page is designed to deceive you.

The CRAAP test limitations become obvious the moment you try to apply it to a claim that arrives in a messaging app. You open the message, you see a screenshot of a headline, and you ask the five questions in your head. How long does that take? Thirty seconds if you are fast, two minutes if you are actually checking. That is two minutes you did not spend leaving the page to see whether the outlet even exists, whether the quote is real, or whether the image has been photoshopped. The test gives you false confidence. It feels rigorous, so you share the claim with a comment like 'not sure if true, but the date checks out'. You have just become part of the problem. The other failure is that CRAAP asks you to judge authority. Authority is exactly what a skilled forger can fake. A fake news site can have a professional layout, a plausible author bio, and a made-up list of credentials. You will check all five letters and give it a passing grade because you never left the page. The research is unambiguous: fact-checkers do not do this. They open a new tab and search for the source itself. They do it in an average of six minutes per source, compared to four minutes for historians and two for Stanford undergraduates, yet they were significantly more accurate than both groups. Speed and accuracy came together because they were doing the right thing, not because they were naturally gifted.

The SIFT Method: Lateral Reading for a Fast-Moving Internet

Four Moves, Not Five Questions

Mike Caulfield, a research scientist and the author of the SIFT method, built his framework in 2017 on the back of the Stanford History Education Group research that showed what fact-checkers actually do. His acronym stands for Stop, Investigate the source, Find better coverage, Trace claims to original context. Each step is a behaviour, not a question. Stop means pause before you engage, notice your emotional reaction, and recognise when a source is unfamiliar. Investigate the source means you do not read it yet; you open a new tab and search for what other people say about it. Find better coverage means you do not argue with a bad source; you search for a trusted outlet that has already covered the same claim. Trace claims to original context means you go back to the original image, quote, video, or data to see if it has been cropped, edited, or stripped of its surrounding paragraph. This is lateral reading. It is the single most effective verification behaviour ever documented in a research setting.

Who Adopted It And When

The SIFT method was adopted into the Stanford Civic Online Reasoning curriculum in 2019, recommended by the Google News Initiative in 2020, and included in the American Library Association's 'Fighting Fake News' toolkit the same year.

The beauty of SIFT is that it does not care whether you are evaluating a scholarly article or a WhatsApp forward. The moves are the same. They take less time than you think. In the time it takes you to read a single paragraph of a suspicious article, you could open a search tab, type the source's name plus the word 'scam' or 'fake', and scan the first three results. That is the entire 'Investigate the source' step. It is over in under a minute. The 'Find better coverage' step is even faster: search for the exact claim in quotation marks and see which reputable outlets have covered it. If the claim is true, someone has reported it; if it is false, a fact-checker has written about it. The 'Trace claims' step is the most effort, since it requires you to locate the original image or quote. It is also the step that catches the most sophisticated fakes. The Stanford study measured fact-checkers taking six minutes per source on average. That sounds slow until you compare it to the two minutes a Stanford undergraduate spends before concluding that the source is probably fine. Six minutes now saves you the six hours you would spend defending a false share to your group chat.

Source Evaluation Framework Comparison: User Burden, Speed, and Depth

Three Trade-Offs To Weigh

When you put the CRAAP test and the SIFT method side by side, the source evaluation framework comparison boils down to three trade-offs. The first is user burden: how many steps must a person take to verify a claim using the framework. CRAAP asks you to evaluate five criteria on the page itself, which means you are doing heavy cognitive work before you have even checked whether the source is real. SIFT asks you to do four moves, but each one is lighter, and the second and third moves can be done in parallel by opening multiple tabs. The second trade-off is verification speed: how many minutes from encountering a claim to reaching a verdict. SIFT is unambiguously faster for news claims because its default is to leave the page; CRAAP's default is to stay on it. The Wineburg and McGrew data shows that the fastest, most accurate evaluators were the ones who went lateral, not the ones who read deeply. The third trade-off is educational depth: which framework teaches you more about the nature of information itself. CRAAP teaches you to interrogate a single document, which is valuable when you are writing a research paper and need to defend your sources. SIFT teaches you that no source is an island, which is more valuable for your life as a citizen and a consumer of news.

