Cognitive Biases That Fool Smart People (and How to Catch Them)
Confirmation bias, anchoring, availability, hindsight and more: ten shortcuts that mislead clever people, the experiments that exposed them, and what helps.
Try the card puzzle at the top before reading on. Four cards, one rule, and a single question: which cards do you need to turn over to check it?
The psychologist Peter Wason devised it in 1966, and it has been run thousands of times since. Fewer than one person in ten gets it right. Most choose E and 4. The answer is E and 7, and the reason most people miss the 7 says something about how all of us check our beliefs: we look for cases that would confirm them, not cases that could prove them wrong.
That’s a cognitive bias. Not a lack of intelligence, not carelessness, but a predictable tilt in how the mind works. Here are ten of the best documented, the experiments behind them, and what actually helps.
Try it
Four cards, one rule
The rule: “If a card has a vowel on one side, it has an even number on the other.” Each card has a letter on one side and a number on the other. Which cards must you turn over to find out whether the rule is true?
Tap the cards you’d turn over, then turn them.
Why biases exist
The mind uses shortcuts, called heuristics, to make fast judgements with little effort. Most of the time they work well enough: judging a risk by how easily examples come to mind, trusting the first answer that feels right, going with the familiar option. The psychologist Daniel Kahneman described the two modes as “System 1”, fast and automatic, and “System 2”, slow and effortful.
A bias is what happens when a shortcut is used in a situation it doesn’t fit. It isn’t random. It goes wrong in the same direction, for most people, again and again. That predictability is what made these errors scientifically interesting, and what makes them possible to guard against.
1. Confirmation bias
We look for, notice and remember evidence that supports what we already think, and we explain away evidence that doesn’t. In the card task, people turn the 4 because an even number with a vowel behind it would fit the rule. But nothing on the back of the 4 can break the rule. The 7 can, and it’s the card people ignore.
It shows up everywhere: in the news we choose to read, in how we interpret ambiguous evidence, in doctors fixing on a first diagnosis and investors reading every headline as support for a stock they own. The fix isn’t to stop having beliefs. It’s to ask the question the 7 represents: what would I expect to see if I were wrong, and have I looked for it?
2. Anchoring
When we estimate a number, we tend to start from whatever number we’ve just seen and adjust too little, even if that number was obviously arbitrary.
Try it
A number that shouldn’t matter
In a famous experiment, people spun a wheel of fortune that stopped at 10 or at 65, then estimated the percentage of African countries in the United Nations. What did the wheel’s number do?
Pick one. You’ll see why straight after.
The same effect has been found in judges’ sentencing, house valuations by estate agents and salary negotiations. It’s why a high first offer tends to pull the final deal upwards, and why “was £400, now £200” makes £200 feel like a bargain whatever the coat is worth.
3. Availability
We judge how common or likely something is by how easily examples come to mind. Things that are vivid, recent or widely reported come to mind easily, so we overrate them. People tend to overestimate deaths from dramatic causes such as tornadoes, murders and plane crashes, and underestimate deaths from common, quiet causes such as diabetes and strokes. In the US, cows kill more people each year than sharks, but which one gets the films?
4. Hindsight bias
Once we know how something turned out, we feel we could have predicted it. In 1975, Baruch Fischhoff gave people accounts of historical events with different outcomes and found that whatever outcome they were told, they rated it as having been the likely one. Hindsight makes the past look more predictable than it was, which makes us too harsh on people who made reasonable decisions that turned out badly, and too confident about predicting the future.
5. Overconfidence
People are, on average, more sure of their judgements than their accuracy justifies.
Try it
Better than average?
In a 1981 study, American students were asked to compare their driving skill with everyone else in the room. What percentage rated themselves as safer than the median driver?
Drag the dot to your guess, then lock it in. Nothing is revealed until you do.
In studies where people give a range they’re “90% sure” contains a true answer, the true answer falls inside far less than 90% of the time. Overconfidence is one reason so many projects run late and so many investors trade too often.
6. Framing
The same choice can be presented in different words, and people choose differently. In a classic experiment by Tversky and Kahneman, a disease is expected to kill 600 people. Offered a programme that “saves 200 people” or a gamble, most chose the sure thing. Offered a programme under which “400 people will die” or the same gamble, most chose the gamble. The two sure options are identical. Only the wording changed.
