Wendover's Media Negativity Argument Has a Serious Problem: It Confuses a Distorted News Feed With a Distorted Society
Wendover Productions builds a compelling case that American journalism systematically exaggerates bad news. Its strongest evidence is a study reporting that 87% of COVID-19 stories from major American media outlets were negative, compared with 50% from major international outlets.
But the researchers behind that finding included a qualification the video never tells its audience: they could not establish that negative coverage produced the public-policy consequences the argument invites viewers to imagine. Their March 2021 paper specifically reports that American counties relying more heavily on major media were no more likely to reopen schools than comparable counties. The difference between measuring a media problem and proving that it changes society is enormous.
Cpb Us E1 Wpmucdn+1
The headline statistic is narrower than the video suggests
The researchers, Bruce Sacerdote, Ranjan Sehgal and Molly Cook, examined English-language reporting from January through December 2020. Their March 2021 revision covered three subjects: vaccines, rising and falling case counts, and reopening businesses, schools and public facilities.
The sample was not a representative survey of everything Americans watched or read. It focused on major media outlets, including The New York Times, USA Today, CNN, Fox News, ABC, CBS and NBC, and compared them with selected international sources. The researchers deliberately divided their sample roughly equally across three subjects and four media categories.
Cpb Us E1 Wpmucdn+1
There is also a version-control problem. The March paper's introduction still describes approximately 20,000 articles, but its methodology reports 43,000, with 42,753 observations in its main results table. The November 2020 NBER version actually used approximately 20,000 articles, but reported 91% negative American coverage against 54% internationally. Wendover combines the earlier sample figure with the revised percentages without explaining the discrepancy.
National Bureau of Economic Research+2
The measurement deserves attention, too. The researchers trained a machine-learning classifier using 200 articles that human readers had identified as strongly positive or negative. Phrases such as "death toll" and "stay at home" helped predict negative classification. They also used a separate negative-word dictionary, which produced similar broad results.
That second measurement strengthens the finding that American coverage was linguistically more negative. But neither measurement establishes that the reporting was inaccurate, sensationalized or harmful.
A journalist reporting an actual rise in deaths can produce a negative article while doing precisely what responsible journalism requires. A report explaining why people should stay home can receive the same negative classification while protecting lives.
Wendover repeatedly treats negative tone as evidence of negative journalism. They are not interchangeable.
The study contains evidence the video leaves out
The researchers found that American major outlets had an above-average interest in promoting what they called prosocial behavior, including mask-wearing and social distancing.
This is an important qualification. American coverage could be more negative while also providing more information encouraging protective behavior than international competitors.
The St. Louis Federal Reserve's November 2021 examination of the research highlighted precisely this problem. Its author, economist Guillaume Vandenbroucke, questioned whether the findings established that American outlets had actually over-reported negative information.
He pointed out that deciding whether coverage is excessive requires defining what the appropriate level of coverage should be. News does not necessarily need to mirror the prevalence of a disease to serve the public.
St. Louis Fed+1
Wendover's hypothetical comparison between a miraculous vaccine and a crime wave illustrates an attention problem, but it assumes that giving one story priority necessarily means neglecting the other.
Real editorial decisions are more complicated. A vaccine breakthrough may require scientific verification, an understanding of clinical trial results and an explanation of its limitations. A crime wave may represent an immediate threat to a community.
The relevant question is not whether a story sounds frightening. It is whether its importance, accuracy and prominence are proportionate to the facts.
The 87% statistic cannot answer that question.
The pandemic argument fails its own causal test
This is the most consequential weakness in the video.
Wendover argues that persistent negative coverage may have prevented people from recognizing improvements during COVID-19. It then develops a scenario in which public tolerance for restrictions is exhausted too early.
The narrator imagines a population willing to tolerate 12 months of mask mandates across a 36-month pandemic. If those restrictions begin too early, the public's limited tolerance might be exhausted before the most dangerous period.
The problem is that this is an invented illustration, not an established consequence of the research.
More importantly, Sacerdote and his colleagues actually examined the relationship between national media consumption and school reopening decisions.
Their findings were uncomfortable for the argument.
Initially, counties with greater consumption of national news media had lower school reopening rates. But after accounting for differences between counties, the apparent relationship weakened substantially. Their analysis found little evidence that negative national media coverage caused fewer schools to reopen.
The researchers explicitly concluded that the observed relationship was driven by selection rather than causality. Their estimates also ruled out a reduction in school openings greater than three percentage points from ten weeks of substantially higher national media exposure.
