Picture a room with eight people in it. Everyone is shown a line, and three more lines, and asked which of the three is the same length. It is obvious. Seven people answer, one after another, and all seven pick the wrong line. Then it is your turn.
Went with the wrong room37 in 100 trials
Said what they saw63 in 100 trials
Asch, 1955 report: 123 people.
That experiment was first run in 1951 and reported in full in 1955, on 123 people. The seven were actors. The eighth person, the only real one, went along with the wrong answer in about 37 of 100 trials, and three people in four went along at least once. The lines had not changed. The room had.
Slide the marker to the share of trials, with a wrong majority in the room, where the real person went along with it.
Slide the marker to where you think the answer is, then press Lock. Nothing is scored.
123
your guess · 0%
the answer · 37%
0%50%100%
1Rare10%
2Half50%
3Most90%
Your guess: —
About 37 in 100. Guesses land low, because the line was obvious and nobody thinks they would pick the wrong one. Alone, people got it wrong less than one time in a hundred. The room, not the line, produced the other 36.
1 voice4 in 100
2 voices14 in 100
3 voices32 in 100
4 or more35 in 100
Asch, 1955. Share of trials that went with the wrong group.
How many voices does it take? In a later run the experiment changed the size of the group. One wrong voice moved almost nobody: about 4 in 100. Two voices, about 14. Three voices, about 32. After that, adding more voices added almost nothing.
Move the control through the number of people giving the wrong answer before your turn. The figures are from the 1955 run, rounded.
Went along4 in 100
Said what they saw96 in 100
People giving the wrong answer first: 1
One person gives the wrong answer before you. People went along about 4 times in 100. One voice barely moves anybody.
Went along14 in 100
Said what they saw86 in 100
People giving the wrong answer first: 2
Two people give the wrong answer. People went along about 14 times in 100.
Went along32 in 100
Said what they saw68 in 100
People giving the wrong answer first: 3
Three people give the wrong answer. People went along about 32 times in 100. The third voice is the one that does the work.
Went along35 in 100
Said what they saw65 in 100
People giving the wrong answer first: 4 or more
Four, seven, nine people give the wrong answer. People went along about 35 times in 100. With fifteen it was about 31. Past three, more voices add nothing.
Going from three wrong voices to fifteen changes the share who go along by about how much?
It falls. A huge group looks like a trick.
Not yet. The share did not fall; it flattened. The flat line is the finding, and a mind that spotted the trick at fifteen would also have spotted it at three.
Hardly at all. Past three, more voices add almost nothing.
Right. Three voices got the share to about 32; four got it to about 35, and fifteen to about 31. A mind treats three as a crowd, and a crowd is a crowd however big it gets.
A lot. Fifteen people is far more pressure than three.
Not yet. That is what everybody expects and what the 1955 run did not find. The jump is between two voices and three; after that the curve goes flat.
Move the control.
Ten voiceshow far it spread
vsvs
The reasonhow true it is
The mind takes how many people say a thing as a reason. It is not one. Ten people repeating what they heard is one claim, said ten times, which was lesson 1. The count of voices tells you how far the claim has travelled. The claim is as true or false as it was with one voice.
40,000 shares
↓→
Measureswhat?
A post has 40,000 shares. By this lesson, what does the share count measure?
How far it spread. Nothing about whether it is true.
Right. Forty thousand people passed it on, and passing on is hearing, not checking. The count is a fact about the spread and says nothing about the reason behind the claim.
How true it is. Wrong posts get corrected before they reach 40,000.
Not yet. Wrong posts travel as well as right ones, and often better, because they are built to be shared. Nothing in the count checks the claim.
How many people checked it. A share is a vote of confidence.
Not yet. A share is a repeat, and the Asch room showed that people repeat a wrong answer they can see is wrong. A vote from a room is not a look at the line.
Each is something said about a claim. Is it about the reason behind the claim, or about the person who made it?
One card at a time. Tap the pile it belongs to.
Card 1 of 6
She is only fifteen. What does she know?
He is a doctor, so his tip about car engines must be right.
The study counted two million people over ten years.
Everyone in my class believes it.
The claim matches what we measured ourselves last year.
A billionaire said it on a podcast.
She is only fifteen. What does she know? → About the person
Not yet. About the person. Her age says nothing about whether her reason holds; a fifteen-year-old can read a chart.
He is a doctor, so his tip about car engines must be right. → About the person
Not yet. About the person. A doctor earned their title on bodies, and an engine is not a body; the tip still needs its own reason.
The study counted two million people over ten years. → About the reason
Not yet. About the reason. The size of the count is part of how strong the evidence is, lesson 5, whoever is reporting it.
Everyone in my class believes it. → About the person
Not yet. About the person, thirty times over. A class agreeing is the Asch room; it measures the spread, not the claim.
The claim matches what we measured ourselves last year. → About the reason
Not yet. About the reason. A match with a separate measurement is a check, lesson 1, and a check is evidence.
A billionaire said it on a podcast. → About the person
Not yet. About the person. Wealth is a fact about the speaker's money, and a microphone is not a reason.
2About the reason
vsvs
4About the person
Right. Four of the six never touch the claim. They are about who said it, how old they are, how rich, or how many agree, and a claim is not made true or false by any of those.
A surgeon on heartschecked for 20 years
vsvs
A surgeon on enginesnever checked
Expertise is real, and it is narrow. A heart surgeon knows hearts, because for twenty years the world has checked their answers. On car engines they are a stranger with a confident voice. The question about an expert is: did they earn it on this kind of question, and would they be caught if they were wrong?
A footballer50 million followers
↓→
The creamcures acne?
A footballer with 50 million followers says a cream cures acne. By this lesson, what have you learned about the cream?
Nothing yet. The follower count and the fame are about the person; the claim still needs its own reason.
Right. Fifty million followers measures how far the claim will spread, and fame is earned on football. The cream needs a test, lesson 7, and the footballer has not given one.
It probably works. Fifty million people cannot all be fooled.
Not yet. Fifty million people have not been asked; they are the audience, not a check. The Asch room fooled people with seven voices, and no voice there knew anything either.
It does not work. Famous people only say what they are paid to.
Not yet. Being paid is about the person too. The cream might work or might not, and the footballer's bank account cannot tell you which.
Lesson complete
How many say a thing, and who, measures how far it spread. The claim is as good as its reason.