Y12W36WR The honest experiment on honesty

Evidence Mapping
The writing prompt

Map what the research on honesty actually establishes about human behaviour, distinguishing robust findings from overextensions and fabrications.

1Retrieval check

Q1.What did the Cohn-Maréchal 2019 wallet study find?

  • AWallets with more money were kept at higher rates
  • BWallets with more money were returned at higher rates — contradicting standard economic prediction
  • CMoney had no effect on return rates
  • DThe study was not replicated

Q2.What happened to Ariely’s honesty research in 2021?

  • AIt was replicated and extended
  • BData-fabrication findings damaged the body of work; specific magnitudes from that tradition should be held with care
  • CIt won a Nobel Prize
  • DIt was republished without change
Show answer key

Q1 → B. Wallets with more money were returned at higher rates — contradicting standard economic prediction.The pattern held in 38 of 40 countries, suggesting an identity-protective mechanism.

Q2 → B. Data-fabrication findings damaged the body of work; specific magnitudes from that tradition should be held with care.The fabrication findings are why mapping robust from fragile here matters.

2Prompt deconstruction

Stimulus
Cohn-Maréchal wallet study; Ariely fabrication.
Scope
Map what the field actually establishes.
Method
Robust / mid-tier / walked-back / open categories.
Thinking
Investigation, not defence of a position; specific findings placed in specific categories.
Output
A map of the field + what the body of evidence actually supports.

3Pick nudge

Which honesty claims will you sort by how well the evidence holds?

Robust
Findings that replicate across studies and countries.
Mid-tier / theoretical
Supported theory (e.g. identity-protective) with more work needed.
Walked-back / fabricated
Ariely’s compromised findings; overextended magnitudes.
Open questions
Where the evidence is genuinely incomplete.

4Planner — categorise the claims

Robustly supported
Cohn-Maréchal 2019: higher-value wallets returned at higher rates, consistent across 38 of 40 countries.
Robustly supported (weaker form)
Cross-national variation in civic honesty correlates with specific institutional quality.
Mid-tier (supported theory)
Identity-protective theory of honesty — people maintain honesty because dishonesty threatens self-concept.
Walked-back / compromised
Specific magnitudes from Ariely-tradition work; any effects relying on the fabricated 2012 ‘signed-at-top’ study.
Overextended popular claims
Confident single-number claims about how much cheating happens in any specific domain.
Genuinely open
Which contextual levers (social norms, identity cues, surveillance) produce what size of honesty effect in what populations.

5Sentence stems

  • The claim that ___ is robustly supported, because ___.
  • The claim that ___ replicates only partially — specifically, when ___.
  • The popular version of ___ has been walked back; the careful version is ___.
  • The genuinely open question is ___.
  • A study that would resolve this would ___.
  • On the weight of evidence, the article’s own position is ___.

6Exemplar paragraph (not about this article)

(1) The claim that wallets with more money are returned at higher rates is robustly supported, because the Cohn-Maréchal 2019 study found it in 38 of 40 countries with a large sample. (2) The claim that identity-protective mechanisms explain the pattern replicates only partially — the theory is well-motivated and consistent with the data, but alternative explanations (social norms, reputational concerns) have not been cleanly separated. (3) The popular claim that ‘signing a pledge at the top of a form increases honesty’ has been walked back, specifically because the underlying data in the 2012 Shu-Mazar-Gino-Ariely paper shows fabrication signatures identified in 2021. The careful version is that cues to identity can affect honesty modestly; the fabricated magnitudes should be set aside. (4) The genuinely open question is which specific contextual levers produce meaningful honesty shifts in which populations. (5) A study that would resolve this would pre-register a suite of interventions with tight magnitudes, published regardless of outcome. (6) On the weight of evidence, the field’s own position — that cross-cultural honesty is higher than standard economic theory predicts, that the identity-protective theory is the best current explanation, and that specific magnitudes from the compromised literature should be held loosely — tracks the evidence.

What this paragraph does, move by move

  1. Places Cohn-Maréchal in ‘robust’ with the sample size named.
  2. Uses ‘replicates only partially’ for the identity-protective theory (supported but not cleanly separated from alternatives).
  3. Names the specific fabricated study (Shu-Mazar-Gino-Ariely 2012) rather than waving at ‘Ariely research’.
  4. Identifies the genuinely open question (contextual levers) specifically.
  5. Specifies what a resolving study would look like.
  6. Ends with a characterisation of the field’s own balanced position.