Here's a thought experiment from 1981 that changed how we understand human decision-making.
Imagine an unusual disease is expected to kill 600 people. Two programs are proposed:
Group 1 was told:
- Program A: 200 people will be saved
- Program B: 1/3 chance all 600 saved, 2/3 chance nobody saved
Result: 72% chose Program A (the safe bet).
Group 2 was told:
- Program C: 400 people will die
- Program D: 1/3 chance nobody dies, 2/3 chance all 600 die
Result: 78% chose Program D (the gamble).
Programs A and C are mathematically identical. So are B and D. The only difference is the words "saved" vs. "die."
That's Tversky and Kahneman's framing effect. And 45 years of research — including the largest replication study in psychology — confirms it's one of the most robust findings in behavioral science.
The Numbers: Four Meta-Analyses Agree
The framing effect has been meta-analyzed repeatedly. Here's what each found:
| Meta-Analysis | Year | Studies | N | Effect Size (d) | |---|---|---|---|---| | Kuhberger | 1998 | 136 | ~30,000 | 0.31 | | Pinon & Gambara | 2005 | 51 | ~13,500 | 0.26-0.44 | | Steiger & Kuhberger (p-curve corrected) | 2018 | 136 | ~30,000 | 0.52 | | Nabi & Walter (emotion-mediated) | 2020 | 25 | 5,772 | 0.22-0.31 |
The 2018 reanalysis is important. Steiger and Kuhberger applied p-curve analysis — a method that detects whether studies show signs of being manipulated — and found no evidence of p-hacking. The corrected effect size actually went UP from d = 0.31 to d = 0.52.
Their conclusion: "Risky-choice framing effects are highly reliable and robust: No replicability crisis there."
The Replications: 102,830 People Across 49 Countries
The framing effect isn't just robust in lab settings. It replicates globally.
Many Labs 1 (Klein et al., 2014): 36 independent samples, 10 countries, N = 6,344. Replication effect size: d = 0.60. Cross-site heterogeneity: I² = 0.0001 — virtually zero. The effect is the same whether you test it in Kansas or Korea.
COVIDiSTRESS Survey (Im & Chen, 2022): 49 countries, N = 102,830. Overall effect: Cohen's h = 0.612. The framing effect replicated in 100% of countries tested. Societal collectivism moderated the size but couldn't eliminate it.
Prospect Theory Replication (Ruggeri et al., 2020, Nature Human Behaviour): 19 countries, 13 languages, N = 4,098. 94% of items replicated, 12 of 13 theoretical contrasts confirmed.
And when researchers tested whether the effect was just a linguistic artifact — people misunderstanding the word "saved" — they added the word "exactly" to eliminate ambiguity. The framing effect persisted at p < .001.
Three Types of Framing (They're Not the Same)
Levin, Schneider, and Gaeth published a critical paper in 1998 showing that "framing" is actually three different psychological phenomena operating through three different mechanisms.
1. Risky Choice Framing (d = 0.44): Framing the OPTIONS as gains or losses. This is the Asian Disease Problem. Gain frame → risk aversion. Loss frame → risk seeking.
2. Attribute Framing (d = 0.26): Framing a single ATTRIBUTE positively or negatively. "75% lean" beef rated significantly better than "25% fat" — identical product. The smallest effect, but it compounds across every feature description, every email subject line, every data point you present.
3. Goal Framing (d = 0.44): Framing the CONSEQUENCES of action or inaction. "You'll gain X by doing this" vs. "You'll lose X by not doing this." Loss/negative goal frames are MORE persuasive for promoting action.
The three types are independent — susceptibility to one doesn't predict susceptibility to another.
Even Physicians Fall For It
McNeil, Pauker, Sox, and Tversky (1982) presented patients and physicians with a choice between surgery and radiation for lung cancer.
- Survival frame ("90 of 100 survive"): 75% chose surgery
- Mortality frame ("10 of 100 die"): 58% chose surgery
A 17-percentage-point shift from logically identical information. Physicians — trained to think analytically about medical statistics — were equally susceptible as patients.
In medical contexts specifically, framing effect sizes range from d = 0.40 to d = 1.21 — substantially larger than in low-stakes laboratory settings.
Your Brain on Framing
De Martino et al. (2006, Science) put 20 participants in an fMRI scanner while they made framed financial decisions.
What they found:
- Bilateral amygdala activated when people made frame-consistent decisions (following the bias)
- Orbital and medial prefrontal cortex activated when people resisted framing (making "rational" choices)
- Anterior cingulate cortex lit up when decisions went against the frame — signaling cognitive conflict
16 of 20 participants were completely unaware of the biasing effect when debriefed.