Do not mistake the comparison for a contest where one framework wins. The CRAAP test is the right tool for the academic setting where you have time, you are writing a paper, and you need to demonstrate that you have thought carefully about each source's quality. Its structure is a gift to students who have never been told what 'critically evaluate' means. It gives them five specific questions to answer. The SIFT method is the right tool for the quick claim, the forwarded message, the breaking news headline that feels designed to make you angry. Its structure matches the way information actually travels: fast, fragmented, and stripped of context. A reader who ignores CRAAP entirely will struggle in a university seminar where the professor asks them to defend the relevance of a source from 1998. A reader who ignores SIFT entirely will share a deepfake video of their prime minister and never know what hit them. The honest answer to 'which should I use' is: learn both, and pick based on the situation, because the situation dictates the method. If you have hours, use CRAAP. If you have eleven seconds, use SIFT. If you are fact-checking, you are already using SIFT, whether you know it or not.

When to Use CRAAP vs SIFT: A Decision Guide for Real Situations

The Academic Paper

The 'when to use CRAAP vs SIFT' question has a practical answer that has nothing to do with ideology and everything to do with the consequences of being wrong. Imagine you are a university student writing a history essay on the fall of Singapore in 1942. You have a reading list, you have time, and you need sources that are authoritative, accurate, and peer-reviewed. This is the CRAAP test's home turf. The five questions map directly onto the criteria your professor will use to grade you. You need currency: a history of the Malayan campaign written in the 1950s will be outdated. You need authority: a blog by a retired army officer is not the same as a monograph from the National University of Singapore Press. You need accuracy: a source that gets the date of the surrender wrong is worse than useless. In this situation, using SIFT is a mistake. SIFT's lateral moves will send you to news articles and fact-checks that have nothing to say about the historiographical debate.

The Group Chat At 9pm

Now imagine the opposite situation. You are in a messaging app on a Saturday night, and your brother has just sent you a video that claims a local supermarket chain has been selling fake rice. You have no time, no stakes, and no obligation to be a careful scholar. The CRAAP test would take you three minutes of vertical reading on a page that is probably fake. SIFT takes you twenty seconds of lateral reading that reveals the video is from a prank channel in Thailand. You have just chosen the right framework for the job. You did it without thinking.

The failure case is the reader who does not have a choice. They are in the middle of a group chat where a claim is spreading and everyone is looking at them for a verdict. They do not have time to run the five CRAAP questions. They do not have the emotional detachment to remember all four SIFT moves. What do they do? They use the smallest version of SIFT that still works: they stop, and they do not share. That is it. The entire method reduces to one reflex that separates the person who spreads misinformation from the person who stops it. You can learn all the frameworks in the world, but if you do not have the self-control to pause before your thumb hits send, the frameworks are decoration. The researchers who studied fact-checkers noted that the most important skill was not speed or accuracy. It was the willingness to say 'I do not know yet'. A reader who shares a claim with a warning label is still amplifying it; a reader who shares nothing buys time for the truth to catch up. The next time you are in that situation, try this: type 'hold on, let me check' and send that instead of the video. The chat will not collapse. Your reputation will not be damaged. You will have done the only thing that matters: not contributing to the noise.

Vertical Reading vs Lateral Reading Evaluation: The Research That Settles It

What The Stanford Study Found

The 'vertical reading vs lateral reading evaluation' debate is not a matter of opinion; it is a matter of data. The data comes from a landmark study by Sam Wineburg and Sarah McGrew at the Stanford History Education Group, published in 2017. They asked professional fact-checkers, historians, and Stanford undergraduates to evaluate a series of online sources while their browsing behaviour was tracked. The results were stark. The fact-checkers were significantly more accurate than both other groups on every single task. They achieved that accuracy by working faster, not slower. The fact-checkers spent an average of six minutes per source, the historians four, and the undergraduates two, yet the fact-checkers were the ones who got it right. How is that possible? Because the fact-checkers were reading laterally: they would open a new tab, search for the source's name, scan the results, and then decide whether to return to the page. The historians and undergraduates mostly engaged in vertical reading. They stayed on the page and evaluated its design, its language, its internal consistency. The undergraduates were particularly susceptible to a professional-looking design, rating a fake environmental organisation as credible because its website was polished. The fact-checkers were not fooled. They never trusted the page in the first place. They knew that a website is not a source; it is a claim about a source, and the only way to verify it is to leave it.

