THE QUESTION LIBRARY HOW WE KNOW
Every human mind takes shortcuts, including the minds of scientists and of the people who wrote this page, so here are the best-known traps and the habits that help.
THE SHORT ANSWER
Everyone falls into thinking traps, including scientists. The best known include confirmation bias (favoring evidence that fits what we already believe), motivated reasoning (reaching the answer we want), judging by what comes to mind easily, ignoring how common something is, seeing patterns in random noise, and survivorship bias (studying only the winners). You can’t switch these off. But habits help: look for what would prove you wrong, ask for the numbers, and let other people check your work.
WHERE THE EVIDENCE STANDS
WHAT WE DON’T KNOW YET
Researchers still debate how big some biases are outside the lab, and how much training can reduce them for good. Some effects that once looked solid have shrunk under larger tests, and more may follow.
WHAT WOULD CHANGE THIS ANSWER
Large, careful replications that failed to find confirmation bias or the availability effect would change this answer. So would strong evidence that simple training removes these biases for good.
Everyone has thinking traps. They are not signs of low intelligence. They are shortcuts the human mind uses all the time, and most of the time they work well. But they also lead to mistakes we can predict. Scientists fall into them. So do experts in many fields, and so do the writers of this page.
People ask about thinking traps for many reasons. Some want to judge a video or a viral post. Some are rethinking what they were raised to believe, or wondering why a relative sees the world so differently. Some want to check their own reasoning. All of these are good reasons to learn how our minds slip.
Here are the traps that show up most often. Confirmation bias is favoring evidence that fits what we already believe. Motivated reasoning is reaching the conclusion we want and then finding reasons for it. The availability heuristic means judging how common something is by how easily examples come to mind. Base-rate neglect means forgetting how common something is to begin with. Patternicity is seeing meaningful patterns in random noise. The Barnum effect is feeling that a vague description was written just for you. And survivorship bias is drawing lessons only from the winners, because the losers are not around to be counted.
These are well studied. In a classic 1960 experiment, most of the bright young adults tested announced at least one wrong rule, and people typically tested only cases that fit their guess. In 1974, two psychologists showed that mental shortcuts make people misjudge how often things happen.
You can’t switch these traps off. Even knowing about them does not make you immune. But habits help. Ask what would prove you wrong. Ask for the numbers. Ask who is missing from the story. And let other people check your work, because that is how science catches its own mistakes.
Confirmation bias is the habit of seeking or reading evidence in ways that favor what we already believe. In a 1998 review, Raymond Nickerson called it “a ubiquitous phenomenon in many guises,” meaning it shows up almost everywhere.
The classic test came from the psychologist P. C. Wason, working in London, in 1960. He told people that the numbers 2, 4, 6 fit a simple rule. Their job was to find the rule by making up new sets of three numbers. After each set, Wason said whether it fit. A common guess, as Nickerson describes it, was “successive even numbers.” People typically tested only sets that fit their guess, such as 8, 10, 12. Every answer came back “yes,” which seemed to confirm it. But the real rule was broader: any three numbers in increasing order. Only 6 of 29 people found it without first announcing a wrong rule. The trick is that a “yes” to a set that fits your guess can’t show that your guess is wrong. Only a test your guess says should fail, such as 1, 2, 3 or 5, 10, 20, can do that. Both get a “yes” under the real rule, which shows the guess is too narrow.
Confirmation bias is close to motivated reasoning. In a 1990 review, the psychologist Ziva Kunda reviewed considerable evidence that people are more likely to arrive at conclusions they want to reach. The review also found a limit: people can only do this as far as they can build reasons that seem sensible. Nickerson describes a related finding. Two people with opposite views can read the same evidence and both come away more sure of their own side.
Smart people are not protected. In their famous 1974 paper in Science, Amos Tversky and Daniel Kahneman wrote that “experienced researchers are also prone to the same biases” when they think intuitively. And in a 2002 study, Emily Pronin and her colleagues found that people see bias much more in others than in themselves. They called this the bias blind spot.
Tversky and Kahneman described the availability heuristic: we judge how common something is by how easily examples come to mind. That often works, because common things are usually easier to recall. But other things affect memory too. In one experiment, people heard a list of well-known men and women. Whichever sex had the more famous names seemed more numerous, even when it wasn’t. In another, people guessed whether English words more often start with r or have r as the third letter. Most chose the first, because words are easier to search by their first letter. Yet r is more common in the third spot.
These classic demonstrations have a mixed record. A 1995 repeat of the famous-names test found a moderate effect, but a 2001 repeat found a smaller one. And a 1998 study by a team in Germany noted that the letter result had almost never been repeated in print. In their own three studies, people’s judgments of whether a letter is more common first or second in words generally matched the real proportions.
