Lateral Reading

Lateral Reading

Mike Caulfield and Sam Wineburg use these examples in Verified to explain how to investigate an unfamiliar source. Organization descriptions, legal counts, sample search results, and search-engine terminology reflect the book’s context.

Key practices

• Anticipate the kinds of results you’re looking for before searching. A mismatch between your expectations and the search results may signal that you need to rethink your strategy.
• Before you click on a result, engage in result mining. Read the snippets that accompany each result—you can learn a lot without ever leaving the search page.
• Specific, unique keywords lead to your desired results much faster than general terms.
• A search engine can locate factual information, but ranking is not a verdict on truth. Check the sources behind the results.
• Search terms influence which results you see. Try neutral wording and alternative descriptions rather than building your preferred answer into the query.
• Rather than spending precious minutes on an unfamiliar website, get off the page. Practice the I in SIFT—investigate the source—by reading laterally.
• Put the name of the group or individual into a search and see what the rest of the web has to say. Use the web to read the web.
• If something looks fishy, lean into the F of SIFT and find a better source. Don’t bother with a source that can’t pass a basic lateral check.
• Once you have a list of results, resist the temptation to click the first thing that catches your eye. Practice click restraint by scanning the full list to get a sense of the information neighborhood, then make a wise first click.

4. Using the Web to Read the Web

You and your friends have complained for years about the bullying at your hometown school. Nothing has been done—until now. The school board is finally ready, and you’ve been nominated by school staff to present the issue.

You’ll have one shot. You want to make sure the school board doesn’t just do something about bullying; you want it to do something that works. You dust off those newly learned research skills and google “bullying at school.”

Before you can even click, Google offers ten autosuggestions. “Bullying at school never acceptable,” the first item in the list, sounds about right. It lands you on the site of the American College of Pediatricians.

“No child should be harassed for his or her unique characteristics.” Check. “Schools should encourage an environment of respectful self-expression for all students.” Yes.

Weighty references accompany the article: “A Multilevel Examination of Peer Victimization and Bullying Preventions in Schools” in the Journal of Criminology. “Association between Bullying and Psychosomatic Problems: A Meta-Analysis” in Pediatrics.

You skim the rest. It’s science-y. Headings include “Abstract,” “Forms of Bullying,” “Prevention,” “Conclusion”; it’s signed by an MD and accompanied by this statement: “The American College of Pediatricians (ACPeds) is a national association of licensed physicians. . . . The mission of ACPeds is to enable all children to reach their optimal physical and emotional health and well-being.”

Before you put your faith in this site, consider the following facts about the American College of Pediatricians:

• • It is labeled a “fringe anti-LGBTQ hate group that masquerades as the premier U.S. association of pediatricians” by the Southern Poverty Law Center.
• • It promotes “conversion” therapy for gay kids, a procedure outlawed in twenty states, the District of Columbia, and all of Canada.
• • It publishes vile and defamatory content, like a post advocating adding “P” for pedophilia to “LGBT” because, it falsely claimed, pedophilia is “intrinsically woven into [the] agenda” of LGBTQ activists.
• • It is condemned by Francis S. Collins, MD, the former director of the National Institutes of Health, for an “ideology that can cause unnecessary anguish and encourage prejudice.”

You probably couldn’t find a worse resource for helping kids who are getting bullied at school—especially LGBTQ kids. Compared to their heterosexual peers, LGBTQ kids are two to four times more at risk for thinking about and attempting suicide during adolescence, a risk to which ridicule and abuse at school can contribute.

You may have assumed that the American College of Pediatricians was a large professional organization, dispensing authoritative information to doctors and the general public. Actually, it’s a tiny splinter group of conservative doctors (estimates range from two hundred to five hundred) who, in 2002, broke from the sixty-seven-thousand-member American Academy of Pediatrics over the issue of adoption by same-sex couples.

Now that you know a bit about the American College of Pediatricians, can you spot the subtle persuasion in the abstract from the “Bullying at School: Never Acceptable” article—the phrase “no group should be singled out for special treatment”? You’d never learn from this site that LGBTQ kids are the main targets of bullying in school. The American College of Pediatricians doesn’t want gay kids “singled out for special treatment” because doing so, it fears, might push “temporarily-confused adolescents into adopting an atypical lifestyle.”

It’s easy to get duped by the site’s professional look, its official-sounding name, its formal logo, and its detached scientific tone. If you got fooled, you’re in good company. When researchers showed this site alongside that of the American Academy of Pediatrics to a group of Stanford students, the majority chose the splinter group as more reliable. Even professors at five different colleges wavered when asked to choose which site was more trustworthy. One professor confidently stated that the splinter group’s site was “just a useful resource for people to learn about bullying.” He had no inkling that something was amiss.

That professor fell into the same trap as Jill. In a parable that illustrates the point, Jill finds a bottle bobbing in a lake. Inside is an article with the headline “N95 Masks Don’t Work.” Jill does what she learned in school: she evaluates the text itself. The tone is scholarly. There are footnotes to impressive-sounding journals. The author has an “NMD,” which sounds medical. The article overflows with data, statistics, and charts. At its core lies a seemingly bulletproof argument: the weave of an N95 mask is 0.3 microns, while the coronavirus is only 0.1 micron, so the virus will “sail through that mask like a marble through a chain-link fence.” Jill sits and sorts through the evidence, coming to her own conclusion that N95 masks don’t work. But here’s the problem: Jill got her information from a bottle floating in a lake. She never asked who wrote the article or what their agenda might be. Jill is all of us when we let a professional-looking page suck us into analyzing content before we understand the source.

