Why we think it's OK to cheat and steal (sometimes) | Dan Ariely

Main ideas, concepts, and lessons

• People are predictably irrational, even with good intentions and relevant experience.
• Intuition often fails because humans don’t process certain decision-making dimensions correctly (e.g., confusing duration with intensity of pain).
• Cheating is widespread and usually incremental: many people cheat “a little,” not only “bad apples.”
• Cheating is influenced by psychology and context, not just classic economic deterrence models (probability of getting caught and the size of rewards/punishments).
• Moral self-image (“personal fudge factor”) allows small cheating while preserving a positive view of oneself.
• Making morality salient reduces cheating, while increasing psychological distance from money increases cheating.
• Social norms matter: seeing an in-group member cheat can increase cheating; seeing an out-group member cheat can increase honesty.
• Behavioral economics and policy: because intuition is often wrong, decision-makers should rely on systematic experimentation rather than assumptions—especially for public policy (education, markets, taxation, healthcare).

---

Methodologies / experiments and what they showed

1) Burn-bandage pain experiment (inspired by Ariely’s personal experience)

• Initial real-world problem

  • Ariely was severely burned (70% of body).
  • Nurses removed bandages quickly and repeatedly (“rip it off quickly”).
  • Ariely suggested alternatives to reduce intensity (e.g., longer removal time to lower intensity).

• Key finding from later experimental research

  • Nurses were wrong in a predictably irrational way:
    • People don’t encode duration the same way they encode intensity.
  • The implication was that less pain would have come from:
    • longer duration with lower intensity (instead of short/high intensity)
    • changing the order of treatment (start with more painful areas like the face and progress toward less painful areas)
    • including breaks during treatment

• Lesson

  • Strong experience and intentions aren’t enough; the relevant cognitive/psychological mechanisms must be tested.

---

2) Cheating baseline experiment (core “how much do people cheat?” setup)

• Task

  • Participants receive a sheet with 20 solvable math problems.
  • They have 5 minutes, not enough for full completion.

• Conditions

  • Non-cheating control
    • After 5 minutes, participants return the sheet for payment of $1 per correct problem.
    • Average result: participants solved about 4 problems.
  • Cheating condition
    • Participants are instructed to shred the paper and put pieces in a pocket/backpack.
    • They report their own number of correct answers.
    • Average result: participants solved about 7 problems.

• Key results

  • Cheating is not concentrated only among “bad apples.”
  • Instead, many people cheat a little.

---

3) Testing economic deterrence predictions (reward size and probability of being caught)

• Reward manipulation

  • Payment per correct answer increases across groups (example values: 10 cents, 50 cents, $1, $5, $10).
  • Economic theory prediction: higher stakes → more cheating.
  • Observed result: people continued to cheat a little, largely insensitive to reward size.

• Probability-of-catch manipulation

  • Different shredding instructions/behaviors change how much evidence remains (examples):
    • shred half the sheet
    • shred whole sheet
    • fully shred everything and leave with money > $100
  • Economic theory prediction: lower probability of being caught → more cheating.
  • Observed result: cheating increased only a little, not strongly following deterrence logic.

• Conclusion

  • People don’t behave like they’re doing strict cost-benefit calculations as economic theory assumes.

---

4) “Personal fudge factor” theory and tests (reducing vs. increasing cheating)

A) Reducing cheating by increasing moral salience

• Ten Commandments recall intervention

  • Participants complete one of two recall tasks:
    • recall 10 books read in high school, or
    • recall The Ten Commandments
  • Observed result
    • Those prompted to recall The Ten Commandments did not cheat at all.
    • The effect appears tied to immediate moral salience, not broad religious identity.
    • Even self-declared atheists who were asked to swear on the Bible and given cheating opportunities did not cheat.

• Honor code signature intervention

  • Participants sign a statement acknowledging the MIT Honor Code (as framed in the study).
  • Observed result
    • No cheating occurs after signing.
  • A humorous framing in the video claims MIT doesn’t have an honor code (as stated by the speaker), emphasizing the impact of moral cues over formal enforcement.

B) Increasing cheating by increasing psychological distance from money

• Distance-from-cash real-world observation

  • Ariely contrasts:
    • 6-pack Cokes in student refrigerators (which disappear quickly)
    • a plate with six one-dollar bills (which do not disappear)
  • This shows people treat money differently from consumption-linked items.

• Controlled token vs. cash experiment

  • Participants do a task like before, then must shred a sheet and report correct answers.
  • Instead of paying in dollars, participants are paid in tokens/another medium.
  • They then walk about 12 feet to exchange tokens for dollars.
  • Observed result
    • Cheating doubles under the token condition.

• Interpretation

  • Psychological steps between the dishonest act and actual money reduce guilt, enabling more cheating.

---

5) Social norms and “in-group vs out-group” cheating experiment (group setting)

• Setup

  • A large group of students participates.
  • Everyone receives an envelope with the experiment’s full money in advance.
  • They’re later told they must pay back what they didn’t earn.

• Cheating baseline

  • With cheating opportunity, participants cheat a little bit (consistent with earlier findings).

• Norm manipulation using an acting participant

  • After 30 seconds, an acting student stands up and says:
    • “I solved everything. What do I do now?”
  • The experimenter replies:
    • If finished, go home (task ends).
  • Others infer that cheating successfully is possible.

• Key manipulation (in-group vs out-group)

  • The acting student wears different university sweatshirts:
    • Carnegie Mellon vs University of Pittsburgh (in a Pittsburgh setting; participants are Carnegie Mellon students).

• Observed result

  • If the actor is from the in-group (Carnegie Mellon sweatshirt): cheating goes up.
  • If the actor is from the out-group (Pittsburgh sweatshirt): cheating goes down.

• Conclusion

  • The effect is framed as norms, not deterrence:
    • in-group cheating feels more acceptable to the group
    • out-group cheating increases awareness of honesty expectations

---

Overarching lessons conveyed

• Cheating is “bounded”: many people cheat slightly but don’t escalate indefinitely, suggesting constraints tied to self-image and moral identity.
• Moral reminders can strongly change behavior (commandments/honor code cues).
• Distancing from money (tokens, steps removed, longer psychological gaps) increases dishonesty.
• Social identity and group norms can either amplify or suppress cheating.
• Policy and business should test intuitions because subjective certainty doesn’t mean correctness.

---

Speakers / sources featured

• Dan Ariely (speaker; researcher discussing experiments and behavioral economics)
• Enron (used as a real-world context example motivating the cheating inquiry)
• MIT (mentioned in relation to the “MIT Honor Code” cue experiment)
• Carnegie Mellon University (participant group / test site)
• University of Pittsburgh (out-group sweatshirt condition / comparative group)
• The Ten Commandments (used as a psychological salience cue)
• Bible (referenced in an experiment where even self-declared atheists did not cheat)