LEVEL is a five-step check for deciding how far to rely on one claim before you act on it: Locate the claim, Estimate the stakes, Value the evidence, Expose the gap, and Lean accordingly. It comes from The Evidence-Aware Life, a book by Len P. van der Hof that is available now in English and Dutch. You do not need the book to use it. Everything you need is on this page.
The idea behind it fits in one sentence: the same evidence can be good enough for a small test and too weak for a big commitment. LEVEL keeps those two questions apart and makes you write down which one you are answering.
What problem does LEVEL solve?
Most bad calls do not start with a lie. They start with a claim that arrived well dressed: a confident tone, a credential, a clean chart, a friend who swears by it. The book calls this borrowed certainty: believing something because of how it was packaged rather than what supports it.
LEVEL is a short routine for looking past the packaging. It does not tell you whether a claim is true. It tells you what the claim can justify right now, for this decision, and what you would need to know before you do more.
What do the five steps mean?
| Step | The question you ask | In plain words |
|---|---|---|
| Locate the claim | What exactly is being claimed? | Write it as one sentence someone could check. If you cannot, you are judging the packaging, not the claim. |
| Estimate the stakes | What does it cost if I act and I am wrong? If I ignore it and it was right? If I wait? | Money, time, safety, and whether you can undo it. Stakes decide how careful you need to be. They do not make the claim more or less likely to be true. |
| Value the evidence | What supports this, and how good is that support? | Name the kind of evidence you actually hold, then check its quality, its independence (are three sources really one?) and whether it fits your situation. |
| Expose the gap | What is missing that could change my decision? | The distance between the evidence you hold and what this decision needs. Stating the strongest case against the claim is a good way to find it. |
| Lean accordingly | How far should I act? | Act only as far as evidence and stakes allow: go ahead, run a small test, ask a qualified person, or wait until a set date. Write down who does what, the limit, and what would make you look again. |
The result is a record of five short lines. It is written to be read later, by you or by someone else, so they can see what you knew when you decided. The LEVEL framework page keeps the one-line version of each step.
A worked example: “This heat pump will halve your energy bill”
Imagine you own a house heated by a gas boiler. An installer visits and says a heat pump will halve your energy bill. The example is hypothetical: the promise belongs to an imaginary installer, not to any real product.
Locate. “Halve your energy bill” hides several claims. A heat pump runs on electricity instead of gas, as Milieu Centraal, the Dutch public information centre on energy, explains. So your gas use goes down while your electricity use goes up. “Halve our gas use” and “halve our total energy cost” are different claims. And compared with which year, at which prices, in which house? A checkable version reads: “In our house, heated the way we heat it now, the heat pump will cut our yearly gas-plus-electricity cost to about half of last year’s.”
Estimate. The purchase is expensive and you will not undo it next month. If you act and the claim is wrong, you have paid for savings that never arrive. If you refuse and it was right, you keep paying more than you need to. Waiting costs you another winter on the current bill. High stakes and little room to reverse justify careful checking. None of this makes the installer more or less likely to be right.
Value. List what you actually hold. The promise itself is an assertion until someone shows how it was worked out. The neighbour whose bill dropped is an anecdote, from a different house and a different winter. The installer is an authority, and also a seller. The mechanism is real: a heat pump uses electricity and often draws heat from the outside air, which is why it is efficient. But a mechanism tells you why something could work, not how much it will save in your house. If three neighbours quote the same installer’s figure, you have one source, not three.
Expose. The installer’s “half” rests on assumptions about your house. Milieu Centraal notes that which type of heat pump suits a home depends on how well it is insulated. So the missing piece is plain: what insulation, energy prices and usage did the calculation assume, and do they match yours? You can get that answer before you sign.
Lean. Do not sign today, and do not drop the idea either. Ask the installer for the calculation in writing, with its assumptions. Put your own meter readings from last year next to it. Get a second quote from another installer. Owner: you. Limit: no deposit before the written calculation. Review trigger: if the assumptions do not match your house, “half” is off the table; if they do, you decide with a real number instead of a sales line.
