Concept
Honesty Priming (Signature Timing)
honesty self-concept forms
Asking someone to sign a declaration of honesty before they fill out a form (rather than after) produces more honest answers, because signing activates a brief sense of personal reputation and duty that fades by the time a signature is requested at the end.
A large field experiment on a car-insurance mileage survey moved the signature box from the bottom of the form to the top; respondents who signed first reported driving about 10% more miles than those who signed at the end — evidently being more truthful rather than under-reporting to minimize their premium (see BIT Make-It-Timely Studies).
Phil Barden cites a related, unquantified UK example: the self-assessment tax return moved its truthfulness/honesty declaration checkbox from the end of the form to the beginning, on the same logic — getting filers into an honest mindset before, rather than after, they've entered their figures.
Replication caveat (Episode 44). Michael Hallsworth reports that the original authors of this finding later co-authored a follow-up paper concluding they could not reliably replicate the signature-timing effect across other settings — a conclusion built partly on the Behavioural Insights Team's own real-world field trials, which didn't find the effect either. This is cited as a specific example of behavioral science's broader Replicability Crisis; treat the original 10%-more-honest finding as unsettled rather than confirmed.
Upgraded to outright fraud, not just non-replication (Episode 195). This is the same Dan Ariely-co-authored car-insurance mileage study cited above (see BIT Make-It-Timely Studies) — not merely a finding that failed to replicate, but one a 2021 statistical audit found could not be real: the reported mileage distribution was implausibly uniform (respondents were as likely to report 49,000 miles as a realistic 10,000), and the dataset contained duplicate rows disguised with added random noise. Treat the original 10%/2,428-mile finding as fabricated, not merely unreplicated.
Discussed in
- Episode 7 — 7- How the UK government nudges British people
- Episode 20 — 20- Debunking a marketing myth (referenced)
- Episode 44 — 44- Do Nudges Really Work (reports failed replication)
- Episode 195 — 195-the-lying-psychologist-who-fooled-the-world (reports the data was fabricated, not merely non-replicating)