Cognitive Reflection Assessment
NOESIS adaptation of the Cognitive Reflection Test paradigm (Frederick, 2005); all 12 items are original NOESIS creations
~12 min · 12 questions · Free partial report
Each of these twelve problems has an answer that jumps out at you, and that answer is wrong. Getting it right takes noticing the impulse, distrusting it, and checking. That skill, cognitive reflection, is what this assessment measures: not how much you know, but whether you verify what feels obvious. The paradigm comes from Shane Frederick's Cognitive Reflection Test (2005), the research behind the System 1 and System 2 vocabulary: a fast, automatic mind proposes an answer, and a slower, deliberate mind either checks it or waves it through. Frederick's three original problems became so famous that they stopped working; by 2016, studies found that about half of online respondents had already seen them, and the answers sit one search away. That is why every item here is original. NOESIS wrote twelve new problems in the same paradigm, each built around a different trap: misleading averages, exponential growth, stacked discounts, sunk costs, and eight more. You will not have seen them before, which is exactly the condition the measure needs. The honest counterpart: new items carry no independent validation studies yet, and your result is a plain count of correct answers out of 12, not a normed percentile. The full report shows every problem again with your answer, the correct answer, and a short explanation of how the trap works. Missing an item and then seeing the mechanism laid bare is where this format earns its keep.
Twelve traps you have not met before, and afterwards the mechanism of every one of them.
- See how often you catch the intuitive wrong answer before it becomes your final answer, across twelve traps you have not met before. Read, for every item, what the lure was and how the correct path goes. Take the System 1 and System 2 distinction home as something you experienced rather than something you read about.
No signup required to start. The Partial Report is free at the end. The Complete costs $9.90, if and when you want it.
What this test measures
Cognitive reflection
Count of the twelve trap problems answered correctly. Each item plants a strong intuitive wrong answer; a point is earned only when that pull is noticed, checked and overridden. A plain count out of 12, not a normed or converted score.
Intuition led
Checked sometimes
Reflective
Your result falls into one of these bands.
Discover where you stand on each dimension.
Start for free →No prior signup. Free partial result immediately.
From first click to report.
The process is simple: you respond, the system calculates, and the result appears. No prior signup, no form, no waiting.
- 1
Click "Start for free"
No signup required to start. You identify yourself at the end, when you want to save the result. The test opens immediately.
- 2
Answer the questions
Multiple choice, 4 options, one correct answer per item. Visible progress bar. You can pause and resume from any device after identifying yourself.
- 3
Get the Partial Report
Immediate and free. A text summary of each dimension, with indication of high, medium, or low.
- 4
Decide if you want the Complete
The Partial already delivers value. The Complete adds percentiles, norm comparison, charts, applied recommendations, and a verifiable Certificate.
- 5
Explore in the Integrative Chat
Talk to an AI about your results. Ask how your profile applies to your career, what a high score means, or compare with other tests.
The Partial is yours right away. The Complete, when it makes sense.
| Free and paid, side by side | Partial Report (free, always) | Complete Report ($9.90) |
|---|---|---|
| Interpretive text per dimension, generated from your answers | ✓ | ✓ |
| Raw score for each dimension | ✓ | ✓ |
| One-paragraph overview reading | ✓ | ✓ |
| Everything in the Partial, expanded to 3–5 paragraphs per dimension | — | ✓ |
| Percentile per dimension, compared to reference norms | — | ✓ |
| Inter-dimensional chart with accessible data table | — | ✓ |
| Applied recommendations: career, relationships, self-care | — | ✓ |
| Certificate with public verification and PDF download | — | ✓ |
- ✓Everything in the Partial, expanded to 3–5 paragraphs per dimension
- ✓Percentile per dimension, compared to reference norms
- ✓Applied recommendations: career, relationships, self-care
- ✓Certificate with public verification and PDF download
You only pay after completing the test, if you want to go beyond the Partial.
- ✓Interpretive text per dimension, generated from your answers
- ✓Raw score for each dimension
- ✓One-paragraph overview reading
- ✓Free trial of the Integrative Chat, unlocked by your first test
Payment via PIX (instant) or credit card. Access for 1 year. After that, the report stays visible in read-only mode.
Why trust this instrument.
Built in the cognitive reflection paradigm introduced by Frederick (2005), this assessment uses 12 original NOESIS problems, each engineered around a different trap: misleading averages, exponential growth, stacked discounts, sunk costs and eight more. Every item offers four options, exactly one of them correct and one of them the dominant intuitive lure, and the score is a plain count of correct answers from 0 to 12, read through three NOESIS bands. There is no time limit and no norm set: the items are original, so no published calibration applies to them, and the reference values in the classic literature describe the classic instruments. The full report shows every problem again with your answer, the correct one, and how the trap works. The items are original rather than borrowed for a measurable reason: by 2016 about half of online respondents had already seen the famous problems, and prior exposure inflates results. What is measured is the disposition to check an intuition, which correlates with cognitive ability without being an intelligence test.
- Authors:
- Shane Frederick (paradigm); NOESIS (items)
- Year:
- 2005
The reliability, norms, and cutoffs shown are those of the original instrument (Cognitive Reflection Test paradigm (Frederick, 2005); all 12 items are original NOESIS creations).