Same five steps, a different answer
The book shows the same method producing different next steps. Mara, a physiotherapist in the book (a composite character, not a real person), runs LEVEL twice.
First on a supplement she takes every morning. The stakes include safety and interactions, not only the price, and her support is a testimonial, a plausible mechanism and her own uncontrolled experience. Her Lean: take the claim and her history to a pharmacist or clinician, and stop presenting the benefit as proven.
Then on whether music helps her finish her administrative notes. Little is at risk and the trial is easy to stop. Her Lean: a short planned comparison of quiet and music sessions, with a rule to stop if the work suffers.
The method stayed the same. The claim, the evidence and the stakes changed, so the next step changed too.
What is the Evidence Ladder behind the Value step?
The Value step uses a second tool from the book, the Evidence Ladder. Its seven rungs name the kinds of support you will meet:
- Assertion: someone said it.
- Anecdote: it happened to someone.
- Authority: a qualified person says so.
- Mechanism: there is a plausible reason it should be true.
- Study: one real test looked at it.
- Convergence: several independent lines of evidence point the same way.
- Your own test: a planned comparison in your own situation.
The order runs from bare assertion toward tested evidence, but the rungs are categories, not scores. A poor study does not beat careful expert judgment just because it is a study. Quality, independence and relevance still have to be checked on every rung.
One well-known result shows why a single study is a rung and not a verdict. When the Open Science Collaboration repeated 100 studies from three psychology journals, 97% of the originals had statistically significant results, but only 36% of the repeats did, and the repeated effects were on average half as large. That finding describes those studies. It is not a failure rate for all of science.
Why keep “how good” and “how costly” apart?
Because mixing them is a classic error, and medicine built a system to avoid it. GRADE rates the quality of evidence separately from the strength of a recommendation; by 2008 more than 25 organisations, including the World Health Organization and Cochrane, had adopted it. Its authors put the point bluntly in the BMJ: “High quality evidence doesn’t necessarily imply strong recommendations, and strong recommendations can arise from low quality evidence.”
LEVEL borrows that separation for everyday claims. Value is about the evidence. Estimate is about the consequences. Lean is where the two meet. The book credits GRADE and states that its Ladder is a teaching tool, not a clinical grading system.
What LEVEL is not
- Not a truth score. It produces a decision record, not a number.
- Not a replacement for experts. For health, legal or money decisions, asking a qualified person is often the right Lean.
- Not a validated intervention. The book presents LEVEL as the author’s practical method and says plainly that the combined method has not been shown to improve outcomes. Judge it by whether it helps you decide.
- Not CALIBRATE. LEVEL checks one claim before one decision. CALIBRATE, from The Calibrated Mind, checks over many decisions whether your confidence matches what actually happens. What calibration means explains that second idea.
If you want to keep LEVEL records over time, they fit naturally into an evidence ledger: a list with one row per claim, its source, and what would retire it.
The LEVEL card
Copy this into a note, a document or the back of an envelope.
LEVEL card Date:
Locate The claim, in one checkable sentence:
Estimate Cost if I act and it is wrong:
Cost if I ignore it and it is right:
Cost of waiting / can I undo it?
Value Evidence I actually hold (which rung?):
Quality, independence, relevance:
Expose The missing piece that could change my mind:
Can I get it in time?
Lean Next step: Owner:
Limit: Review date or trigger:Try this today (15 minutes)
Pick one claim you are about to act on this week: a supplier’s promise, a statistic in a slide deck, a tip a colleague swears by. Fill in the card.
If you cannot write the Locate line in one sentence, stop there. That is your finding, and it is worth more than the other four lines. If your Lean line says “test”, put the review date in your calendar before you close the note.
Cite this:What is LEVEL? A five-step check before you act on a claim.Len P. van der Hof. https://lenvanderhof.com/en/blog/what-is-level/ ·