Lateral reading works because the web is a web—a galaxy of electronically connected sources. To understand a node in a network, you probe its connections to other nodes. If you wanted to know how spiders catch prey, you wouldn’t just stare at a single strand; you’d see how that strand connects to others to form an elaborate geometric pattern. For an unfamiliar website, examining connections can provide context that its own presentation omits. You evaluate a single website by seeing what the rest of the web has to say. You use the web to read the web.

Here’s how that works with the American College of Pediatricians. You land on the site, but before giving it your full attention—and before you risk falling down a rabbit hole by reading articles that con you with scientific language and fancy references—you open a new tab and search the group’s name. Immediately you see something’s askew. Wikipedia calls the organization a “socially conservative advocacy group” with one employee. The American Civil Liberties Union labels it a “fringe group.” Even without knowing that glma.org is the site of GLMA: Health Professionals Advancing LGBTQ Equality, you get clues from a snippet that says the group is “not to be confused with the 60,000 member American Academy of Pediatrics.” Psychology Today tells you the American College of Pediatricians “is an anti-LGBT group.” From there, the picture only gets darker.

The shift from vertical reading (dwelling on a single page) to lateral reading (searching the source across the web) makes an enormous difference. Consider a second example: research into minimum wage policy. Does raising the minimum wage increase the standard of living, or does it cause prices to rise and hurt the very people it’s supposed to help? In the book’s example, a prominent search result is epionline.org, run by the Employment Policies Institute.

The site piles on cheap signals of credibility: a professional layout in tasteful maroon and gray, a nonprofit designation from the IRS, a physical address and contact number, a dot-org URL, and a “Research” tab listing reports by professors from the University of New Hampshire, University of Connecticut, University of California, Irvine, Carnegie Mellon, and others. Rather than clicking on any of those research reports or dwelling on the “About” page, a skilled fact-checker does something different. He leaves the site in a split second and googles “Employment Policies Institute.” In half a minute he has seven tabs open. One, from the online magazine Salon, wastes no time calling out the group: “Corporate America’s New Scam: Industry P.R. Firm Poses as Think Tank.”

The Employment Policies Institute dresses up in the costume of “nonpartisanship.” In actuality, it’s the handiwork of a public-relations firm, Rick Berman and Company, which specializes in cooking up websites for corporate and political clients. As for the research reports that conveniently stack up against raising the minimum wage? One of the professors received a $180,000 grant from the institute to fund his study. The funding and organizational connections give readers reasons to examine the study’s methods, disclosures, and independent assessments. They do not, by themselves, establish that its data or conclusions are false.

When you dwell on a site before you know who’s behind it, you play into the hands of the Rick Bermans of the world. But when you marshal the web’s web-like qualities, you’re the one in the driver’s seat. As journalist Steve Daly put it, “Think of lateral reading as sending up multiple signal flares in a barrage instead of one at a time. The bombardment lights up a wide swath of terrain. It lets you see quickly which part of this landscape might provide the fastest gauge of reliability.”

The big message of lateral reading is: get off the page. Once you’ve entered your keywords and stare at the results, how do you decide which result to click? Many of us never look beyond Google’s first few results—some studies have found that over 50 percent of searchers never go beyond the third result. There’s a joke among disgruntled web advertisers: “Where’s the best place to hide a dead body? Page 2 of the search results.”

Google tells us that by using the criteria of expertise, authoritativeness, and trustworthiness (E-A-T), it tries to array the most relevant results at the top of the search. And that sort of works—sometimes, for certain subjects, given the right keywords. But Google’s search algorithms betray the flaws of human design: racism, sexism, ableism, antisemitism, and every other ism that afflicts humankind. (To its credit, Google generally tries to fix these problems when they are pointed out.) Each year the company adjusts its algorithms with hundreds or thousands of tweaks. Yet the process is never-ending because Google is locked in a cat-and-mouse game with legions of site owners trying to outgame the system by manipulating keywords, planting links on other sites, spamming, reconfiguring content, and doing everything imaginable to get a leg up on the competition. The business of pushing results to the top—search engine optimization—is an eighty-billion-dollar-a-year industry focused on a singular goal: landing a site on Google’s prime real estate while kicking the competition to page 2 or worse.

That’s why you must practice click restraint. Before you click on anything, scan the full list of results. Get a sense of the information neighborhood. Look beyond the first three entries and consider what the snippets and URLs tell you about the credibility and perspective of each source. A wise first click—informed by a quick survey of the landscape—saves you from tumbling into a site that has gamed its way to the top.

Once you click, let lateral reading become your default. Investigate the source immediately. Put the name of the group or individual into a search alongside relevant keywords, and let the web tell you who they are and what others think of them. If the source turns out to be unreliable, move on. Find better coverage. Compare independent sources and trace important claims to their evidence. Several pages repeating the same assertion may still lead back to a single unsupported source.

A source check helps decide where to spend your attention. Leave an unfamiliar page long enough to establish who is behind it, then examine the evidence relevant to the claim. If the background check reveals serious problems, look for better coverage.