Tversky and Kahneman’s 1974 paper also described base-rate neglect. People read short descriptions said to come from a group of engineers and lawyers. One group was told there were 70 engineers and 30 lawyers. The other was told the reverse. Those numbers should have changed people’s guesses a lot, but the two groups gave almost the same answers. Even for a description that said nothing useful about the man’s job, people rated the odds of an engineer at 50-50, whatever the mix.
In a 1996 review, Jonathan Koehler argued that the base-rate problem has been oversold. He found that people almost always use base rates to some degree, and use them more when the numbers are given as simple counts. In short, people often give base rates too little weight, depending on how a problem is framed.
Writing in Scientific American in 2008, Michael Shermer gave the name patternicity to the tendency to find meaningful patterns in meaningless noise. Psychologists also use the word apophenia, a leaning toward false-positive errors. Shermer points to a 2008 paper in which the biologists Kevin Foster and Hanna Kokko built a simple model. It shows how natural selection can favor this habit when missing a real pattern costs more than believing a false one. Shermer’s example: if an animal treats a rustle in the grass as a predator and it turns out to be the wind, it loses little; if it makes the opposite mistake, it might not survive. Where the example stops fitting: it is about split-second reactions to danger, while many of the patterns people argue about today are beliefs we have time to check.
The Barnum effect is a close cousin. In 1949, the psychologist Bertram Forer gave his class a personality test. A week later, each of the 39 students got a sketch with their own name on it. Every sketch was the same, built from general lines such as “You have a tendency to be critical of yourself.” Forer noted the lines came largely from a newsstand astrology book. Asked how well the sketch revealed their personality on a scale of 0 to 5, only 5 students rated it below 4. A 1985 review of later studies found that people accept such profiles more when they seem relevant and flattering. Our answer on astrology looks at how this plays out in horoscopes.
Survivorship bias means drawing conclusions only from the cases that made it through, because the ones that didn’t are not there to be seen. The most famous example comes from World War II. At Columbia University’s Statistical Research Group, the statistician Abraham Wald worked on estimating how vulnerable aircraft were, using data from planes that came back. The hard part, as a 1984 account of his work explains, was that the damage on downed aircraft could not be observed. Planes hit in the most dangerous places were more likely to be missing from the count. Wald built methods to reason about those missing planes. In that account’s worked example, which uses made-up numbers, the engine area turned out to be the most vulnerable part.
The same trap shows up in daily life. If we only hear from people who quit school and got rich, we learn the wrong lesson. Ask: who didn’t make it, and why can’t I see them?
The strongest objection is that psychology has had a replication problem, so maybe these biases are overblown too. Some famous findings did fail. The idea of “ego depletion” said that self-control runs down like fuel after a short effort. But a 2016 test across 23 labs with 2,141 people found an effect near zero. A 2021 test across 36 labs also found no evidence for it in its planned analyses.
But a failure in one place does not sink everything. Confirmation bias has been found again and again in many settings, as Nickerson’s review shows. And when a 2014 project retested 13 classic and newer psychology effects across 36 samples, 10 replicated consistently, while two did not replicate.
“Biases are always bad.” Tversky and Kahneman called these shortcuts “highly economical and usually effective.” The trouble comes when we trust them where they fail.
“Knowing about biases makes me immune.” Pronin’s studies suggest the opposite risk: we spot bias in others and miss it in ourselves.
“Only other people fall for this.” Believers and skeptics, experts and beginners all use the same mind. Nickerson notes that science succeeds less because each scientist doubts their own ideas and more because scientists are highly motivated to show that other scientists’ ideas are false.
For a quick routine you can use on any viral post, see our guide How to Check a Viral Science Claim in Five Minutes. The answer stays the same: these traps are real, well studied, and part of every human mind.
THREE THINGS TO REMEMBER
WHERE THE AUTHOR’S RESEARCH TOUCHES THIS
Author’s hypothesis. In the book’s Introduction, Ricardo Maldonado, author of GOD PLAYS DICE, describes a falsifiable hypothesis that our Big Bang may be the aftershock of a higher-dimensional event. On this site it goes by the name HD-Blast. It is an Author’s hypothesis, not an established result. Two passages in the book deal with traps on this page.
In Chapter 5, the author sets aside a number pattern he has explored in particle masses. He calls it “a line of thought rather than a claim” and writes that with enough freedom, you can find a pattern almost anywhere. A real pattern, he argues, has to predict new measurements, not only explain old ones. In Chapter 11, under the heading “The Hardest Result to Correct Is Your Own,” he writes that correction is hardest when the idea is yours, and that a corrected claim is not a corrected human being. Whether HD-Blast survives is up to the data and to other people checking it.
WORDS WORTH KNOWING
KEEP ASKING
See a mistake? Report a problem. Corrections are made openly.
CONTINUE THE THREAD
How We KnowStory
A claim that breaks with a mountain of past evidence has to climb that mountain first, so here is where the rule came from, when surprising claims won, when they failed, and how to use the rule fairly.
