HPC · Science Deep Dive 18 June 2026 · revised 15 October 2026

The Framing Effect: How Identical Information Changes Your Decisions.

Wording can sway a choice even when the options are mathematically the same. Brain studies show part of why, and tests show the pull can be cut. Here is what the science actually says, and what to do with it.

Section: Decisions · Cognition
Reading time: 22 min read
Sources: 25 · reviewed

01What Framing Is

How the Wording of a Choice Changes the Choice

The framing effect is a change in choice caused by wording alone. Give people the same information, presented as a gain or as a loss, and many will choose the other way. It matters to you because choices often reach you already framed, as with a doctor's odds, a fund's returns or a question on a ballot. The evidence also shows the pull can be weakened in some tasks.

Medical training did not prevent it in one classic test. Physicians, patients and students imagined they had lung cancer and chose between two treatments, surgery and radiation.^[3] One version of the questionnaire gave the results as how many patients would live, the other as how many would die, and nothing else changed. Physicians preferred surgery more often when the outcomes were described as survival rather than death.^[3] The pattern, choices that shift with how facts are put rather than what they say, has been found many times since.^[6]

The founding experiment shows the scale. In the Asian Disease Problem, people chose between two plans to fight an outbreak. When the safe plan was described by lives saved, most people took it. When the same plan was described by deaths, most turned it down.^[1] Same numbers, opposite majorities.

The history

The standard account of framing is prospect theory.^[2] Its core idea is simple: people judge outcomes not in absolute terms but against a reference point, a baseline such as what they have or expect. Whether an outcome feels like a gain or a loss against that point changes how it feels, how much it weighs and what gets chosen.^[2]

The two sides are not equal: in prospect theory, a loss weighs more than a gain of the same size, a tilt called loss aversion.^[2] On this account, switching from "lives saved" to "lives lost" moves the reference point, so the same option now reads as a loss.

The effect also shows up in the brain. In a small brain-imaging study, activity in an emotion-linked area tracked choices that followed the frame.^[4]

How solid is the effect? For risky choices, solid. A meta-analysis, which pools the results of many studies, combined the field's risky-choice experiments and corrected them for publication bias (the habit of printing only striking results). It found a medium-sized effect, and the correction raised the estimate rather than lowering it.^[5]^[6]

Framing has also held up through the replication crisis, the wave of failed repeats of famous studies. A preregistered project spanning 36 countries replicated a shopping-discount framing effect, a different task from the disease problem.^[9] A repeat of the Asian Disease Problem itself, run as a hypothetical-choice survey in many countries during the COVID-19 pandemic, found the effect no weaker.^[17]

This article traces the effect from brain studies to pooled data to a protocol. It asks two questions: if the frame is not the fact, what is the brain reacting to, and can you resist it?

02The Mechanism

The Neural Circuit That Tracks the Frame

Brain imaging gave the first clues to how framing works. In one small study, twenty participants played a gambling game inside a brain scanner. The game set up gain and loss frames while the scanner tracked brain activity.^[4] The results tied framing to activity in specific brain areas.

When participants made the choice the frame pushed them toward (playing safe under a gain frame, taking the risk under a loss frame), their amygdala, an area linked to threat and reward, was more active.^[4] The odds were the same in both frames; the brain response was not.

The key finding lay in the individual differences. Those with more activity in the ventromedial prefrontal cortex (vmPFC), a front-of-brain area that weighs value, and in the nearby orbital frontal cortex made choices that followed the frame less.^[4] One proposed model reads this as two systems: a fast feeling system that tags options, and a slower control system that can override those tags. The study shows associations; it did not test that model directly.

Imaging linked amygdala activity to choices that followed the frame, and prefrontal activity to resisting it. A later study linked the amygdala to framing from a gain start and the striatum to framing from a loss start.

Diagram · HPC

The circuit is not as simple as "amygdala bad, prefrontal good." The first scan study tied the amygdala to frame-driven choices in both frames.^[4] Later imaging splits the work by starting context: the amygdala tracked framing mainly when people started from a gain, and the striatum, the brain's reward-value hub, when they started from a loss.^[13] Behaviour looked alike in both contexts, so different brain areas may carry the same bias.

Genes add another layer: in a genetic study of more than a thousand Chinese Han people, variants in the COMT gene and the SLC6A4 gene, which shape dopamine and serotonin signals, were associated with how prone people were to framing.^[12] Earlier work found that people with two short copies (ss) of the serotonin transporter gene (5-HTTLPR) had a more reactive amygdala and more framing bias than people with the long form.^[8]

Prospect theory supplies the idea the scans are read against: the reference point, the baseline that the wording can move.^[2] Whether an option then reads as a gain or as a loss changes how it is weighed.