Cpb Us E1 Wpmucdn
That does not prove media coverage had no effect on any pandemic decision. The researchers did not test every mask mandate, school closure, border restriction or individual risk assessment.
But it directly undermines the video’s use of school reopening as evidence of the consequences of negative reporting.
Wendover takes a hypothetical explanation for public behavior and presents it alongside research that explicitly failed to establish the relevant causal relationship.
That is not a minor qualification. It changes what the audience should conclude.
Media influence is real, but tone is the wrong measurement
There is strong independent evidence that news coverage can change public behavior.
But it demonstrates why Wendover needs to distinguish negative reporting from misleading reporting.
A particularly useful comparison is the 2024 Political Analysis study by Elliott Ash, Sergio Galletta, Dominik Hangartner, Yotam Margalit and Matteo Pinna.
The researchers investigated the influence of Fox News during the early COVID-19 pandemic. They exploited differences in cable channel positioning to identify changes in viewership that were not simply the result of people choosing their preferred political news source.
They found that greater exposure to Fox News reduced social-distancing behavior and purchases of protective supplies. Their survey of 1,480 Americans also identified program-specific coverage that minimized the threat of COVID-19.
Cambridge
This is much closer to the kind of evidence Wendover needs.
It examines identifiable media exposure, specific messages and measurable behavioral consequences.
And the direction is revealing: the problem was not excessive negativity, but coverage that understated a genuine danger.
This does not prove that alarming journalism cannot cause harm. It demonstrates that the consequences depend on what the reporting actually communicates.
A negative headline can save lives by conveying a genuine danger. A reassuring headline can cost lives by concealing one.
The sentiment score alone cannot distinguish between them.
The public broadcasting explanation remains unproven
Wendover identifies the BBC in Britain and ABC in Australia as important differences between their media systems and America's predominantly commercial environment.
The argument is plausible: publicly funded broadcasters have less direct pressure to maximize advertising revenue and audience engagement.
But the original researchers themselves were more cautious.
Their March 2021 paper proposed three possible explanations for American media negativity:
- Other countries have influential publicly owned broadcasters.
- American media markets are less concentrated, potentially creating greater competition.
- The United States abolished its Fairness Doctrine in 1987, while the UK and Canada retain different forms of broadcasting regulation.
The authors explicitly described these as possible explanations deserving further investigation. They did not establish which explanation was responsible for the measured difference.
Cpb Us E1 Wpmucdn
Wendover concentrates on public broadcasting and builds a persuasive narrative around it.
But the research did not compare public and private broadcasters under otherwise equivalent conditions, establish that public funding caused less negativity, or measure what would happen if the United States introduced a stronger public broadcasting system.
The video also acknowledges the obvious weakness: public broadcasters depend on political institutions for funding and can face pressure from the governments they are supposed to scrutinize.
There is a genuine trade-off between commercial incentives and political independence. The evidence does not establish that one funding model reliably produces better journalism.
The evolutionary explanation goes further than the science
The video begins with a compelling account of prehistoric survival and eventually argues that negativity bias was evolved rather than learned.
It proposes that men benefited reproductively from taking risks to find food, while women faced stronger evolutionary pressure to survive pregnancy and protect children.
That is an evolutionary hypothesis, not a demonstrated explanation of modern psychological differences.
The 2019 PNAS study by Stuart Soroka, Patrick Fournier and Lilach Nir provides substantial evidence for an average human response to negative news.
It examined 1,156 participants across 17 countries and six continents. Participants watched seven BBC World News stories while researchers measured physiological responses.
The study found greater average physiological activation and attention during negative stories. But it also found considerable individual variation: only five countries showed a statistically significant average effect in one of its principal measures, and nine of seventeen did so in another.
PMC
Most importantly, the authors explicitly distinguish evolutionary explanations from cultural and institutional explanations.
They state that these explanations are not mutually exclusive and acknowledge that researchers do not yet know the extent to which negativity bias is culturally determined.
Wendover moves from that uncertainty to a much firmer conclusion: the bias was not learned; it evolved.
The evidence supports an average human sensitivity to negative information. It does not establish the particular prehistoric division of reproductive responsibilities proposed in the video, nor prove that modern differences between men and women originated through that mechanism.
The distinction matters because a plausible evolutionary story can sound like a scientific explanation long before the evidence establishes it.
The NordVPN advertisement simplifies the internet problem
There is another revealing example near the end of the video.
Wendover argues that national restrictions increasingly fragment the internet. The narrator describes American websites being inaccessible while travelling in Europe because of GDPR legislation, before introducing NordVPN as a solution.
