Picture a post. New study PROVES late-night phone use causes bad grades. Scientists surveyed 40 students at one school; the ones with a phone in bed scored 12 per cent lower. Dr Famous, a dentist, agrees. Every parent knows this. Two million shares. All of it is made up, and every piece is a kind you have met.
Where it came from→What the count hides→The shape of the reason→Move the number
The course is a kit of checks, and this is the whole of it. Where did the claim come from, and what did the mind do with it? What does the count hide? Does the reason reach the claim? And then: set a number, and move it as each check lands.
Four pieces of the post. Tap the one that is a small count.
Tap the part of the picture that answers it.
40 students, one school
++
Phones in bed, lower grades
++
Dr Famous, a dentist
++
Two million shares
40 students, one school
Right. Forty is a sample, and a small one at one school. A 12 per cent gap in 40 people is the kind of swing lesson 5 showed a fair coin producing on its own.
Phones in bed, lower grades
Not yet. That is two things together, lesson 7. It is a real finding about the 40 and says nothing yet about cause, but it is not a count problem.
Dr Famous, a dentist
Not yet. That is about the person, lesson 10. A dentist earned their title on teeth, and the claim still needs its own reason; but it is not a count.
Two million shares
Not yet. That is a count of how far the post travelled, lesson 1 and lesson 10. It is a big number about the spread, not a small count about the students.
Phone in bed
++
Lower grades
The ones with a phone in bed scored 12 per cent lower. Which of lesson 7's four stories fits what the post has shown?
They go together. The phone might cause it, a third thing might push both, or it might be chance in 40 people. Nothing was tested.
Right. Nobody changed one thing on its own, so all four stories are open. The post picked the first story and wrote proves over it.
The phone causes it. That is what the survey found.
Not yet. The survey found two things together in 40 students. Together is the fact; cause is one story of four, and the one the post needed.
Bad grades cause students to take phones to bed.
Not yet. Possible, and as untested as the post's own story. Choosing the backwards story is the same mistake with the arrow turned round.
Each is a piece of the post. Which check does it need?
One card at a time. Tap the pile it belongs to.
Card 1 of 6
Surveyed 40 students at one school.
Dr Famous, a dentist, says it is obvious.
Phone in bed, grades 12 per cent lower.
Two million shares.
Every parent knows this.
The students who filled in the survey.
Surveyed 40 students at one school. → Who got counted
Not yet. Who got counted. Forty students at one school is a small sample through one gate, lessons 5 and 6.
Dr Famous, a dentist, says it is obvious. → Who said it
Not yet. Who said it. A title earned on teeth is about the person; the claim about phones needs its own reason, lesson 10.
Phone in bed, grades 12 per cent lower. → Together or because
Not yet. Together or because. Two things rising together, with four stories open and no test run, lesson 7.
Two million shares. → Who said it
Not yet. Who said it, two million times. A share count measures spread, lesson 10, and says nothing about the claim.
Every parent knows this. → Who said it
Not yet. Who said it. Everyone knows is a count of hearings, lesson 1, and lesson 10's room full of voices.
The students who filled in the survey. → Who got counted
Not yet. Who got counted. The survey was answered by those who chose to answer it, which is a gate, lesson 6.
2Who got counted
++
3Who said it
++
1Together or because
Right. Two pieces are about the count, three are about the people saying it, and one is the finding itself. The three about people were never evidence; the two about the count shrink the evidence; the finding is together, not yet because.
PROVES
vsvs
What would have sunk it?not said
Then the word proves. Lesson 11 asks: what result would have sunk this? A study where phones were taken out of half the bedrooms at random and grades did not move. Has anybody run it? The post does not say. The word proves is doing the work the study did not do.
Move the control through the checks, one at a time. The number is made up and round; the readout says which lesson moved it, and by about how much.
How sure, in 10070 in 100
Where it started70 in 100
Checks applied: Before any check
The post feels right. Phones at night, bad grades: the fast guess says yes, and the number starts where the feeling put it, at 70.
How sure, in 10055 in 100
Where it started70 in 100
Checks applied: 40 students at one school
Forty students is a small count, lesson 5, and one school is one sample. A 12 per cent gap in a group that size swings easily. Down to 55.
How sure, in 10050 in 100
Where it started70 in 100
Checks applied: Only those who answered
The 40 are the students who filled in the survey. Who answers a survey about phones is a gate, lesson 6. Down to 50.
How sure, in 10040 in 100
Where it started70 in 100
Checks applied: Together is not because
Phones in bed and lower grades go together. A busy home or late nights could push both, lesson 7, and nothing was changed on its own. Down to 40.
How sure, in 10040 in 100
Where it started70 in 100
Checks applied: A dentist and two million shares
A dentist's opinion and a share count are about the person and the spread, lesson 10. Neither was evidence, so neither moves the number. Still 40.
How sure, in 10035 in 100
Where it started70 in 100
Checks applied: Nothing that could sink it
The post says proves and names no test that could have failed, lesson 11. Down to 35. Not false; not shown.
Which check moved the number the least?
The dentist and the two million shares. They were never evidence either way.
Right. The number did not move at all on that check, because who said a thing and how many shared it are not clues about the claim. The other checks moved it because they were about the evidence.
The small count. Forty students is still forty.
Not yet. Forty students took the number from 70 to 55, the biggest single move. A small count is a weak count, and the number knows it.
The word proves. One word cannot matter.
Not yet. Proves cost five points, because it claimed a test that was never named. One word is one claim, and a claim with no test behind it is lesson 11's horoscope.
Move the control.
0False
++
about 35The post
++
100Proved
The number ends near 35. That is not false, and it is not proved. Phones in bed may well hurt sleep; it would not be surprising. What the post has shown is a small count, through a gate, at one school, of two things together. Not proved and false are different places on the line.
More coffee shops
++
More new companies
A new post: cities with more coffee shops have more new companies. So coffee makes people start businesses. Which check lands first?
Who said it. Find out who wrote the post.
Not yet. Who wrote it is about the person, and the claim fails on its own shape first. A cause drawn between two things that go together is lesson 7, whoever drew it.
Together is not because. Young, well-off people with money to spend push both.
Right. Two things rising together across cities, with a third thing standing behind both, is the ice cream and the sea. The coffee never reached the company.
The small count. There are not enough cities.
Not yet. There are thousands of cities, so the count is not the weak point. The weak point is the arrow drawn between two things that a third thing pushes.
Lesson complete
A post is taken apart one check at a time, and the number ends somewhere between false and proved.