The practical takeaway from the research is that vertical reading is not 'the wrong way'; it is 'the way that does not scale'. When you have a single source and unlimited time, vertical reading gives you a deep understanding of that source's strengths and weaknesses. When you have an unknown source and limited time, vertical reading is a gift to the people who are trying to deceive you. The Wineburg and McGrew study demonstrated that fact-checkers do not read less carefully; they read differently. They start with the source, not the content. They use the vast, messy, self-correcting ecosystem of the internet as their reference library. A reader who wants to adopt the fact-checker's mindset does not need to become a professional. They need to learn one habit: when you encounter a claim, your first move is to leave the page, not to read it. Open a new tab, search for the source's name plus a word like 'fake', 'scam', or 'about', and read what other people say. If you find a fact-check that debunks the claim, you are done. If you find a Wikipedia article about the source, you have your answer. If you find nothing, that absence of information is itself information: a source with no digital footprint is a source you should not trust.

The S.U.R.E. Framework: Singapore's Answer to Information Literacy

How The Acronym Works

Singapore's National Library Board developed its own framework, called S.U.R.E., which stands for Source, Understand, Research, Evaluate. It is the closest thing to an official government-endorsed media literacy curriculum. The S.U.R.E. framework is taught in schools, public libraries, and universities, and it was designed to be compatible with both the CRAAP test and the SIFT method. The 'Source' step asks you to identify who created the information; the 'Understand' step asks you to grasp the meaning and context; the 'Research' step asks you to verify the facts; and the 'Evaluate' step asks you to form a judgement. Look closely and you will notice that S.U.R.E. is essentially a bridge between CRAAP and SIFT: it includes the CRAAP test's concern with authority and accuracy, but it also incorporates the SIFT method's emphasis on verification.

A Decade Of Classroom Use

Since 2021, the National Library Board has updated its workshops and guides to include lateral reading concepts. A Singaporean schoolchild learning about S.U.R.E. is also learning the same moves that professional fact-checkers use. If you are a reader outside Singapore, the S.U.R.E. framework is still worth knowing about. It is one of the few state-supported information literacy programmes in the world that has been running for over a decade without being co-opted by political interests.

The connection between S.U.R.E. and the two Western frameworks is not accidental; it is a deliberate design choice. The National Library Board wanted a programme that would work for Singapore's multilingual, multiracial population. It recognised that neither CRAAP nor SIFT alone was sufficient. CRAAP was too slow for the average social media user. SIFT was too unstructured for someone who wanted a clear set of steps to follow. S.U.R.E. splits the difference: it gives you a memorable acronym, just like CRAAP, but it points you towards lateral moves, just like SIFT. The 'Research' step is the one that matters most. It is the step that forces you to leave the page and check the source against external evidence. The 'Evaluate' step then asks you to weigh that evidence and make a decision, which is something that neither CRAAP nor SIFT does explicitly. If you are a teacher or a parent looking for a framework to teach a teenager, S.U.R.E. is a better starting point than either of the Western models. It was designed for a generation that lives on their phones and trusts forwarded messages more than they trust newspapers. The National Library Board has published a range of free resources, including posters, worksheets, and online tutorials, all of which are available on its website.

