Loading Pledged…
Research Notes
These sources inform product hypotheses. None of them proves that Pledged works, predicts an individual outcome, or justifies an unaffordable consequence.
Last editorial review
2 August 2026
Linked sources
10 selected papers
Published Pledged efficacy data
None yet
Research on smoking cessation, weight, exercise, or clinical populations should not be repackaged as a general Pledged success claim. Pledged is not medical treatment, and consequences should never replace professional care.
Theme 1
These papers examine how advance constraints or incentives affected specific choices. They do not establish a universal stake amount or a Pledged success rate.
Kahneman & Tversky
A foundational model of decisions under risk. It supports discussing asymmetric valuation of gains and losses, not a fixed 'losses hurt 2×' rule for every context.
Open sourceGiné, Karlan & Zinman
A deposit-contract study in a particular smoking-cessation setting. Clinical and population context matters; this is not evidence for generic habit completion.
Open sourceRoyer, Stehr & Sydnor
A workplace exercise intervention combining incentives and commitment options. Its design and follow-up are relevant context, not a direct test of Pledged.
Open sourceHalpern et al.
Acceptance differed sharply between reward and deposit designs. That tradeoff is important: efficacy among accepters and population-level usefulness are not the same question.
Open sourceTheme 2
Planning and repetition research can inform rule design. It does not identify one universal timeline or prove that external enforcement creates automatic behavior.
Gollwitzer & Sheeran
Reviews if-then planning across studies. It supports making cues and responses specific while leaving effect size and applicability tied to task and study quality.
Open sourceWood & Neal
A theoretical account of habits and recurring contexts. It does not show that a check-in product can make a behavior effortless.
Open sourceLally et al.
Observed substantial variation in automaticity trajectories. The often-repeated 66-day figure is a sample median, not a countdown that applies to every behavior or person.
Open sourceOyserman, Fryberg & Yoder
Examines identity accessibility and health-related behavior in defined study contexts. It does not support a universal compliance percentage for identity language.
Open sourceTheme 3
The replication record matters. Pledged should not market the classic 'finite willpower tank' as settled science.
Hagger et al.
A 23-lab preregistered project found an effect estimate close to zero for the tested protocol. It challenged a simple, universal depletion account.
Open sourceVohs et al.
A later multisite project also did not find the preregistered confirmatory effect. Self-control remains an active research area, not a product guarantee.
Open sourceA longer bibliography is not stronger evidence. Traceability, scope, and honest uncertainty matter more.
As real product data becomes available, results should be reported separately from this literature review—with methods, caveats, and corrections visible.