Time pressure is one condition that changes the effect. In one preregistered eye-tracking experiment, a short deadline amplified framing, mainly in gain-framed trials.^[15] Shifts in early attention and shorter information search offered an explanation.^[15]

The founding experiment's swing is no outlier. Repeat studies have confirmed its direction, though its size varies with the task.^[1]^[6]^[17]

Can people reduce it? Some experiments suggest so: making the same choice in a foreign language reduced framing, and emotional distance is one proposed explanation.^[7]

03Evidence

The Five Strongest Studies on Framing Effect Psychology

01The claim

The single load-bearing finding

The hero study finds 72 vs 22 % choosing safe option.

If a change of language can weaken the frame, the next question is how much weight each link in this story can bear. Five studies carry most of it, though they are not the only important papers in a field of thousands. Together they cover the effect's existence, size, brain links, reach into medical choices and openness to change. Ranking evidence is a judgment call that weighs design against sample size, clear cause against real-world fit, and repeat results against historical weight. Each study is scored out of 100 on six criteria, weighted toward sound method and listed in the table below.

Key study result

72 vs 22%

choosing safe option

02How we measured

Grading the framing-effect trials

Studies scored on design, sample, rigour, causality, replication, citations.

For framing research, sample breadth weighs heavily: the bias should hold across physicians, patients and other groups, not only in one laboratory task, before broad claims are made.

Rubric weights

Design/30

Sample/20

Rigour/15

Causality/15

Replication/10

Citations/10

03The spread

Heterogeneity across 5 studies

Methodological quality across the ranked studies.

The hierarchy shows what a list of single studies can hide: agreement on the core effect. Risky-choice framing is a medium-sized, reliable bias, linked to activity in specific brain areas and seen both in lab tasks and in hypothetical clinical choices put to physicians.^[6]^[3]^[4] The size of the effect varies by task and design.^[5]^[6] In variations of the classic disease problem, the type of disease changed the size of the effect, and, as in the eye-tracking test, short deadlines made it stronger.^[25] The effect remained; these factors changed how large it was.

Rubric spread

→ 78 /100

Highest to lowest rubric score across the ranked studies.

04Beyond the top five

Wider evidence

What the wider evidence adds.

One more distinction sharpens the picture. A classic typology splits framing into three types.^[10] Risky-choice framing is the classic design: a sure thing against a gamble. Attribute framing changes how a trait of an object is described. Goal framing stresses what you gain by acting or lose by not acting. Each type is thought to work by a different route.^[10] The proposed routes differ.

The studies

5 studies, ranked.

Below: the anchor study in full; then every study's rubric score; then the supporting studies in ranked order.

01Anchor

The founding experiment: same outcomes, opposite choices

Tversky & Kahneman 1981 Controlled Experiment · Between-Subjects · Replicated

The effect is not tied to one kind of content or setting. Choices flip in a regular way when the same outcomes are described another way.

It is the founding demonstration on which the rest of the argument builds. Its core finding has been repeated since, including in a survey in many countries during the pandemic.

Rubric breakdown

Design27/30

Sample14/20

Rigour13/15

Causality14/15

Replication10/10

Citations10/10

Total 88/100

The strongest studies, in the order of the argument.

Each scored 0–100 against a six-criterion rubric, tagged by design and year; the anchor leads.

Anchor (Rank 1) Supporting

Rank Authors & title Journal · Year Finding Score

Steiger & Kühberger

The pooled estimate, corrected for publication bias

2018

Using p-curve, a method that corrects for publication bias, the authors re-analysed the field's earlier database of 136 studies and about 30,000 people. The corrected estimate rose from d = 0.31 to d = 0.52: a medium-sized effect, larger than first measured. Effects vary by task and design.

87/100

De Martino, Kumaran & Seymour

The first brain scans of framing: an emotion and control circuit

2006

Amygdala activity was significantly linked to choices that followed the frame. Across people, more vmPFC activity went with choices that resisted it.

79/100

McNeil, Pauker & Sox

Doctors, patients and students choosing a treatment for imagined lung cancer

1982

Asked to imagine they had lung cancer, physicians preferred surgery more often when the results were framed as survival than when the same data were framed as death rates. Patients and students shifted in the same direction.

82/100

Keysar, Hayakawa & An

The foreign-language effect: less bias at a distance

2012

In one real-money betting experiment, acceptance averaged 54% in the native language and 71% in the foreign language, consistent with less loss aversion. Four separate experiments found less risky-choice framing.