This NOESIS adaptation has not yet been independently validated.
- Frederick, S. (2005). Cognitive Reflection and Decision Making. Journal of Economic Perspectives, 19(4), 25-42. See also Toplak, West & Stanovich (2014); Thomson & Oppenheimer (2016); Primi, Morsanyi, Chiesi, Donati & Hamilton (2016); Sirota & Juanchich (2018).
The theory behind this test.
Dual-Process Theory of Reasoning
Framework describing reasoning as the interaction of two modes: a fast, automatic process that proposes an answer, and a slow, deliberate process that either checks the proposal or waves it through. Frederick (2005) made the second one measurable with problems engineered so that the intuitive answer is confidently wrong, so a correct response requires noticing the pull, distrusting it and verifying. The disposition this measures, cognitive reflection, correlates with cognitive ability without being it: the items demand almost no arithmetic and no knowledge, and errors cluster on the planted answer rather than spreading at random. The paradigm's own central threat to validity is prior exposure — the classic items became so famous that by 2016 roughly half of online respondents had already seen them.
Learn more about this theory →Target Audience
Anyone curious whether they check their first impulse or trust it: puzzle lovers, readers of behavioral economics, and people who make quick calls all day and want to know what that speed costs.
Recommended ages: 18+
Questions the report answers.
When the obvious answer appears, do you check it or ship it?
Read the result as a count of caught impulses, 0 to 12. High means the planted lure rarely became your final answer: you paused, distrusted, verified. Low means the traps mostly worked, which is the ordinary human default and says nothing about intelligence. The band matters less than the pattern in the answer key: most people miss one particular family of traps, percentages, averages or scaling, and finding out which family is yours is the part you can actually use.
Who benefits most.
The behavioral economics reader
You know the theory: a fast mind proposes, a slow mind checks. But the famous demonstration problems stopped working on you the day you read about them. Twelve problems you have never seen give you back the real experience of being measured, instead of a memory quiz about answers you already know.
The confident quick thinker
Fast answers are your reputation, and most of the time they are right. This is a quarter of an hour against twelve problems built to punish exactly that speed. Either you finish vindicated, or you find the one family of traps that reliably gets you. Both results are worth having.
The second-guesser
You already distrust your first impulse, sometimes to a fault. A result here separates useful checking from anxious rechecking: if you land high, your slow habits are buying real accuracy. The report also shows on which kinds of problem the double-checking actually changed the outcome.
Frequently asked questions
Is this an IQ test?
No. It measures one narrow thing: the tendency to check an intuitive answer before committing to it. In the research literature this tendency correlates with cognitive ability measures, but a 12-item multiple-choice assessment is not an intelligence test and your result is not an IQ. A low count here means the traps worked on you that day, nothing more.
Can I take it again?
You can, but the second result means something different. These items lose part of their bite once you have seen the explanations: you are no longer suppressing an intuition, you are remembering an answer. Prior exposure inflates results; that is exactly what happened to the classic reflection problems over two decades of fame, and it is why we wrote new ones. Treat a retake as revision, not as a fresh measurement.
Why did you not use the famous original problems?
Two reasons. First, the classic items are among the most widely circulated puzzles in the world; studies in 2016 found that about half of online respondents had already seen them, and prior exposure measurably inflates results. Charging you for a score on problems you may already know would be charging you for noise. Second, those items sit in journal articles under publisher copyright, with no license for commercial reproduction. Original items solve both problems at once.
Does a low score mean I am not smart?
No. It means that on these twelve problems, the planted intuitive answers got past your checking more often than not, which is the most common human outcome. People with strong formal training fall for these when tired or rushed. The useful part of a low count is the answer key: each explanation shows the exact move the trap made, and that pattern can be learned.
Is there a time limit?
No. Reflection is the whole point, so nothing here pushes you to answer fast. Take the time to feel the pull of the obvious answer and then check it. What we do ask is that you answer without a calculator and without searching: the arithmetic is deliberately light, and an outside lookup would turn a reflection measure into a typing exercise.
Can a company use this to screen candidates?
We advise against it and flag this test as a poor fit for hiring decisions. It runs unsupervised, the answer to any reflection-style problem is searchable, and a motivated candidate can prepare. It works as self-knowledge and as a starting point for talking about decision habits on a team, not as a selection gate.
Cognitive Reflection Assessment
No signup, no time limit. The Partial is free at the end.
The Complete costs $9.90, if and when you want it.
This test is a self-knowledge tool for informational purposes. It does not constitute a psychological or clinical diagnosis and does not replace evaluation by a qualified professional. The report indicates the AI model used in text generation.
All twelve items are original NOESIS creations in the cognitive reflection paradigm introduced by Frederick, S. (2005), Cognitive Reflection and Decision Making, Journal of Economic Perspectives, 19(4), 25-42. No item reproduces or adapts the classic CRT item sets (CRT-3, CRT-7, CRT-2, CRT-Long), whose questions and answers circulate publicly. © 2026 NOESIS. All rights reserved.