One New Year, someone forwarded a line into our family chat.
“Studies show that one glass of red wine a day is good for your heart.”
I did not reply.
Not because I doubted it.
Because I did not know how to answer.
Saying “that sounds right” left me unsure.
Saying “that sounds wrong” left me without a reason.
Later I understood.
The problem was not the message.
The problem was that I had no way to look at it.
This book is that way of looking.
It became my default move for lines like this.
1. The book dealer does not read the text
A dealer in old books opens a book and checks three things.
Whether the paper belongs to that decade.
Whether the type matches the printer.
Whether anyone wrote real notes in the margins.
He never starts with the text.
Because he knows content can be faked, and the traces of how a book was made cannot.
A line that says “studies show” has its own paper and type.
Where it came from, who made it, whether anyone took it apart.
Those traces are more reliable than the conclusion itself.
What he teaches is exactly this way of looking.
The author says it plainly in the preface.
This is not a standard introduction to psychology, and it does not summarise what the field has found.
It does not tell you what psychology discovered.
It tells you how a sentence was made.
That is why the title says “see through”.
Not because you should doubt everything.
Because before you believe it, you should look once more at where it came from.
Someone who cannot read the traces buys more and gets it wrong more often.
〔图 1：一句“研究表明”，过几道手〕
2. The master who explains everything
A patient walks in and says he cannot sleep.
A doctor asks three questions: when do you go to bed, what do you do before, do you nap during the day.
After asking, the doctor may still not know.
Change to a “master”, and he asks one question: how was your relationship with your mother as a child?
Then everything can be explained.
Insomnia comes from anxiety, anxiety comes from the mother, the mother comes from her time.
Explaining everything means explaining nothing.
The book calls this the Freud problem.
The numbers are cold.
Fewer than one in ten members of the American Psychological Association identify with psychoanalysis.
Five Nobel laureates the public has barely heard of are the real face of modern psychology.
They are Kahneman, Simon, Hubel, Wiesel and Sperry.
What the public remembers is the man who explained everything.
Then there is common sense.
“Look before you leap” and “strike while the iron is hot” are both proverbs.
“Many hands make light work” and “too many cooks spoil the broth” are both proverbs.
Two proverbs contradict each other. Which one do you trust?
Neither.
A proverb’s talent is explaining afterwards.
Before the fact, it told you nothing.
The author’s line is harsh: people love these sayings because they are hard to refute.
Whoever can explain everything has usually said nothing.
3. The forensic examiner’s line
A forensic examiner works with one rule.
To call a death a homicide, he must say what would make him change his mind to “accident”.
If he cannot say it, the judgement does not stand.
An examiner who always writes “natural causes” is no examiner at all.
The book’s counterexample is Freud.
The author says he “made his theory impossible to falsify”.
Whatever the patient says, it fits mother, repression, or the unconscious.
Now a positive example.
“Gay people are more likely to be open about themselves in more tolerant societies.”
This sentence can be slapped down by data.
If tolerance rises and openness stays flat, it is wrong.
Only a claim that can be slapped down in public deserves a hearing.
There is a trap here.
Some people turn “falsifiable” into “I do not believe it, so it is not science”.
That is turning a rule of inspection into an excuse for refusing delivery.
Falsifiable means this: in principle, an observation could prove it wrong.
It does not mean: I refuse.
〔图 2：能解释一切，和能被推翻〕
4. The blender’s “a pinch” and the lab report’s degrees
A master blender says “a pinch”.
The apprentice asks: how much is a pinch?
The master says: use your hand.
Inside the plant, nobody dares argue.
But once you need to export, or go to court, or repeat it, the hand is not enough.
The lab report says: fifty two degrees.
From then on, the argument becomes a test.
The book says scientists are not essentialists.
They do not ask what intelligence really is.
They ask: which task, which score, which condition measures it.
Once it lands on a task, the argument has a referee.
“Whether this measure reflects intelligence” can still be argued.
But at least people are arguing about the same thing.
There is a cost here too.