Information Literacy Framework Singapore: How S.U.R.E. Fits In

The Legal Backdrop

The broader information literacy framework Singapore has built around S.U.R.E. extends far beyond the classroom and into the country's approach to online safety and regulation. Singapore's Infocomm Media Development Authority issued a Code of Practice for Social Media Services that took effect in 2023, requiring designated social media services to minimise exposure to harmful content, including cyberbullying and content facilitating scams. This is a top-down approach to a problem that most countries are trying to solve with education alone. It reflects a uniquely Singaporean philosophy: the state sees itself as responsible for protecting citizens from harm, not just from foreign interference but also from each other. The Media Literacy Council, a charity formed in 2012, runs public education campaigns that complement the legal framework. It works closely with the National Library Board to promote S.U.R.E. The result is a two-pronged strategy: the law punishes the worst offenders, and education tries to make everyone else resilient enough to spot the danger before it reaches the courts.

Factually And Its Limits

The Singaporean government has also created a fact-checking portal called Factually, run by the Singapore POFMA office. It publishes clarifications on false claims circulating locally. Factually is not an independent fact-checker; it is a government publication, and it is legally protected from defamation lawsuits by the Protection from Online Falsehoods and Manipulation Act (POFMA) that it serves. If you see a claim that a local minister is corrupt, and the government says it is false, you can check Factually to see the official response. But Factually is not a substitute for the SIFT method. It only covers claims that the government has decided to address. It does not provide the transparent sourcing that the International Fact-Checking Network code requires. A reader who wants to verify a claim about Singapore should use SIFT first. Lateral reading will find the news articles, the fact-checks, and the official statements. Only then should they consult Factually as one more source. This is not a criticism of Factually; it is a reminder that no single source, no matter how authoritative, should be the end of your inquiry. The S.U.R.E. framework's final step, Evaluate, asks you to make a judgement. That judgement is only as good as the range of sources you have considered.

Frequently Asked Questions

Is the CRAAP test still useful if I use SIFT instead? Yes, but only in the context where it was designed: academic research. If you are writing a paper, CRAAP gives you a structured way to defend your sources to a professor. If you are reading the news, SIFT is faster and more effective. Use both, but know which one you are using and why.

How long does it take to learn the SIFT method? Ten minutes to learn the four steps, and a lifetime to master the habit. The steps are simple: Stop, Investigate, Find better coverage, Trace. The difficulty is remembering to use them when you are emotionally triggered by a headline. Practice on low-stakes claims first, like a celebrity rumour, before you need it for a real crisis.

Does the SIFT method work for academic articles? Not directly, and you should not force it. A scholarly article is not a claim you need to verify; it is a source you need to evaluate. Use CRAAP for the article itself, and use SIFT for the sources it cites. The combination is powerful: CRAAP gives you depth, SIFT gives you breadth, and together they cover most of your inquiry needs.

What should I do if I am not sure whether a source is trustworthy? The honest answer is that you are not sure, and you should not pretend otherwise. The single most useful sentence in all of media literacy is 'I do not know'. Say it out loud, then open a new tab and check. If you cannot find a trustworthy source in five minutes, that is a strong signal that the claim is false, exaggerated, or missing important context.

One final note on the limits of any framework. The CRAAP test, the SIFT method, and Singapore's S.U.R.E. programme are all tools. A tool is only as good as the person wielding it. A skilled deceiver can build a website that passes all five CRAAP questions, can game the SIFT method's lateral moves by creating fake social media accounts, and can make a forwarded message look like it came from a trusted friend. The research is clear that no framework is foolproof. The people who are most confident in their ability to spot misinformation are often the ones most likely to be fooled. The Wineburg and McGrew study found that Stanford undergraduates were the most confident and the least accurate of the three groups. That should be a warning to anyone who thinks that being a digital native makes them immune. Verification is a skill, not a personality trait. It requires constant practice. If you use these frameworks once a year, you will lose the reflex. If you use them every day, on small claims and large, they become as automatic as checking the weather before you leave the house. The cost of the habit is a few extra seconds per claim. The cost of not having it is everything from a ruined reputation to a stolen life savings.

Evaluating Sources: The CRAAP Test Compared to the SIFT Method and When to Use Each