78/100

04Stakes

The Invisible Tax on Decisions You Think You're Making Rationally

Framing reaches beyond lab gambles. In experiments and surveys, it shifted treatment preferences, retirement plans, claiming plans and consent to organ donation.

01 System 01 · Medical Decisions

Treatment Choices Shifted by Wording

In an imagined lung-cancer choice, physicians, patients and students leaned toward surgery when outcomes were framed as survival, and less so when the same data were framed as deaths.^[3] A separate drug-rating experiment found doctors and patients similarly swayed by framing.^[19]

In practice

choosing a treatment because of how the data were shown, not what the data said

02 System 02 · Financial Decisions

Retirement Plans Swayed by Narrow Framing

In one study, people who focused on current spending rather than future retirement spending (narrow framing) planned smaller increases in retirement contributions.^[23] In an experiment, "breakeven" wording led respondents to expect to claim Social Security earlier, especially those with lower financial literacy.^[24] Both results concern plans, not observed savings or claims.

In practice

planning to save less because of how the choice was framed

System 03 · Policy Choices

Organ Donation Consent Shaped by Defaults

In one online experiment, about 82% consented to organ donation under an opt-out default and about 42% under an opt-in default.^[22] The opt-out question was identical to the opt-in question except for its default.

In practice

going along with a choice because it was already made for you

04 System 04 · Individual Differences

Anxiety and Age: Mixed Evidence

Traits of the person choosing may matter as well. A neuroimaging commentary describes an association between trait anxiety, greater framing susceptibility and framing-related amygdala activity.^[14] Age results are mixed: an 18-study meta-analysis found no overall age difference for negative-framed risky choices.^[16] A separate study found greater susceptibility with age under negative wording.^[20] How age affects framing remains unsettled.

In practice

choices that feel heavier and more threatening than the data justify

05Protocol

A 4-Step Framing Debiasing Protocol

The science offers no magic bullet against framing. It points to a sequence of steps meant to make a frame easier to notice and to weigh. The sequence is a synthesis of the cited studies, untested as a whole.

The protocol, as a sequence.

Before Decision → During Analysis → Under Emotional Pressure → After Initial Assessment

01 Step 01 · Before Decision

Invert the Frame

Restate every important choice in the opposite frame before you act. If it is framed as a loss, rewrite it as a gain, and compare your preference in both versions.

Why

It may show the frame's pull before you commit. Choices shift when the frame flips,^[1] and in one small scan study, people with more prefrontal activity followed the frame less.^[4] Whether restating a choice engages that activity is not known.

Common mistake

Only doing the inversion in your head. Writing both versions down is a suggested way to compare them.

02 Step 02 · During Analysis

Convert to Natural Frequencies

Swap probabilities for counts. "30% chance of mortality" becomes "30 out of 100 people die." Keep the base rate and the counts within each subgroup visible. Ask for the absolute risk reduction (ARR), not the relative risk reduction (RRR).

Why

Counts that keep the base rate in view, called natural frequencies, improved reasoning on diagnosis-style problems in a set of experiments.^[11] That was shown for diagnosis, not framing itself. Counts may make the numbers easier to compare; reduced framing was not tested here.

Common mistake

Accepting relative risk figures from doctors, salespeople or news reports without asking for absolute numbers. In one drug-rating experiment, results given as a relative risk reduction made doctors and patients rate a drug as more effective than other formats did.^[19]

03 Step 03 · Under Emotional Pressure

Create Deliberate Distance

When loss-framed facts trigger strong emotion, ask: "What would I advise a close friend here?" Taking a third-person view is meant to create distance from the frame.

Why

Four experiments using a foreign language found less risky-choice framing.^[7] Emotional distance is one proposed explanation; whether a third-person view has the same effect is not known. Limits apply: these were specific tasks, and the effect is not a proven everyday tool.

Common mistake

Treating the emotional reaction as useful information rather than an effect of the wording. The urgency of a loss frame is real, but it may partly reflect the wording.

04 Step 04 · After Initial Assessment

Premortem the Frame

Assume the loss-framed outcome has already happened. Ask: what would I have done differently?

Why

The premortem is a planning exercise that imagines failure in advance. Here it is a suggestion: picturing the feared outcome as already present may loosen loss aversion's grip. A different aid has some data: one before-and-after study links training in statistics to less framing, though randomised trials are needed to confirm cause.^[18]

Common mistake

Confusing the premortem with pessimism. The goal is not to predict failure. The aim is to explore how the loss frame affects your preference; whether this exercise reduces framing is not known.

