Loading
You sit down with a specialist. She looks at your scan and says operate. You get a second opinion. He looks at the same scan and says wait. Both of them trained for years. Both are looking at the same body.
You have to choose. You are not qualified to referee them. This module is about that moment. It is not about who is right.
The first instinct is to go and read everything. That instinct is a good one. But you have a weekend. They have had ten years. Reading harder does not close a gap that size.
It feels true because reading does help. You will learn the words and the shape of the argument. It does not give you what took them ten years. That is a feel for which studies are any good. At the end you still choose which expert to trust. You just do it with more words.
You cannot check the chemistry. You can check the shape of the fight. Who is arguing. About which exact claim. With what evidence. And whether anybody keeps score.
This module asks four questions. Is the doubt real, or was it bought? How much of the field is on each side? About what, exactly? Is this lone voice the rare one who turns out to be right? Who has a record you can go and look up? Then it asks what to do when none of that settles it.
It will not make you the judge of the science. It cannot. It will not hand you a side on anything. And it is not medical, legal or money advice. It is about how a disagreement is built.
On a screen, two arguments can look identical. Both have people with titles. Both have charts. Both say the science is not settled yet.
One of them is a real scientific argument. One of them was paid for. They are not the same thing. And the difference shows from outside, with no chemistry at all.
About 66 million years ago, life on Earth crashed. About three quarters of all species died out. A huge rock hit what is now Mexico. Around the same time, volcanoes in India erupted. They poured out lava for a very long time. Scientists agree that both things happened. They argue about how much each one mattered.
This argument has run for over forty years. It is still running today. Both sides publish. Both sides go and collect new rock.
The two camps agree on almost everything. They agree the crater is there. They agree the lava is there. They agree roughly when the rock layer formed.
The fight is about weight. How much of the dying goes to each cause? Nobody is selling anything. New dates keep arriving, and both sides answer them. The papers argue with each other by name. Other geologists can check the work.
Two teams have argued for forty years about what killed the dinosaurs. How much of the story do they agree on?
In 1964 a US Surgeon General's committee reported on smoking. It had read more than 7,000 published articles. It said cigarettes cause lung cancer. Inside the cigarette companies, people already knew. Five years later, someone inside one company wrote down the plan.
“Doubt is our product since it is the best means of competing with the 'body of fact' that exists in the mind of the general public.”— Brown & Williamson internal memo, "Smoking and Health Proposal", 1969. Truth Tobacco Industry Documents, UCSF, document xqkd0134.
The plan worked for a long time. In 2006 a US federal judge ruled on it. She found a fifty-year scheme to deceive the public about smoking. The doubt was not a scientific position. It was a product with a budget.
None of it made the science less settled. It only made the public argument louder.
Researchers have read the internal papers and named the moves. Fund your own studies. Attack the studies you did not fund. Start groups with science-sounding names that will not say who pays them. Feed the press. Lobby. Sue.
A live disagreement lives in journals, at conferences and in data. Bought doubt is built to reach you, not the field. Adverts, opinion pages, hearings. Ask where the doubt is aimed. That one question sorts a surprising amount.
Real disagreement moves. A better measurement arrives and somebody shifts. Bought doubt repeats. The same objection comes back for thirty years, whatever turns up. A position that no evidence can change was never about evidence.
Which kind of doubt is each one?
Tap an item, then a bucket
These signs move the odds. They do not close the case. A researcher paid by a company can still be right. A man who repeats himself can still be right. And a real disagreement can be picked up and used by someone selling something.
Sorting tells you how much weight the doubt has earned. It is not a way to stop listening.
"Scientific consensus" sounds like a show of hands. It is not. Nobody sends out a ballot. It is what survives years of people trying to break a claim. The 1964 smoking committee is a clear case. Its answer came from a whole pile of work, not one study.
Agreement stays invisible until somebody measures it. Measuring it does not create it. There are four common ways. Ask the researchers. Read every paper. Put a committee in a room. Or look at what the textbooks teach.
Each one gives a real answer. Each one has a hole in it.
This one is a survey. The whole game is who you ask. If you pick who answers, you have already picked the answer.
A standing panel closes some of that door. One economics panel in Chicago keeps a standing list of senior economists. It puts each question to all of them. It publishes every reply, including the ones who answer "uncertain".
A systematic review is the opposite of picking the studies you like. One team writes the rules before it starts. Then it hunts down everything published on the question. It reports what the whole pile says.
A review usually ends with a label on how sure it is. One system sorts evidence into four levels. High, moderate, low and very low. More than 120 organisations in over 20 countries now use it.
So "low quality evidence" is often a real label with rules. It does not mean the claim is false. It means the reviewers are not sure yet.
The US National Academies pick a committee for a question. They check the members for conflicts of interest. They gather evidence, some of it in public. Then outside experts review the draft before anyone may publish it.
The committee has to reach agreement before the report goes out. That is what the word "consensus" means on the cover of those reports.
The last way is the slowest. Look at what the textbooks now teach. A claim in the textbooks has already won its argument. The catch is timing. Textbooks are written years after the fight.
Every way of measuring agreement has a weak spot. Match them up.
Tap one side, then its match
A consensus is rarely one solid block. The centre can be settled while the edges stay wide open. That cigarettes cause lung cancer is settled. The harm from a few cigarettes a month is much less clear.
Both of those sentences are about smoking. Only one of them is settled. So both of these can be honestly said about smoking. The science is settled. Scientists disagree. They are pointing at different parts of it.
"Scientists agree" is a headline, not a claim. The useful move is to write the claim out in full. Often the honest version arrives with a range attached, not one number.
