Contributions
- Introduces "lateral reading" as a core strategy used by expert fact-checkers to distinguish reliable from unreliable information online
- Contrasts expert approaches (rapid investigation across multiple sources) with students' and historians' reliance on close reading of individual pages
- Shows that domain expertise (being an experienced fact-checker or historian) does not guarantee accurate evaluation of online sources
- Demonstrates that knowledge of internet structure, search behavior, and organizational backgrounds is critical to credibility assessment
Method
The study employed a think-aloud protocol with three participant groups: 10 academic historians (PhD-holding faculty), professional fact-checkers working at news organizations (n=10), and 25 Stanford undergraduates. Participants completed six web evaluation tasks addressing social and political issues (bullying at schools, minimum wage policy, teacher tenure in California). Sessions were video-recorded and audio-transcribed. Researchers used rubrics to score the quality of participants' conclusions and coded their verbal protocols to identify evaluation strategies. Qualitative analysis focused on the different approaches each group used to assess trustworthiness.
Results
Fact-checkers' performance: Fact-checkers achieved near-perfect accuracy scores on evaluating two organizations' trustworthiness (American Academy of Pediatrics vs. American College of Pediatricians). They completed tasks in under 2 minutes on average and used lateral reading consistently—opening new tabs to investigate organizational claims, funding sources, and institutional affiliations before returning to the original site.
Historians' performance: Historians took 3–4 times longer than fact-checkers to reach conclusions and showed mixed accuracy. Two of ten historians resembled fact-checkers in their lateral reading approach; the rest relied more on close reading of the original site. Several were influenced by superficial features (logo appearance, professional design, .org domain) and made errors comparable to students.
Students' performance: By contrast, undergraduates rarely distinguished between the two organizations evaluating the same topic. They spent time engaged in "fluttering" (aimless clicking), read pages vertically and sequentially, and judged trustworthiness based on layout, graphic design, and official appearance. In the minimum-wage task, 60% of students incorrectly chose the less reliable site as more trustworthy.
Lateral reading as a key differentiator: Fact-checkers who took bearings (a navigation metaphor from Wineburg & McGrew) before diving into unfamiliar online territory learned substantially more in less time. When fact-checkers left a page early and searched independently for information about an organization's background, they arrived at warranted conclusions. Fact-checkers showed knowledge of digital architecture (how search results are organized, how sites attempt to appear authoritative) and institutional knowledge (recognizing funded advocacy as distinct from independent research).
Notes
Strengths: The paper makes a strong case that commonly taught close-reading strategies (scrutinizing text features, analyzing rhetoric) are ineffective online because they focus inward on a single text rather than outward on context and provenance. The "lateral reading" framework is intuitive and actionable. Videotaped think-aloud protocols provide rich evidence of expert reasoning.
Limitations: Sample size is small (45 participants total), and all undergraduate participants came from Stanford—a highly selective institution. The study did not test how easy lateral reading is to teach, nor did it examine whether generalists (e.g., librarians, Wikipedia editors) might develop similar strategies. Task-based studies constrain ecological validity; real-world evaluation may look different under time pressure, cognitive load, or emotional arousal.
Educational implications: The paper challenges curriculum models that emphasize close reading of web content and suggests instead that students need instruction in lateral reading: using browser tabs, search results, and external sources to verify claims. This reframes digital literacy from an individual-page skill to a navigation-and-context skill.