Operational logic

The protocol is short by choice. That is an editorial decision about debiasing, the effort to reduce a bias, not a tested finding. In two experiments, people's measured thinking styles showed no significant link with how often they followed the frame.^[21] The lead the protocol builds on is distance: a change of language reduced framing in the experiments above.^[7]

The protocol is not shown to remove framing, and the language experiments did not test everyday choices. Proneness also varies between people and is associated with genetic variation.^[12]^[8] The steps start with noticing the frame because each of them needs it: you cannot restate, count or step back from a frame you have not seen. Framing's most dangerous trait is that it can feel like a sensible response to the facts, not to their packaging.

06Verdict

The verdict.

"The frame is not a wrapper around a decision. It is part of the decision.", Editorial synthesis

Bottom line

The frame is not the fact. Without a habit that separates them, your brain may treat the frame as though it were.

The argument

The evidence points one way: in risky choices, framing is a medium-sized bias that shifts choices between identical outcomes based only on how they are described. Brain scans link it to specific areas. In tested tasks, it reached trained physicians as well as lay people, and it shows up in hypothetical medical, money and policy choices.

A medium effect across the pooled studies means this is no marginal lab quirk. It can act on choices put in terms of gains or losses, though its size varies with the task.

The bias can shift: in some controlled experiments, using a foreign language reduced it in specific tasks. The protocol's other steps, such as frame inversion and counts in place of percentages, build on related findings.

Marketers, politicians, doctors, algorithms and your own instincts can all frame your choices. Noticing the frame is the step this protocol starts from.

One reading of framing research is that judging equal options equally is not automatic, even for experts. In a hypothetical treatment task, physicians shifted with the wording, as patients and students did.^[3]

The whole argument, on one axis

Same numbers. Opposite majority.

01Claim

The Bias Is Reliable

In risky choices, framing is a reliable, medium-sized effect of wording, backed by experiments and pooled analyses. Brain scans link it to activity in specific areas.

02Consequence

Training Did Not Prevent It

Framing affected physicians, patients and students in a hypothetical treatment task.^[3] A separate drug-rating experiment found similar susceptibility in doctors and patients.^[19] The reach is broad: it shows up in treatment preferences, retirement plans and consent under different defaults.

03Lever

Distance Can Reduce It

In experiments, using a foreign language reduced framing in specific tasks; emotional distance is one proposed explanation.^[7] The protocol adds frame inversion, a third-person view and a premortem, which are editorial suggestions, and counts in place of percentages, a format studied in diagnosis problems.^[11] Its effect on framing as a whole is untested.

Put it to work

Where this science goes next on HPC

AssessmentCognitive Biases AssessmentScore where you actually stand Protocol90-Day Cognitive Debiasing ProtocolTurn the mechanism into practice GlossaryCognitive Behavioural TherapyThe definition, held to the evidence

07Bibliography

The bibliography.

25 sources · ~3h est. corpus read · 25 visible

RCT · 1 Meta · 3 Review · 1 Journal · 20

Type

Sort

01 Journal

Tversky, A., & Kahneman, D1981

The framing of decisions and the psychology of choice

Science211(4481) · 453–458

doi: 10.1126/science.7455683
a b c
02 Journal

Kahneman, D., & Tversky, A1979

Prospect theory: An analysis of decision under risk

Econometrica47(2) · 263–292

no doi
a b c d
03 Journal

McNeil, B. J., Pauker, S. G., Sox, H. C., & Tversky, A1982

On the elicitation of preferences for alternative therapies

New England Journal of Medicine306(21) · 1259–1262

doi: 10.1056/NEJM198205273062103
a b c d e f
04 Journal

De Martino, B., Kumaran, D., Seymour, B., & Dolan, R. J2006

Frames, biases, and rational decision-making in the human brain

Science313(5787) · 684–687

doi: 10.1126/science.1128356
a b c d e f g
05 Meta

Kühberger, A1998

The influence of framing on risky decisions: A meta-analysis

Organizational Behavior and Human Decision Processes75(1) · 23–55

doi: 10.1006/obhd.1998.2781
a b
06 Meta

Steiger, A., & Kühberger, A2018

A meta-analytic re-appraisal of the framing effect

Zeitschrift für Psychologie226(1) · 45–55

doi: 10.1027/2151-2604/a000321
a b c d e
07 Journal

Keysar, B., Hayakawa, S. L., & An, S. G2012

The foreign-language effect: Thinking in a foreign tongue reduces decision biases