This is a dual-system phenomenon: an emotional amygdala-based system drives frame-consistent decisions, while an analytical prefrontal system enables resistance. When the prefrontal cortex is impaired — by stress, burnout, sleep deprivation, or cognitive overload — the amygdala wins more often.
The Connection to Burnout and Business Decisions
Here's where this connects to everything else we've covered:
- Burnout impairs prefrontal function by g = -0.39 to -0.53 (Gavelin et al., 17 studies)
- Lower HRV correlates with reduced prefrontal control (Thayer's neurovisceral integration model)
- Impaired prefrontal → amygdala dominance → INCREASED susceptibility to framing
- Burned-out solopreneurs are MORE susceptible to how information is presented, not less
This creates an ethical responsibility. If your audience is cognitively vulnerable — and burned-out business owners are — framing should serve their genuine interest, not exploit their vulnerability.
What This Means for How You Present Your Business
Goal framing for action (d = 0.44): When you want someone to DO something (sign up, try a tool, build a system), frame what they'll LOSE by NOT acting. "Every week without a knowledge base costs you X hours in repeated answers" is more persuasive than "A knowledge base saves you X hours per week."
Attribute framing for evaluation (d = 0.26): When describing what your product does, frame attributes positively. "80% of customers find answers without contacting support" works better than "Support contacts reduced by 80%." Same fact, different evaluations.
Risk framing for trial decisions (d = 0.44): The trial-to-paid transition is a risky choice frame. The customer has built content, set up courses, invested time. Frame the decision as "keep what you've built" (loss avoidance) rather than "continue paying" (gain calculation). The endowment effect (Post #166, WTA/WTP = 3.28x) amplifies this.
The ethical guardrail: Frame honestly. If your product genuinely helps, loss framing is describing reality — they WILL lose the time and effort invested. If your product doesn't deliver, loss framing is manipulation. The frame should match the truth.
The Practical Reality
One A/B test found that changing a CTA from "Sign up for free" to "Trial for free" produced a 104% increase in trial starts. Same offer. Different frame.
McKinsey found that a 1% improvement in pricing yields an 11-12% increase in profits. Yet only 24% of SaaS companies regularly run pricing experiments.
The framing effect is probably the most under-leveraged finding in business psychology. Not because people don't know about it — but because applying it requires the uncomfortable admission that your customers aren't making purely rational decisions. And neither are you.
The Nervous System Connection
The framing effect operates through the same amygdala-prefrontal circuit that governs the stress response. When you're in a ventral vagal state (calm, connected, HRV up), your prefrontal cortex can evaluate frames analytically. When you're in sympathetic overdrive (stressed, rushed, HRV down), the amygdala processes the frame emotionally and you follow it without awareness.
This means: how you present information to stressed, overwhelmed solopreneurs matters MORE than how you present it to calm, deliberate decision-makers. The very people who need the most help are the most susceptible to how that help is framed.
Frame toward their genuine interest. The research says your words will matter.
---
Sources:
- [Tversky, A. & Kahneman, D. (1981). The Framing of Decisions and the Psychology of Choice. Science, 211(4481), 453-458.](https://sites.stat.columbia.edu/gelman/surveys.course/TverskyKahneman1981.pdf)
- [Steiger, A. & Kuhberger, A. (2018). A Meta-Analytic Re-Appraisal of the Framing Effect. Zeitschrift fur Psychologie, 226(1), 45-55.](https://econtent.hogrefe.com/doi/10.1027/2151-2604/a000321)
- [Levin, I.P., Schneider, S.L. & Gaeth, G.J. (1998). All Frames Are Not Created Equal. OBHDP, 76(2), 149-188.](https://worthylab.org/wp-content/uploads/2020/12/levinetal1998allframesarenotcreated_equal.pdf)
- [Klein, R.A. et al. (2014). Investigating variation in replicability: A "Many Labs" replication project. Social Psychology, 45(3), 142-152.](https://econtent.hogrefe.com/doi/10.1027/1864-9335/a000178)
- [De Martino, B. et al. (2006). Frames, Biases, and Rational Decision-Making in the Human Brain. Science, 313(5787), 684-687.](https://www.science.org/doi/10.1126/science.1128356)
- [Im, H. & Chen, C. (2022). Disease risk framing across 49 countries. Journal of Behavioral Decision Making, 35(2).](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0257151)
- [Kuhberger, A. (2023). A systematic review of risky-choice framing effects. EXCLI Journal, 22, 1012-1031.](https://pmc.ncbi.nlm.nih.gov/articles/PMC10620856/)