Ed Yong
About
Science writer, author of 'I Contain Multitudes'
Cast within
No topic-region cast yet — this appears once Ed Yong's compiled claims are aligned into a topic region's argument tree.
Claims by Ed Yong (20 of 42)
Happy-Clappy Science Communication Is Dangerous
One of the most dangerous ideas in science communication is that the only way to promote science to people is by presenting it as uniformly exciting and happy—'sitting around the campfire singing Kumbaya'—when in fact showing where scientists disagree, where studies are flawed, and what remains to be done is an equally valid and genuinely fascinating way to engage people.
Discerning Readers Are the Key Lever
The most powerful force that could improve science journalism is more discerning readers: if people paid attention to and rewarded material that went the extra mile to provide context and verification, and ignored content that merely rehashes press releases, the incentives would shift—otherwise audiences get the low-quality media they deserve, and scientists and science communicators are as guilty as the public because of their tendency to want science presented as uniformly exciting and fun.
Failure Mechanisms Reinforce Each Other Synergistically
The individual problems in science—over-reliance on conceptual replication, researcher degrees of freedom, publication bias, and the difficulty of interpreting replication failures—are each problematic on their own, but synergistically they work together to enhance each other's failings rather than adding up linearly.
Conceptual Replication Masks Weak Findings
Psychology relies heavily on 'conceptual replication'—taking many different experiments that test the same underlying idea and treating them as supporting each other—but this only validates a concept if each individual study is itself solid; if the underlying studies were produced through researcher degrees of freedom or selected by publication bias, then conceptual replication becomes the epitome of confirmation bias rather than genuine corroboration.
Researcher Degrees of Freedom Manufacture False Positives
Researchers commonly make on-the-fly analytical decisions—checking significance levels during data collection and stopping when significant, not pre-specifying sample size or statistical tests, and discarding outliers—that virtually guarantee positive results; Joe Simmons demonstrated these 'researcher degrees of freedom' could 'show' that listening to the Beatles' 'When I'm Sixty-Four' physically made people younger.
Bargh Priming Study Failed Direct Replication
John Bargh's highly cited 1996 priming study—showing that people unconsciously exposed to old-age-related words walked more slowly down a corridor—failed to replicate in a careful 2012 attempt by Axel Cleeremans's Belgian group, which found the slowing effect appeared only when the experimenters administering the test had been told to expect it.
Experimental Skill Claim Is a Dangerous Excuse
While there is genuine experimental skill and artistry in pulling off subtle psychology experiments (Kahneman argues priming requires a 'knack' to avoid making subjects aware they've been primed), the existence of this idea is dangerous because it provides an easy excuse for researchers whose results cannot be replicated to attribute failures to others' lack of skill rather than to the falsity of the original finding.
97 Percent of Psychology Papers Report Positive Results
Publication bias toward positive findings is long-standing and severe: statistician Theodore Sterling found in 1959 that 97% of studies in four major psychology journals reported statistically significant positive results, and when he repeated the analysis decades later in the 1990s the figure was virtually unchanged.
Journalism Treats Each Study as Isolated Point
A central failing of science journalism, especially for epidemiology, is treating every new individual discovery as an isolated point rather than placing it within the accumulated body of evidence; epidemiology relies on strength in numbers, and the alcohol-cancer link is actually strong because the IARC graded it as a cause about 20 years ago and dozens to hundreds of studies plus plausible mechanisms have since confirmed it—context that journalists routinely omit.
Psychology's Openness Is a Strength Not Weakness
Psychology stands out positively not because it has fewer problems but because the people openly discussing and criticizing the field's replication failures are themselves psychologists—insiders taking an introspective look—rather than outside critics, which exemplifies the self-correcting character that science is supposed to have.
Mechanism Plus Correlation Strengthens Causal Claims
A causal claim in epidemiology is much stronger when, beyond observing a correlation, there is a plausible biological mechanism; the alcohol-cancer link is supported both by dozens of observational studies finding the same effect and by studies showing it is biologically plausible that alcohol causes cancer, making the link stronger than most other risk factors even though weaker than smoking.
British Culture Argues Against Anti-Alcohol Bias
Russ's hypothesis that anti-alcohol research is 'primed' by cultural moralizing against drinking is undercut by the fact that much of the strongest alcohol-cancer evidence comes from Britain, whose drinking culture is not compatible with a temperance-driven bias.
Verification Is Hard in Interdisciplinary Science
Verifying whether a scientific result is reliable is increasingly difficult because science is becoming highly interdisciplinary and networked, so a journalist can no longer just hand a paper to a couple of experts—those experts may know only one side of the field, may lack the statistical skills, or, worst case, an entire field may be in collusion with a received line of thinking (such as priming researchers all buying into the same concept).
Internet Spread of False Facts and Counter-Network
The internet both spreads false 'known facts' (like the claim that NFL players' life expectancy is 58, which traces to the biased NFL Players Association and is contradicted by evidence that players are healthier than average) and simultaneously plugs journalists into a previously invisible network of scientific criticism that makes them better at spotting methodological flaws—but the open question is whether corrections ever travel as far as the original mistakes, which they generally do not.
Replication Is the Under-Rewarded Backbone of Science
Replication should be the backbone of science—the way of checking that what enters the public literature is actually true—but in practice failed replication attempts are rarely done, and when they are, they circulate only as gossip and never see the light of day, forming a hidden undercurrent that the formal literature does not capture.
Taste vs Smell Distinguished by Use
The true distinction between taste and smell is not the organ or the distance of the stimulus but their use: taste is a largely innate, inflexible, near-binary sense used almost entirely to decide if something is good to eat (spit or swallow), while smell is complex, experience-dependent, tailorable, and used for a wide range of purposes from navigation to social interaction.
Smell Perception Is Top-Down and Context-Dependent
Smell is partly top-down rather than bottom-up: the same odor can be repulsive or attractive depending on context (e.g., cheese versus something awful), and perception depends on experience and the cultural value imparted on things, both within humans and dramatically across species where the same smell carries different meanings.
My Notes
Loading notes...