Define intelligence as one test score, and you lose creativity, street sense, and judgement under pressure.
The book says operational definitions are temporary and revisable.
It is not the truth, it is a temporary lab report.
Lab reports get new editions, and that is not frightening.
What is frightening is mistaking one edition for the disease itself.
5. The claims investigator asks who knows whom
A claims investigator has a habit.
Three people all say they saw the car hit.
He does not ask what they saw first. He asks: do you three know each other?
If they do, the testimony gets discounted.
This is not cruelty, it is experience.
The book starts with a television show.
For ten years the most famous “Amazing Randi” made a whole audience believe they had been healed.
The placebo effect tells us one thing.
If people believe they will get better, their symptoms can really ease.
So “he took this and got better” has at least three possibilities.
The thing worked, the placebo worked, or the body healed itself.
One person cannot tell these three apart.
Then there is vividness.
A living story is easier to remember than a table of numbers. It is also easier to believe.
That is a feature of the human brain, not a defect.
But using it as evidence turns “easy to remember” into “true”.
In court, eyewitness testimony is vivid and fingerprints are dull.
Cases are decided by fingerprints, not tears.
Still, a single case is not garbage.
A case is the best source of hypotheses.
“This person ate a mushroom and got better, so maybe the mushroom contains something.”
That is where research starts.
It is just not where it ends.
A personal story is bright. Bright enough to cast your own shadow.
6. Ice cream, drowning, and the pump
In summer, ice cream sells more, and more people drown.
Someone concludes at once that ice cream is dangerous.
Anyone with a little statistics asks one more question: what about heat?
Heat sells ice cream, and heat sends people into the water.
Heat is the third variable.
The book lists three traps: the third variable, direction, and selection bias.
Direction asks: does A cause B, or does B cause A?
Does anxiety cause insomnia, or does insomnia cause anxiety?
Both directions work.
Selection bias asks: did the sample choose itself?
Women who used hormone replacement therapy were healthier.
Maybe because healthier women were the ones who went for it.
The book’s “ovens prevent pregnancy” example is savage.
Households with ovens had fewer children.
Because households with ovens were more likely to have two working adults and less time.
The oven has no contraceptive power.
Two things moving together.
It often only means they heard the same footsteps.
〔图 3：两件事一起动，先别急着下结论〕
7. The horse that could count, and the pump on the map
In the early twentieth century, a horse could do arithmetic.
A question was asked, and the horse tapped the answer with its hoof.
The crowd marvelled.
Then someone ran an experiment: the person asking stood where the horse could not see him.
The horse could not answer.
It was not calculating, it was reading faces. When a person tensed, it stopped.
Then move the scene to London, eighteen fifty four. Cholera was everywhere, and Snow marked each death on a map.
The dense dots gathered around one pump. He had no microscope.
He had comparison: the water from that pump, and water from elsewhere.
Finally, the coin toss.
Assign participants to two groups at random.
The groups end up alike on every variable, known and unknown.
This is the hardest way to remove a third variable.
You want to know which fertiliser is better on two plots.
The fairest method is not “pick two plots that look alike”. It is a coin toss.
Because the sameness you can see is not the sameness you cannot see.
If you want fairness, hand the choice to a coin.
8. The test track and the rush hour
You want to test a car’s brakes. Do you brake on a test track, or in rush hour traffic?
The test track is not real life.
But it can repeat, compare, and remove the road.
Rush hour is more real.
But you cannot tell whether the car stopped because the brakes are good, or because the road was empty that day.
The book says naturalness is not always necessary.
Psychology experiments simplify on purpose.
A button, a paragraph, a staged scene.
Simplification is not laziness.
Simplification pulls the variable out.
The cost is that lab conclusions have to be filled in by fieldwork.
The lab says the intervention works; the field may disagree.
To see a thing clearly, you sometimes have to make it unnatural first.
9. A crate, three items
An inspector checks a crate by pulling three items.
All three pass, and still he will not write “all good”.
He writes: this batch passed sampling.
This is not word games.
