Seven puzzles with an obvious wrong answer. Measures whether you stop before you answer.
7 questions across 1 scale, about 5 minutes|Published scoring key from Cognitive Reflection Test (3-item original plus extended items), not invented
Authors: Frederick, S. (2005); extended items from Toplak, West & Stanovich (2014) and Thomson & Oppenheimer (2016)
Licence: Published in the academic literature and widely reproduced for research and education
The tendency to catch and check an intuitive answer rather than acting on it.
High. You consistently notice when an answer arrived too easily, and you verify before committing.
Low. Your first answer usually wins. That is fast and often fine, and it is exactly what these items exploit.
In 2005 Shane Frederick published three questions. Each had an answer that arrives instantly and is wrong, and each had a correct answer available to anyone who paused. Those three items went on to predict time preference, risk preference and susceptibility to a range of reasoning errors better than much longer cognitive batteries did.
The finding that made the test famous was not that people get them wrong. It is who gets them wrong. Frederick administered the items at elite universities and still saw majority failure rates on the bat-and-ball question. Reflection and intelligence turned out to be related but clearly separable.
Because the original three became so well known, researchers published extended sets to replace items that respondents had already seen. The seven here are Frederick's three plus four standard extensions.
7 answers
Each answer you give is a number on the published response scale.
One scale
Answers belonging to the same scale are added together.
A band
Each total is read against the bands printed for that scale.
One point per correct answer, zero for anything else, out of seven. Each item's distractor set includes the intuitive wrong answer that the item was designed to elicit.
Scoring runs in your browser from your own answers. It is deterministic, so the same answers always produce the same result, and no part of it is generated, estimated or inferred.
Seven items is a short test, and a one-point difference between two people means very little. What the CRT measures reliably is the broad distinction between someone who habitually checks a first answer and someone who does not.
The real threat to validity is exposure. Once you have seen the bat-and-ball problem, your answer next time measures memory rather than reflection, which is precisely why extended item sets exist. Take this once and treat a retake as entertainment.
The correlation with general intelligence is real but modest. People with high IQ scores fail these items regularly, which is the more interesting finding and the reason the test is not a proxy for ability.
Worth taking if
Not the right tool for
Not an intelligence test. The correlation with IQ is real but modest, and plenty of very able people score zero.
Single-use: once you have seen an item, your answer to it next time measures memory rather than reflection.
Seven items is a short scale, so a one-point difference means nothing.
This is a self-assessment questionnaire, not a diagnostic instrument and not medical advice.
7 questions, about 5 minutes. Your answers are scored on this device and stored on this device. There is no account to create and nothing is sent to a server. Progress saves as you go, so a long assessment can be finished across more than one sitting.
No sign up, ever. Every item in this assessment comes from Cognitive Reflection Test (3-item original plus extended items), and the page above tells you exactly how each score is worked out before you answer anything.
A bat and a ball cost $1.10 in total and the bat costs $1.00 more than the ball. The answer that springs to mind is 10 cents; the correct answer is 5 cents, because a 10-cent ball would make the bat $1.10 and the total $1.20.
No. It correlates modestly with measures of general intelligence, but it is measuring a different thing: whether you override an intuitive answer rather than whether you can work out the right one.
Most adult samples average well under half correct on the original three-item version. On the seven items here, four or more puts you above the usual range, but see the note about having encountered the puzzles before.
Yes. Each item explains its keyed answer as soon as you respond, including why the intuitive answer fails.
About five minutes. There is no time limit, which is the point. The test is about whether you use the time, not whether you have it.
Items 1-3 are Frederick's original three; the remainder are standard extended-set items. Keyed answers checked against the published solutions. Check the source.