A range is not weakness. It is the shape of the actual finding.
A news story says scientists are divided. What is the most useful next question?
In the early 1980s two doctors in Australia said something odd. They said stomach ulcers were caused by bacteria. Almost everyone believed ulcers came from stress and acid. Almost everyone believed nothing could live in stomach acid.
In 2005 the two of them won the Nobel Prize in medicine.
Barry Marshall could not get people to take it seriously. So he drank a glass of the bacteria himself. Within days he was sick. He showed the bacteria had inflamed his stomach. Then he published it, where anyone could check the work.
He did not ask to be believed. He produced a result the old story could not explain.
In 1912 Alfred Wegener said the continents move. He was largely right. His critics had a fair objection. He could not say what moved them. That is not stubbornness. That is a real hole in an idea.
The hole was filled in the 1960s by mapping the sea floor. Then the field changed its mind quickly. The idea won when the missing piece arrived.
It feels true because it has happened. They laughed at Marshall. They laughed at Wegener. The trouble is that almost every new idea gets laughed at. Nearly every wrong idea gets rejected too. So rejection, on its own, tells you almost nothing.
“The fact that some geniuses were laughed at does not imply that all who are laughed at are geniuses. They laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown.”— Carl Sagan, Broca's Brain, 1979, chapter 5, p. 75
Linus Pauling won two Nobel Prizes. Late in life he said big doses of vitamin C stopped colds. Trials have now tested that in more than eleven thousand people. Taking it every day did not stop people catching colds. It shortened them a little, and that was all. One group was different. Marathon runners and soldiers did catch fewer colds.
A prize in chemistry is not a licence in medicine. Expertise is narrow, even at the very top.
The right ones and the wrong ones both stood alone. Standing alone is not the signal. What separates them is what they did next.
A claim that could have failed, and did not, has earned something. A claim that fits every possible outcome has earned nothing.
Marshall's claim could have failed loudly. If stress really caused ulcers, killing the bacteria would change nothing. People were given antibiotics. The ulcers mostly stopped coming back.
You already have the word for a claim that cannot fail. It is unfalsifiable. That kind of claim never earns anything, however confident the voice is.
Most lone dissenters turn out to be wrong. Not all of them. Most. That is exactly why the ones who were right are famous. Famous is what rare looks like. A story you hear often can still describe something rare.
So the honest starting position is: probably not. The useful follow-up is: and here is what would change that.
A doctor with real qualifications says a standard treatment is useless. He has a website, a book and a lot of confidence.
You have twenty minutes and no medical training. Where do you start?
What do you look at first?
An opinion cannot be wrong later. It just gets replaced by the next opinion. A forecast can be wrong. It carries a date and a number. Then it happens, or it does not. That is why forecasts can be scored and opinions cannot.
Weather services score themselves. Every rain forecast gets checked against what happened. So you can ask a real question. On the days they said 70%, how often did it actually rain? People have gone and counted. The answer comes out close to 70.
A psychologist spent about twenty years collecting predictions. They came from hundreds of people paid to comment on world events. He gathered tens of thousands of dated forecasts, then scored every one.
The average was poor. It was not much better than spreading your bets evenly. Simple rules that just followed recent trends beat the people.
Among those experts, who scored better?
A later tournament ran for four years. Thousands of people took part. Some were much better than the rest, year after year. Putting the best of them on teams together raised accuracy further. A short course on handling probability raised it too.
The point is not that these people are wizards. The point is that a score existed at all. Without one, nothing can be seen to improve.
Which of these claims could somebody check later?
Tap an item, then a bucket
Most public experts have no score. Nobody wrote the predictions down, so nobody could check them. That is not a crime. It is a missing measurement, and it is worth noticing.
A fair question for anybody. What did you say last time? What happened? The answer tells you more than the confidence in their voice.
When two experts disagree, doing nothing feels safe and neutral. It is neither. Doing nothing means carrying on as you are. That has a cost too. The cost is just quieter than the cost of acting.
Two experts who disagree usually agree about most of it. Write down the part they both accept. It is often longer than the part they fight over. Sometimes your decision is already sitting inside the agreed part.
Two specialists disagree. What is each disagreement actually about?
Tap an item, then a bucket
Two panels read the same breast screening studies. One says a scan every two years, from 40 to 74. The other says a scan every year from 45. It offers yearly scans from 40 as an option.
They are not arguing about the numbers. They weigh missed cancers against false alarms differently. No new study settles that. The question contains a trade somebody must make.
Ask each side one thing. What result would change your mind? If both can answer, the disagreement is about evidence. Evidence can move it. If neither can, it is about values. No study is coming to help.
It is a polite question. It is also the most informative one available to you.
Sometimes one option is fine under both stories. Take that one, and the disagreement stops mattering to you.
When no option is safe under both, ask a different question. Which mistake can I undo? You already have that idea from decision-making. Some doors only open one way.
Two doctors trained by the same teacher are nearly one doctor. Ask each of them where they learned it. Ask what they read. Agreement is only worth something if it was reached separately.
Note what you decided. Note how sure you were, as a number. Note what you expect to see, and by when. Then go back and look at it later.
That is how a record gets made. You have spent this whole module asking other people for one. This is the version you can build yourself.
Is the doubt real, or was it bought? They agree that what, exactly? Is this lone voice doing the things the right ones did? Who has a score you can look up? And when it stays unresolved. What do we both agree on? Which mistake can I undo?
Two specialists disagree. Neither can say what would change their mind. What does that most likely mean?