What is the framing effect?
A framing effect occurs when the presentation of a decision changes preferences even though the underlying outcomes are equivalent. Amos Tversky and Daniel Kahneman’s 1981 Science paper is a landmark demonstration. It studies decisions under risk; it is not evidence that every positive headline outperforms every negative one. Read the original paper.
The idea at a glance
The experiment behind the headline
In a hypothetical disease problem involving 600 people, 72% of a group of 152 chose a programme described as certainly saving 200 people. In a separate group of 155, only 22% chose the equivalent option described as 400 people dying; 78% selected the risky alternative. The problem held the possible consequences equivalent while changing the description. 1981 study, problems 1 and 2.
These are classroom choice percentages, not advertising uplift. The participants did not buy a service, and the study did not test a local business landing page. Its practical value is a question to investigate: does your wording help people understand the same offer differently?
Framing is not permission to change the facts
Compare “8 of 10 jobs met the agreed deadline” with “2 of 10 jobs missed the agreed deadline.” Both describe the same hypothetical record. Compare either with “We always deliver on time,” and the claim has changed. That is no longer a framing experiment.
Numbers need their denominator, timeframe and definition. “Save 20%” requires a meaningful comparison price. “Only $33.30 a day” for a $999, 30-day offer should not obscure that the actual purchase is $999. Showing an equivalent unit cost can explain a price, but it should not replace the total.
When the frame becomes misleading
A reader may reasonably assume that a highlighted benefit is typical, certain or included. Check those assumptions explicitly. A possible outcome should not look like a guarantee; a relative percentage should not hide a tiny absolute difference; a selected testimonial should not stand in for representative evidence.
The original paper itself describes its decision model as approximate and incomplete. Business audiences bring context, experience and priorities that a short experimental problem does not reproduce. The responsible next step is to test comprehension and behaviour in your own setting—not to declare people predictable.
How to apply it in your business
This is an original Maqtrix workflow to test. It is not a procedure proven by the cited study or a claim about results we have delivered.
1. Write the neutral factual version
List the actual price, period, deliverables, eligibility and uncertainty without sales language. This becomes the reference document. Every variation must remain consistent with it, including footnotes and the enquiry email.
2. Name the customer’s decision
Are they comparing suppliers, deciding whether a problem matters or choosing a first step? A statement about saving time may answer a different question from a statement about reducing risk. Start with the decision you need to explain.
3. Create two equivalent versions
Emphasise the benefit in one version and the avoided inconvenience in the other. For a documented workflow, “keep your enquiries organised” and “reduce missed follow-ups” may be two angles. Do not add an unsupported result to make the second version stronger.
4. Check equivalence line by line
Keep the service, price, evidence, deadline and call to action the same. If one version promises extra work or removes a condition, it is a different offer. Ask someone unfamiliar with the project to explain what they believe each version guarantees.
5. Make the relevant tradeoff visible
Present limitations near the claim rather than in an obscure footer. If results depend on customer input or advertising budget, say so. A reader should not have to contact sales to learn a fact that materially changes the purchase decision.
6. Run a controlled comparison
Keep the audience and destination consistent. Record exposure to each version and a preselected outcome. Avoid changing the image, price and headline simultaneously if the question is specifically about framing.
7. Choose the version that attracts the right customers
A fear-led headline may earn clicks while producing poor-fit enquiries. Review the enquiry content, sales conversations and later satisfaction. Prefer the message that customers accurately understand and that your delivery can support.
A practical service-business example
Hypothetical example: a business has a documented process that sends a weekly progress summary. Version A says, “See your next steps every week.” Version B says, “Stop wondering what happens next.” Both should lead to the same explanation of the weekly summary. Neither justifies claiming that revenue will increase.
For Maqtrix’s starting offer, a practical frame is “Use 30 days to identify the next marketing priorities.” The price remains CAD $999 with advertising spend included and no ongoing commitment. Explain the agreed work and review process; do not imply that every channel will be fully rebuilt in that period.
What to measure
| Measure | What to record |
|---|---|
| Comprehension | Can readers accurately state the deliverables, price and limitations? |
| Qualified action | Relevant enquiries per eligible visitor, not clicks alone. |
| Expectation mismatch | Enquiries or complaints based on something the offer does not include. |
| Later fit | Proposal acceptance and satisfaction for comparable customers. |
Define the outcome and observation period before launching. Where volume allows, compare randomly assigned, comparable groups and keep other changes limited. With a small customer base, begin with a usability pilot and report the actual counts. A before-and-after change is useful to investigate, but it does not by itself prove causation.
Use the first-party measurement guide to plan the records, and the campaign brief template to document the test.
Common questions
Does negative framing always work better?
No. The cited experiment concerns risky choices in a particular problem. It does not establish a universal marketing winner.
Is framing the same as lying?
A frame can highlight one truthful aspect. It becomes misleading when it conceals material information or encourages an unsupported interpretation.
Can a tiny sample tell us which headline wins?
It can reveal confusion and provide a direction for further testing. A few extra clicks do not establish a stable effect.
Original sources and research notes
Researched October 7, 2026. Historical claims and findings are linked to the sources below. Company accounts are labelled as such. Business workflows, examples and interpretations are Maqtrix’s proposals, not independently established performance outcomes.
- Tversky & Kahneman (1981), The Framing of Decisions and the Psychology of ChoiceOriginal Science paper; the two disease-problem groups and their choices.
Named brands and researchers are discussed for educational analysis. No affiliation or endorsement is implied.
Continue exploring
Turn the idea into a useful test
Choose one customer decision you want to make clearer. Email us with your business, the current problem and the change you are considering.
Start an email enquiry →