A 20-question risk tolerance test scored separately across five domains: financial, health and safety, recreational, ethical and social. The finding this is built on is that risk-taking does not generalise — being bold in one domain says almost nothing about the others — so a single risk score would destroy the only information worth having. Not a financial suitability assessment. Scored in your browser.
Twenty statements, about two and a half minutes. Risk tolerance is not one trait — being bold in one part of life says almost nothing about the others — so this scores five domains separately and shows you where they differ. Scored entirely in your browser.
How to use this test
Answer for what you do, not what you admire
Every one of these domains has a socially preferred answer, and they point in different directions. A profile built on the answers you wish were true describes nobody.
Take the first reading and move on
Twenty statements, roughly two and a half minutes. Deliberating usually means constructing a justification rather than recalling what you actually did.
Read the shape, not the top score
The informative thing is the gaps between domains. High recreational with low social is a real and common shape, and it tells you more than either number alone.
Follow the leading domain to its page
Each of the five has a page on what it predicts, what it costs and how to work with it. The domain at the BOTTOM of your profile is usually worth reading too.
What this test measures
Ask someone whether they are a risk-taker and you will get one answer. The research says that answer is close to meaningless. Weber, Blais and Betz (2002) set out the finding this test is built on: risk tolerance does not generalise. The same person is routinely bold in one part of life and cautious in another, and the pattern is stable enough to measure. Collapsing it to a single "risk score" — which is what almost every popular quiz does — destroys the only information worth collecting.
So this scores five domains separately and never combines them.
Financial risk
Willingness to accept a worse money outcome for a shot at a better one: volatility, concentration, illiquidity. High tolerance here predicts starting things and sitting through drawdowns. Its characteristic error is not taking a risky bet but taking a bad bet that happens to be exciting — mistaking a tolerance for variance for an edge.
Health and safety risk
The domain where the downside is your own body. It correlates least well with the other four, so people are often surprised to find it leading their profile — they think of themselves as careful, and are, about money. What makes it different is the asymmetry: financial losses are recoverable in principle, and many health outcomes are not. A tolerance calibrated in a domain with reversible outcomes behaves badly in one without them.
Recreational risk
Wanting the version of an activity with genuine uncertainty in it. This is the most visible domain — it is what people mean when they call someone a risk-taker — and among the least predictive of the rest. Someone can be genuinely fearless on a mountain and unable to send a confrontational email.
Ethical risk
Tolerance for being in the wrong: bending a rule, shading a description, keeping an advantage you were not owed. It is the only one of the five where the downside is mostly borne by somebody else, which is why it deserves more deliberate attention than the others. The social-desirability pull on these items is strong, so an honest high score is unusually informative.
Social risk
Willingness to be disagreed with, disapproved of, or thought wrong. This is most often the domain at the bottom of the profile. Its costs are small individually and substantial in aggregate: objections never raised, decisions shaped around how they will look, preferences withheld until they harden into resentment. It is also the most trainable of the five, because the feared outcome is usually both survivable and immediate, so the estimate updates quickly.
How to read your result
Each domain is scored from four statements on a five-point agreement scale, giving a range of 4 to 20. The useful reading is the shape: which domains sit high, which sit low, and how far apart they are. High recreational with low social is one of the most common profiles there is, and naming it tells you more than either number on its own.
Half the statements are worded in reverse, so agreeing with everything does not push every score up. If you agree — or disagree — with nearly all of them, including pairs that contradict each other, the test says so, because the shape is then much weaker evidence than it looks.
What this test is not
⚠️ It is not a financial suitability assessment. Regulated risk-profiling for investment advice is a separate exercise with different obligations and a different evidence base, and nothing here should be used in its place. It is also not a medical or safety assessment.
It uses an original RECATOOLS item set written for this site from the published domain definitions. It is not DOSPERT and makes no claim to that instrument's validation or norms. There is no norm sample behind these numbers, so each score is a position on a stated scale rather than a comparison against anyone else.
No domain here is the good one and none is a diagnosis. High tolerance is not courage and low tolerance is not wisdom; both are dispositions with costs, and the costs differ by domain. Risk tolerance also moves with circumstance more readily than personality traits do — dependants, a mortgage, an injury or a windfall all shift it.
Everything is scored in your browser. Your answers are held only on this device so you can resume if you stop partway, they are cleared when you finish, and they are never uploaded.
The five domains
Financial risk
Willingness to accept a worse money outcome for a shot at a better one. Predicts comfort with volatility — and a characteristic confusion between variance and edge.
Health & safety risk
The domain where the downside is your own body, and frequently not reversible. Correlates least well with the other four.
Recreational risk
Wanting the version with uncertainty in it. The most visible domain and among the least predictive of the rest.
Ethical risk
Tolerance for being in the wrong. The only domain that prices someone else's downside rather than your own.
Social risk
Willingness to be disagreed with or thought wrong. Most often found at the bottom of the profile — and the most trainable of the five.
Frequently asked questions
-
About two and a half minutes. There are 20 statements and you rate your agreement with each on a five-point scale.
-
Because risk tolerance does not generalise, which is the central finding of the research this is built on. The person who will free-climb is not reliably the person who will hold a volatile investment. A single number would throw away the only information worth collecting.
-
No. This is an original RECATOOLS item set written over the same domain framework. DOSPERT is a separate instrument with its own validation and norms, and none of its item text was used, adapted or consulted.
-
No. This is not a financial suitability assessment and must not be used as one. Regulated risk-profiling for investment advice is a different exercise with different obligations and a different evidence base.
-
That the statements did not pull you strongly in any direction. It is a genuine result and the test reports it rather than picking a winner.
-
Half the statements are reverse-worded, so agreeing with nearly all of them means agreeing with pairs that contradict each other. That pulls all five scores toward the middle and makes the ranking much weaker evidence, so the test says so.
-
No, and neither is a low one. These are dispositions, not grades. The exception worth naming is the ethical domain, which is the only one where the downside is mostly borne by somebody else.
-
Yes, more than personality traits do. Dependants, a mortgage, a redundancy, an injury or a windfall all move it. A result from a few years ago may describe someone you no longer are.
-
No. Everything is scored in your browser. Answers are held on this device only so you can resume if you stop partway, and are cleared when you finish.
-
No. Each score is your position on a stated 4–20 scale, not a percentile against other people. This item set has not been normed and makes no claim to have been.
-
Because it is the one where the socially preferred answer is most obvious. That pull is exactly why an honest high score there carries more information than a high score anywhere else — and why this test does not moralise about the result.
Pick up where you left off
Stored only in this browser — never sent to our servers.