Science
97% of scientists agree — so why does the public hear a 50–50 shouting match?
On a great many questions — that the climate is warming, that vaccines do not cause autism — something close to 97% of the relevant experts have landed on the same answer. Yet the version that reaches you, on your screen and at dinner, often sounds like a dead-even brawl with two equally loud sides. Somewhere between the lab and your living room, a settled question got re-staged as a fight. The interesting thing is that this is sometimes done on purpose, and that you can learn to spot the wires.
Start with a confusion that does most of the damage: people hear the word "consensus" and picture a vote, a show of hands, a popularity contest among experts. It is nothing of the sort. A scientific consensus is not an opinion, a preference, or a belief. It is the weight of the evidence — what is left standing after a claim has been challenged, re-checked, re-tested, and torn at by rival hypotheses, and still keeps giving the same answer. That is why overturning an established consensus takes more than a clever objection or a loud voice: it takes evidence strong enough to survive the same gauntlet.
Consensus is a survivor, not a vote
Here is the translation that makes it click. A consensus is not the crowd that gathered around an idea; it is the idea still standing after everyone took their best swing at it. Picture a bridge that has been load-tested by every engineer who hated the design, in every storm, for decades — and has not fallen. You do not cross it because a committee likes it. You cross it because it has refused to break. That refusal, accumulated over time, is what scientists mean by consensus. Carl Sagan called science "a candle in the dark" for exactly this reason: it is not a set of conclusions you must accept, but a method for keeping the few claims that survive relentless testing and blowing out the rest.
And the method has teeth. Sagan's "baloney detection kit" is a short list of moves any honest inquiry should welcome: insist that facts be independently verified; encourage open debate of the evidence; remember that an appeal to authority carries little weight, since in science there are no authorities, at most experts; spin up multiple competing hypotheses instead of marrying the first appealing one; quantify where you can; and prize falsifiability — always ask what evidence could prove the claim wrong. A real consensus is what is left after a claim runs that whole gauntlet and survives. Which gives us a sharp test for the impostor: a manufactured controversy is one that refuses to run the gauntlet, yet insists the verdict is still open.
Five wires behind the fake fight
Once you know what a real consensus is, the counterfeit is easy to back-light. Science communicators have catalogued the recurring tricks used to make a settled question look like it is still being debated. There are five worth naming. The disguised expert: a confident voice with a title that sounds relevant but lies far from the actual field — a physician opining on climate physics, say. Cherry-picking: parading the handful of studies or data points that fit, while the mountain that does not is quietly left offstage. The Galileo gambit: "they laughed at Galileo too" — as if being mocked were itself a credential, when being ridiculed has never once made a claim true. Doubt-mongering: not arguing a counter-case at all, just spraying vague uncertainty so you feel you should "wait until the science is in." And false balance: staging the side with almost no scientific support as the equal-and-opposite of the side with overwhelming support, two chairs of identical size on the same stage. Sagan would recognize every one of these as an old fallacy in a fresh suit — selective observation, appeal to authority, ad hominem dressed up as humility.
The good news: rebuttals work
If that sounds like a losing battle, the data say otherwise. In 2007 a scientist named John Cook started a project, Skeptical Science, to answer climate myths one by one; it now holds more than 250 rebuttals, each written at three levels — basic, intermediate, advanced. (It is also home to the often-cited paper finding 97% expert agreement on human-caused warming, Cook and colleagues, 2013.) Research on those rebuttals found something heartening: a good debunking measurably lowers belief in the myth — and it helps the people who were most wrong to begin with the most of all. The trick is not to shout louder. A strong rebuttal calmly supplies the "alternative fact" — the clear, correct explanation — and then names the specific fallacy the myth was using to fool you.
The core line
A real consensus is what survives the gauntlet; a fake controversy is a fight that performs doubt but never runs the test. Ask which one you're watching.
But science does overturn consensus — so where's the line?
Now the objection that has to be answered, or this piece just trades one credulity for another. Doesn't science sometimes overthrow a reigning consensus — continental drift, ulcers caused by bacteria, the expanding universe? Absolutely, and that is the system working, not failing. But notice how those reversals won. Not by a louder voice or a wounded "they laughed at me too," but by evidence strong enough to drag the experts, often against their will, to a new answer. That is the razor's edge Sagan asked us to walk: stay genuinely open to the new idea, and stay rigorously skeptical of it, at the same time. The fake-controversy machine wants you to fall off one side — "the experts are all bought, trust no one." The lazy opposite wants you to fall off the other — "the experts spoke, stop thinking." Neither is skepticism. Skepticism is asking, every time, what would change my mind, and whether the loud new claim has even tried to clear that bar.
What it means for you
You will not personally re-run the climate models or the vaccine trials, and you do not have to. What you can do is carry one cheap, fast test into the next argument that lands in your feed. When two sides are presented as evenly matched, stop and ask a single question: is this side overturning the evidence, or is it performing the fact that a fight still exists? Look for the wires. Is the "expert" actually in the field? Are they showing you all the data or a hand-picked few? Is "they mocked Galileo" doing the work that evidence should be doing? Is anyone selling you a fog of doubt without a single testable claim inside it? You do not need a PhD to run that checklist — Sagan built the kit precisely so that ordinary citizens could. So the next time a settled question is wheeled out as a 50–50 cage match, you can ask the only question that sorts the real fight from the show: is anyone here actually trying to break the bridge — or just renting two chairs and pretending the crossing is still in doubt?
Framework from Carl Sagan, The Demon-Haunted World (the "baloney detection kit" — independent verification, debate, no special weight for authority, multiple hypotheses, quantification, Occam's razor, falsifiability), via the The Demon-Haunted World knowledge skill, with sticky-writing self-checks from Made to Stick. The five misleading tactics draw on ScienceUpFirst's public explainer on scientific consensus (2026); the rebuttal findings and the 250+ rebuttals / three-level structure / Cook et al. 2013 "97%" paper draw on public reporting about Skeptical Science (2026). Popular-science commentary; check specific studies against the primary sources.