The underlying event is real. In May 2018, numerous American news websites, including the Los Angeles Times and Chicago Tribune, blocked or restricted European visitors.
But the explanation needs correcting.
The GDPR regulates how organizations collect and process personal information. It applies to certain organizations outside Europe when they offer goods or services to people in the EU or monitor their behavior. It does not generally prohibit Europeans from reading American websites.
Your Europe
A 2022 study by Neil Thurman, James Sly, Bartosz Wilczek and Richard Fletcher examined what happened.
It found that publishers themselves chose to block European visitors rather than adapt their websites to the new requirements. Other American publishers, including The New York Times, complied without blocking access.
The researchers also found that websites that restored access subsequently experienced between 44% and 61% fewer monthly European visitors than before the restrictions.
Sage Journals
This distinction matters.
A European government did not simply erect a censorship wall around American journalism. Individual publishers made commercial and compliance decisions about whether European readers were worth serving.
NordVPN can change a user's apparent IP location and may help bypass some geographical restrictions. But that does not make the underlying problem government censorship.
The sponsorship is disclosed, and there is no evidence that NordVPN controlled the editorial argument. Nevertheless, the presentation turns a complicated publisher-compliance problem into a much simpler story about government restrictions on internet freedom.
That is a convenient framing for a company selling a VPN.
Closing verdict
Wendover identifies a genuine problem: major American news organizations produced substantially more negative COVID-19 coverage than their international counterparts.
The research deserves attention. The difference is large, the broad pattern survives alternative measurement methods, and commercial competition provides a plausible explanation.
But the video makes a series of consequential leaps.
It treats negative language as evidence of bad journalism, turns an observational media comparison into a theory of distorted public decision-making, and uses hypothetical pandemic scenarios to imply consequences that the underlying researchers could not establish.
It then elevates public broadcasting into a proposed institutional solution without testing whether that arrangement actually causes better coverage. Its evolutionary explanation is similarly more confident than the science permits.
The central problem is not that Wendover found the wrong statistic. It is that the statistic is asked to prove far more than it can.
A media system can be excessively negative. It can also be insufficiently alarmed, politically distorted, or commercially compromised. Measuring the number of negative stories does not tell us which failure is occurring.
And if we cannot distinguish accurate warnings from manufactured panic, we risk creating precisely the kind of distorted understanding of reality that this video claims to expose.
The video makes a strong case that news deserves scrutiny. It does not make a sufficiently strong case that negativity itself is the disease.
References
1. The System that Fuels Media Negativity — Wendover Productions, March 21, 2023. Original video and primary source for the arguments examined here.
2. Why Is All COVID-19 News Bad News? — Bruce Sacerdote, Ranjan Sehgal and Molly Cook, March 22, 2021. The revised study supplies the 87%-versus-50% comparison, its sampling and classification methods, and the critical qualification concerning school reopening.
3. Why Is All COVID-19 News Bad News? — NBER Working Paper 28110 — Sacerdote, Sehgal and Cook, November 2020. The earlier version documents the original 20,000-article sample and 91%-versus-54% finding; NBER identifies it as a working paper that has not undergone peer review.
4. Perspectives on U.S. Media Coverage of the COVID-19 Pandemic — Guillaume Vandenbroucke, Federal Reserve Bank of St. Louis, November 23, 2021. An independent economic assessment highlighting the distinction between negative reporting and genuinely excessive reporting.
5. Cross-national Evidence of a Negativity Bias in Psychophysiological Reactions to News — Stuart Soroka, Patrick Fournier and Lilach Nir, Proceedings of the National Academy of Sciences, September 2019. Establishes average physiological responses to negative news across 17 countries while documenting substantial individual variation and uncertainty about cultural versus evolutionary origins.
6. The Effect of Fox News on Health Behavior During COVID-19 — Elliott Ash, Sergio Galletta, Dominik Hangartner, Yotam Margalit and Matteo Pinna, Political Analysis, published online July 31, 2023; journal issue 2024. Provides quasi-experimental evidence that specific misleading coverage can change public-health behavior.
7. Data Protection Under GDPR — European Commission, official guidance, checked August 2026. Establishes the actual scope of GDPR obligations for organizations processing European residents' personal information.
8. Forbidden Fruit or Soured Grapes? — Neil Thurman, James Sly, Bartosz Wilczek and Richard Fletcher, International Communication Gazette, July 15, 2022. Documents publishers' decisions to restrict European access and the long-term audience losses that followed.