Boosting
Building people's own competence to decide well, rather than steering their choice from the outside.
What it means
Boosting is an intervention approach that aims to foster lasting competences — skills, knowledge, and decision tools people can use on their own — so they make better choices by their own lights, in contrast to nudges that arrange the environment to steer behavior. Examples include teaching simple decision rules, statistical literacy for reading risk, and rules of thumb for spending or eating. Its proponents argue it respects autonomy and is durable because the capability stays with the person even when the choice architect is gone, whereas a nudge's effect typically vanishes if the nudge is removed. The trade-offs: boosts demand more motivation, time, and cognitive engagement than nudges, and they can fail for people unwilling to learn. It matters because it reframes the goal of behavioral policy from compliance to empowerment, and offers a counterweight to 'commercial nudging' that boosted citizens can resist.
Two views of the mind
Boosting and nudging grow from opposed readings of the mind. Nudging descends from the heuristics-and-biases program, which treats intuition as a source of systematic error to be worked around. Boosting descends from the simple-heuristics, or ecological-rationality, program, which treats those same intuitions as adaptive tools that fit their environments and can be trained. The difference sets what each intervention targets: a nudge changes the environment and leaves the person unchanged, while a boost changes the person and can leave the environment alone. That is why boosters call cognition educable rather than defective, and often read biases as symptoms of bad information formats rather than fixed flaws of reasoning.
What counts as a boost
Boosts come in families. Cognitive boosts hand people a reusable procedure: translating percentages into natural frequencies, a fast-and-frugal decision tree, a checklist for spotting misinformation. Self-control boosts help people enforce their own goals, such as implementation intentions or 'self-nudges' that redesign one's own kitchen or phone. Some install quickly; others, like statistical literacy, take sustained teaching. What matters is not the effort a boost demands but where the competence ends up: with the person, portable across contexts, rather than baked into a choice architecture they do not control. That is what separates a boost from an 'educative nudge,' which informs without necessarily leaving a durable skill behind.
What the evidence shows
The best-evidenced boost is risk literacy. A meta-analysis of natural-frequency formats found they raise correct Bayesian answers sharply, but from a very low base, with most people still failing — the tool helps without being a cure. Teaching studies show representation training produces gains that survive over weeks and transfer to new problems. Head-to-head field tests are rarer: one seven-month residential trial found a boost package cut energy use more than a matched nudge package, though in a single small setting. Overall the boost literature is younger and thinner than the nudge literature, with fewer large-scale replications, so confident claims about its comparative power run ahead of the evidence.
Limits and critiques
Boosting inherits problems it is sometimes said to escape. A choice architect still decides which competence counts as 'good' and how to frame it, so the autonomy gain is partial; critics argue the nudge–boost line does not by itself carry the normative weight often claimed for it. Boosts also demand attention, motivation, and frequently numeracy, which invites a targeting problem: those who would benefit most are often least likely to engage, so a poorly designed boost can widen the gap it meant to close. In practice the categories blur, since many interventions inform and steer at once, making boost and nudge endpoints of a continuum rather than a clean binary.
Choosing a boost over a nudge
The approaches suit different problems. Prefer a boost when a decision recurs, a transferable skill exists, people are motivated to learn, and preserving autonomy is itself a goal — reading medical risk, managing money, judging sources. Prefer a nudge when a decision is one-off, attention is scarce, or the population is unmotivated, because a boost no one engages with changes nothing. The two also combine: a nudge can carry someone through today's choice while a boost builds the competence for next time. The design test is simple — will the skill still work once the designer, the app, or the program is gone?
Examples
Teaching patients to translate '10% in 10 years' into natural frequencies (10 of 100 people) is a boost — a transferable skill — whereas pre-selecting their treatment would be a nudge.
Teaching people to open a second tab and check who is behind a website is a boost; an algorithm quietly demoting that site for them is a nudge, and it leaves when the algorithm does.
A course that teaches 'move savings on payday, before anything else' leaves people able to save at any bank; an app's automatic transfer stops working the day they switch apps.
Teaching shoppers to compare the price per 100 grams, not per package, is a boost they carry to any store; moving cheaper staples to eye level is a nudge that ends at the checkout.
Training interviewers to score each candidate against fixed criteria before any discussion is a boost against halo effects; software that hides names on résumés is a nudge that only works while it runs.
First described in Ralph Hertwig & Till Grüne-Yanoff (2017).
Key references
- Herzog, S. M., & Hertwig, R. (2025). Boosting: Empowering citizens with behavioral science. Annual Review of Psychology, 76, 851–881. doi.org/10.1146/annurev-psych-020924-124753
- Paunov, Y., & Grüne-Yanoff, T. (2023). Boosting vs. nudging sustainable energy consumption: A long-term comparative field test in a residential context. Behavioural Public Policy, 1–26. www.cambridge.org/core/journals/behavioural-public-policy/article/boosting-vs-nudging-sustainable-energy-consumption-a-longterm-comparative-field-test-in-a-residential-context/39875649817B5B6AF64DB77E6FC3EBDE
- Sims, A., & Müller, T. M. (2019). Nudge versus boost: A distinction without a normative difference. Economics & Philosophy, 35(2), 195–222. www.cambridge.org/core/journals/economics-and-philosophy/article/abs/nudge-versus-boost-a-distinction-without-a-normative-difference/5AF46AB56B403452F43FEC7F32608A05
- Hertwig, R. (2017). When to consider boosting: Some rules for policy-makers. Behavioural Public Policy, 1(2), 143–161. www.cambridge.org/core/journals/behavioural-public-policy/article/abs/when-to-consider-boosting-some-rules-for-policymakers/047550D639F89EEB137FE61BA7C09DEF
- Hertwig, R., & Grüne-Yanoff, T. (2017). Nudging and boosting: Steering or empowering good decisions. Perspectives on Psychological Science, 12(6), 973–986. doi.org/10.1177/1745691617702496
- McDowell, M., & Jacobs, P. (2017). Meta-analysis of the effect of natural frequencies on Bayesian reasoning. Psychological Bulletin, 143(12), 1273–1312. doi.org/10.1037/bul0000126