Psychological Science23(6) · 661–668

doi: 10.1177/0956797611432178
a b c d
08 Journal

Roiser, J. P., de Martino, B., Tan, G. C. Y., Kumaran, D., Seymour, B., Wood, N. W., & Dolan, R. J2009

A genetically mediated bias in decision making driven by failure of amygdala control

Journal of Neuroscience29(18) · 5985–5991

doi: 10.1523/JNEUROSCI.0407-09.2009
a b
09 Journal

Klein, R. A., et al2018

Many Labs 2: Investigating variation in replicability across samples and settings

Advances in Methods and Practices in Psychological Science1(4) · 443–490

doi: 10.1177/2515245918810225
↩
10 Journal

Levin, I. P., Schneider, S. L., & Gaeth, G. J1998

All frames are not created equal: A typology and critical analysis of framing effects

Organizational Behavior and Human Decision Processes76(2) · 149–188

doi: 10.1006/obhd.1998.2804
a b
11 Review

Gigerenzer, G., & Hoffrage, U1995

How to improve Bayesian reasoning without instruction: Frequency formats

Psychological Review102(4) · 684–704

doi: 10.1037/0033-295X.102.4.684
a b
12 Journal

Gao, X., Liu, J., Gong, P., et al2017

Identifying new susceptibility genes on dopaminergic and serotonergic pathways for the framing effect in decision-making

Social Cognitive and Affective Neuroscience12(9) · 1534–1544

doi: 10.1093/scan/nsx062
a b
13 Journal

Sun, S., Hu, J., & Yu, R2022

Domain-specific neural substrates underlie the framing effect

NeuroImage: Reports2(4)

doi: 10.1016/j.ynirp.2022.100119
↩
14 Journal

Jepma, M., & López-Solà, M2014

Anxiety and framing effects on decision making: Insights from neuroimaging

Journal of Neuroscience34(10) · 3455–3456

doi: 10.1523/JNEUROSCI.5352-13.2014
↩
15 Journal

Roberts, I. D., Teoh, Y. Y., & Hutcherson, C. A2022

Time to pay attention? Information search explains amplified framing effects under time pressure

Psychological Science33(1) · 90–104

doi: 10.1177/09567976211026983
a b
16 Meta

Best, R., & Charness, N2015

Age differences in the effect of framing on risky choice: A meta-analysis

Psychology and Aging30(3) · 688–698

doi: 10.1037/a0039447
↩
17 Journal

Rachev, N. R., Han, H., Lacko, D., et al2021

Replicating the disease framing problem during the 2020 COVID-19 pandemic

PLOS ONE16(9)

doi: 10.1371/journal.pone.0257151
a b
18 Journal

Borracci, R. A., Arribalzaga, E. B., & Thierer, J2020

Training in statistical analysis reduces the framing effect among medical students and residents in Argentina

Journal of Educational Evaluation for Health Professions

doi: 10.3352/jeehp.2020.17.25
↩
19 RCT

Perneger, T. V., & Agoritsas, T2011

Doctors' and patients' susceptibility to framing bias: A randomized trial

Journal of General Internal Medicine26(12) · 1411–1417

doi: 10.1007/s11606-011-1810-x
a b c
20 Journal

Xu, L.2024

Advanced age predicts increased susceptibility to attribute, goal, and risky-choice framing in negative frame valences

Humanities and Social Sciences Communications11(1) · 217

doi: 10.1057/s41599-024-02658-6
↩
21 Journal

Mandel, D. R., & Kapler, I. V2018

Cognitive style and frame susceptibility in decision-making

Frontiers in Psychology

no doi
↩
22 Journal

Johnson, E. J., & Goldstein, D2003

Do defaults save lives? Science, 302(5649), 1338–1339

Science302(5649) · 1338–1339

doi: 10.1126/science.1091721
↩
23 Journal

Shin, S., Kim, H., & Heath, C. J2019

Narrow framing and retirement savings decisions

Journal of Consumer Affairs53(3) · 975–997

doi: 10.1111/joca.12211
↩
24 Journal

Brown, J. R., Kapteyn, A., & Mitchell, O. S2016

Framing and claiming: How information-framing affects expected Social Security claiming behavior

Journal of Risk and Insurance83(1) · 139–162

doi: 10.1111/j.1539-6975.2013.12004.x
↩
25 Journal

Author not identified in source2018

Moderators of framing effects in variations of the Asian Disease problem

13(6), 529–54613(6) · 529–546

no doi
↩

The protocol card