Cognitive Reflection Test

Take our 7-question version below. There is no time limit and no sign-up. You see every answer, with the reason the tempting one fails, at the end.

Every item has two answers: one that arrives instantly and feels obvious, and one you only reach by checking. The maths is never harder than primary school. What the test picks up is whether you notice that your first answer needs checking.

Psychologists call the fast, automatic mode System 1 and the slow, deliberate mode System 2, a framing popularised by Daniel Kahneman. The CRT is a small window onto how readily your System 2 steps in when System 1 is confidently wrong.

That is why your review shows two numbers: how many you got right, and how many times you picked the intuitive answer. Most wrong answers on a CRT are the intuitive one, not random slips.

CRT scores are moderately correlated with scores on intelligence and numeracy tests. People who do well on one tend to do better on the other, but the overlap is far from complete.

Researchers such as Maggie Toplak and Keith Stanovich have found that the CRT predicts avoiding common reasoning errors even after intelligence is taken into account. In other words, it captures a habit (checking yourself) as well as ability.

So this is not an IQ test, and your score here does not convert into an IQ number. It is a check on one thinking habit. An IQ test asks a different question: when a problem is genuinely hard, can you find the rule?

Shane Frederick, then at MIT, published the original three-question CRT in 2005 in the Journal of Economic Perspectives. Its best-known item is the bat and ball: together they cost $1.10, the bat costs $1.00 more than the ball, so what does the ball cost? Most people say 10 cents. The answer is 5 cents.

In Frederick’s sample of about 3,400 people, the average score on those three questions was 1.24 out of 3. More than half of the students tested at Harvard, MIT and Princeton gave the intuitive answer to the bat-and-ball question.

Because the original questions became famous, researchers later wrote new items, often called CRT-2, that work the same way. Ours are new versions with different numbers and settings, so results are not directly comparable with published scores.

Getting a CRT item wrong feels different from getting an ordinary question wrong. You did not lack knowledge. You had the answer within reach and walked past it. That small jolt is what makes these puzzles fun to send to a friend.

One honest caveat: once you have seen the trick, it stops working. A second attempt tells you what you remember, not how reflective you are. If you want a fair read on a friend, send them the link before you tell them about the bakers.