7. The sunk cost fallacy
We keep investing in something because of what we’ve already put in, even when it no longer makes sense. The ticket you’ve paid for, the course you’re halfway through, the project that’s years over budget. What’s spent is gone either way; the only question is what’s best from here. There’s more on it in the sunk cost fallacy.
8. The halo effect
One good trait colours our view of unrelated ones. The term comes from the psychologist Edward Thorndike, who found in 1920 that military officers rating their soldiers gave correlated scores across very different qualities: a soldier rated handsome was also rated intelligent, loyal and good at his job. Attractive defendants receive lighter sentences in some studies; charming interviewees get rated as more competent.
9. The planning fallacy
We underestimate how long our own tasks will take, even when we know that similar tasks took longer in the past. In one study, students predicted they’d finish their theses in about 34 days on average. The average was about 55. There’s a whole article on why projects always take longer than planned.
10. The bias blind spot
And here’s the one that makes the others hard to fix. People readily see biases in others and rate themselves as less biased than average. Emily Pronin and colleagues found this pattern across many biases in 2002. Reading a list like this one tends to produce the thought “other people do this”, which is itself the bias at work.
Do smart people fall for them?
Yes, often just as much. Some biases are only weakly linked to intelligence, and some barely at all. Worse, clever people can be better at building arguments for what they already believe, which can make confirmation bias stronger. Expertise helps in fields with quick, clear feedback, such as weather forecasting, and much less in fields without it.
What actually helps
Knowing about a bias rarely fixes it in the moment. The shortcuts run automatically. What works is changing the situation the decision is made in.
- Consider the opposite. Before deciding, deliberately argue for the other conclusion. Studies have found this simple step reduces several biases more than being told to be fair.
- Look up the base rate. How did similar cases turn out? This “outside view” is the best single fix for the planning fallacy and overconfidence.
- Write predictions down. It’s the cure for hindsight bias, and it lets you check how well calibrated you are.
- Run a pre-mortem. Imagine it’s a year from now and the plan failed. What happened? It surfaces risks people are reluctant to raise.
- Score things separately. Rate each quality before you form an overall view, to keep a halo from spreading.
- Set your own anchor first. Before you see an asking price or an offer, decide what you think it’s worth.
None of this makes anyone perfectly rational. It just tilts the odds back towards good judgement, which is the most anyone can do.
Test yourself
The free thinking traps test has twenty questions on biases and bad arguments, and the psychology myths test covers popular beliefs about the mind. If you want the most famous puzzle of all, start with the bat and ball problem.
Questions people ask
- What is a cognitive bias?
- A systematic error in judgement that comes from the shortcuts the mind uses to make quick decisions. The shortcuts, called heuristics, are usually useful, but in some situations they lead people astray in predictable ways, which is what makes them biases rather than random mistakes.
- What are the most common cognitive biases?
- Some of the best documented are confirmation bias, anchoring, the availability heuristic, hindsight bias, overconfidence, framing effects, the sunk cost fallacy and the planning fallacy. Each has been shown in many experiments, often in experts as well as students.
- Can you get rid of cognitive biases?
- Not entirely, and simply learning about them helps less than people hope. What works better is changing how decisions are made: deliberately considering the opposite case, using checklists and base rates, getting outside views, and recording predictions so you can check them later.
- Are smart people less biased?
- Not much. Some biases are weakly related to intelligence, but many aren’t, and clever people can be better at constructing arguments for what they already believe. People also tend to see biases in others more easily than in themselves, which psychologists call the bias blind spot.
- What is the difference between a bias and a heuristic?
- A heuristic is a mental shortcut, such as judging how common something is by how easily examples come to mind. A bias is the systematic error a heuristic produces in situations where it doesn’t fit, such as overestimating the risk of plane crashes because they’re vivid and widely reported.
Sources and further reading
- Tversky, A. and Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131.
- Wason, P. C. (1966). Reasoning. In Foss, B. M. (ed.), New Horizons in Psychology. Penguin.
- Svenson, O. (1981). Are we all less risky and more skillful than our fellow drivers? Acta Psychologica, 47(2), 143–148.
- Fischhoff, B. (1975). Hindsight is not equal to foresight. Journal of Experimental Psychology: Human Perception and Performance, 1(3), 288–299.
- Pronin, E., Lin, D. Y. and Ross, L. (2002). The bias blind spot: Perceptions of bias in self versus others. Personality and Social Psychology Bulletin, 28(3), 369–381.
- Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