He knows three good items are not one good crate.
The book’s formulation: suppose results from many different experiments all support one particular conclusion.
Then it adds a finer judgement.
If all experiments fail in the same way, that is when to worry.
If experiments with different flaws all point the same way, confidence rises.
The second half is paraphrase, not quotation.
There is a quieter trap here.
If only significant results get published, the pile you gather is “significant”, not “true”.
That is publication bias.
So aggregation must look at all the evidence, not at the evidence you can find.
Evidence is not one clean strike.
It is many hands pulling the same rope from different directions.
People who wait for the “decisive experiment” have a name in this book.
Einstein syndrome.
A single result that explains everything at once never arrives.
10. Heart disease, and two pills that only hurt together
Heart disease has more causes than one hand can count.
Genes, diet, stress, sleep, income.
Change only the diet and leave the rest.
The effect will be limited.
The book calls this the myth of the magic bullet.
There is a harder layer. Interaction.
Two drugs are fine alone and dangerous together.
Two pressures are bearable alone and crushing when stacked.
Causes amplify each other, and they cancel each other.
So “find that cause” is often an illusion.
What you find is usually the cause you thought of first.
11. A false positive on the lab report
A disease affects one in a thousand people. The test is ninety nine percent accurate.
You test positive. What is the chance you actually have it?
Most people say ninety nine percent.
About one in ten.
Because healthy people are so many, a one percent false alarm rate is enough to swamp the sick.
This is the base rate problem.
The gambling table holds the same illusion.
Ten reds in a row, and the players who bet black feel more confident.
The wheel has no memory. The gambler does.
Now another line.
“My uncle smoked for forty years and lived to ninety.”
That is a real case. But it does not defeat “smoking raises the risk of lung cancer”.
Because the latter is about probability, not about any one person.
The forecast says seventy percent chance of rain, you bring an umbrella and stay dry. That does not mean the forecast was wrong.
It said seventy.
Probability does not tell stories. So people do not remember it.
12. After five heads
A coin lands heads five times in a row. You feel the next one should be tails.
The coin has no memory. Your brain does. It turns five heads into the start of a story.
The same on a court. A player makes three shots and the team feels he is “hot”.
The data says his next shot is no more likely to go in. The team still passes to him.
The book is clear about this. People tend to explain chance events.
The tendency has a use. It lets us learn fast in a chaotic world.
It also has a cost.
It lets us see cause where there is only randomness.
Do not swing to the other extreme either.
Treating every pattern as chance is also a kind of laziness.
The difference is this: can the pattern be predicted, and can it be repeated.
〔图 4：听完“研究表明”，先停三秒〕
At the end, four questions
The book teaches twelve tools. In daily life, they shrink to four questions.
First question: under what condition would this claim be wrong?
If you cannot say, it is a sentence that can never be wrong.
A sentence that can never be wrong has said nothing.
Fortune tellers understand this best, which is why they say “you are about to face a change”.
Second question: how was that word measured?
Anxiety, intelligence, addiction, burnout, these words all sound familiar.
But ask: what measures it?
If nothing does, everyone is just talking past everyone else.
Third question: what does it offer as evidence?
One person’s story, or a comparison group?
“My friend took it and got better” is a story.
“Two groups, one takes it, one does not, and here is the difference” is a comparison.
Stories pick their people. Comparisons do not.
Fourth question: when two things move together, is a third thing pushing both?
Ice cream and drowning.
The weather pushes both, not the ice cream.
After four questions, you will find most “studies show” lines cannot survive them.
Here are a few phrases worth pausing on.
“Studies show”: no author, no journal, no sample size.
“Experts say”: no name, no field, no conflict of interest.
“Everyone I know”: that is a sample, not evidence.
“Tripled”: no base rate. One to three is also triple.
The book will not make you clever.
It will make you pause for three seconds.
Those three seconds are its whole value.
Reference
Keith E. Stanovich, How to Think Straight About Psychology, 10th edition.
Dou Donghui and Liu Xiaocen, trans., Renmin University Press.
