What the questions are built to do
Each problem offers a tempting shortcut. You might subtract two prices and forget that the difference must leave two separate amounts, scale the work but not the number of workers, or average two speeds without checking the time spent at each one.
The practice has seven questions, typed numerical answers and no time limit. Its six question families cover price differences, rates, doubling, percentage reversals and two kinds of average. After each answer, you see the check, whether you were right or wrong.
These are questions written for QuickenBrain in the style of the cognitive reflection test. They are not the original research test, and a result out of seven should not be converted into the original score out of three.
The original research question
Shane Frederick introduced the three-item CRT in Cognitive Reflection and Decision Making (2005). The questions were designed to invite an incorrect first response that could be checked with a little further thought. He examined how scores related to choices involving time and risk.
That is a research relationship between measures, not a guarantee about a particular person's judgment. The original paper did not validate QuickenBrain's question bank. Here, the useful object of attention is the check you make on a particular problem, not a label such as intuitive person or rational person.
A worked example with two checks
Suppose a ticket and a wristband cost $6.60 together. The ticket costs $6.00 more than the wristband. What does the wristband cost?
Sixty cents is tempting because $6.60 minus $6.00 is $0.60. But try that answer in the whole question. A $0.60 wristband would leave $6.00 for the ticket. The total works; the difference is only $5.40. One condition passed and the other failed.
Instead, take the $6.00 difference out of the total. The remaining $0.60 contains two copies of the cheaper price, so divide it in half. The wristband costs $0.30 and the ticket $6.30. They add to $6.60 and differ by $6.00.
This example is written for the guide. The arithmetic can be checked directly; it is not a result reported by a study. The same method works with other totals and differences, provided the two prices satisfy both conditions.
What the two result counts mean
Right answers counts one point for each correct response. Fast wrong answers counts responses that match the specific tempting wrong number assigned to a question. Other incorrect answers are still incorrect, but they are not included in that second count.
The word “fast” names the answer pattern. It does not measure how quickly you responded. You could spend a minute calculating that number and still receive the label. A correct answer given immediately is still correct. The clock records active time but does not change either count.
A score alone cannot show whether you noticed the trap, remembered an earlier explanation, used a written equation or arrived by a different route. Nor can seven questions establish how you make decisions in other settings. A completed set can be saved in this browser; an unfinished one is not saved as a completed result.
What seeing a question before changes
Familiarity is worth noting, but the evidence is not as simple as “seen it before, therefore meaningless”. Białek and Pennycook (2018) found higher accuracy among participants who reported prior exposure, while the test's relationships with their other measures were not generally weakened.
Meyer, Zhou and Frederick (2018) followed repeated administrations rather than relying only on reported familiarity. Across 14,053 participants, the average change per exposure was small. They also distinguished repeated testing from exposure accompanied by feedback.
Here you receive an explanation after every question, so repeating a family is deliberately a learning situation. Do not read a later result as if you had never met the structure before. Instead, ask whether you can explain why the answer fits, including when the numbers change.
A method you can apply before pressing Check
- Keep the first answer provisional. You do not have to reject it because it arrived quickly. You do need to test it.
- Write down the conditions. For a price problem, keep both the total and the difference. For a rate problem, track the amount of work, the workers and the time.
- Put your answer back into the question. Calculate what else must then be true. Look for a condition that no longer works.
- Check the requested unit. A correct dollar amount entered into a cents field is a different number.
Read the explanation even when you are right. If it uses another route, compare the conditions both routes preserve. There is no benefit in racing the clock on this practice.
The Check the fast answer lesson works through the original bat-and-ball structure. The Cognitive reflection practice gives you a new seven-question set. For the related habit of checking what a statement guarantees, try Logical deduction.
Questions people ask
What is the cognitive reflection test?
The cognitive reflection test is a set of three questions that Shane Frederick introduced in 2005. Each is built to invite a quick wrong answer that a little further thought can catch. Frederick looked at how scores related to choices involving time and risk. The practice on this site uses seven questions written in the same style, not the original three.
What is the answer to the bat and ball question?
The ball costs 5 cents. The question says a bat and a ball cost $1.10 together, and the bat costs $1.00 more than the ball. The quick answer, 10 cents, fits the total but leaves a gap of only 90 cents. Take the $1.00 gap out of $1.10 and split the remaining 10 cents in two: a 5 cent ball and a $1.05 bat.
Why do people get cognitive reflection questions wrong?
Each question offers a tempting shortcut that fits part of the wording. You might subtract two prices and forget that the difference must leave two separate amounts. You might scale the work but not the number of workers, or average two speeds without checking the time spent at each. The quick answer passes one condition and fails another.
Does seeing the questions before change the result?
A little. Białek and Pennycook found in 2018 that people who reported seeing the questions before scored higher, while the test's links with their other measures were not generally weakened. Meyer, Zhou and Frederick followed repeated testing across 14,053 participants and found the average change per exposure was small. Feedback after each question, as this practice gives, is a different situation.
How do you check a tempting answer?
Put it back into the question. First write down every condition, such as both the total and the difference in a price problem. Then test your answer against each condition and look for one that no longer works. Check the unit the question asks for. Keep the first answer provisional: you do not reject it for being quick, but you do test it.