The Cookie Experiment

Three people working together. A plate of five cookies available on the break. Social norms dictate that no one take the last (fifth) cookie, but what about the fourth?


Such was the inquiry of the “cookie experiment” reported by Keltner, Gruenfeld & Anderson (2000). Groups of individuals were placed into teams of three, with two responsible for creating a social policy paper and one randomly assigned to evaluate the output.  After  30 minutes, “the experimenter arrived with a plate of five cookies. This allowed each participant to take one cookie, and at least one participant to comfortably take a second cookie, thus leaving one cookie on the plate.” No one was expected to take the fifth cookie and no one did. But what about the fourth, the one that could be taken without awkwardness, or negotiation? The results of the study showed that the “boss” (the evaluator) was clearly more likely to take a second cookie.  (Videotapes of the interactions also showed them more likely to eat with their mouths open and scattering crumbs widely.) The conclusion is that individuals in power are more “disinhibited” and prone to focus on their own needs to the exclusion of others.


The authors proceed to identify some candidate behaviors that are more “prosocial” in nature, with less focus on self and more on others. One of those, expressing approval,  led me to think about ways in which to improve and be more effective in this area.

  1. Take the time to point out areas of positive contribution of employees. Don’t assume they “just know” they are appreciated.
  2. Be specific, demonstrating authentic knowledge of the contribution. (In large organizations you often see congratulatory emails work their way up the chain to increasing levels of abstraction until you see a final reply all of the nature, “I want to thank each and every one of you for your efforts in this endeavor.” Not exactly Hoosiers-like inspiration.)
  3. Give them some insight as to how this benefits the company, especially beyond the obvious. Ideally everyone has the vision of how what they are doing benefits the company. Ideally.
  4. Resist the temptation to temper it with some constructive criticism, as if their performance might degrade due to the poor aerodynamic properties of an engorged head. Authentic (and rigorous) feedback about improvement areas is vital, but it does not have to permeate every conversation. Share five large positive things and one minor area of criticism with someone and what will he remember? Just the one negative thing. Save it for another conversation.
  5. Don’t fish for return compliments. It should be more sincere than simply a bartered transaction at an accolade flea market.

Simple things requiring little investment, yet with potentially high returns in terms of motivation. And who knows, if you do it enough you might start to notice there’s a few less crumbs on your shirt at the end of the day. :)

“I have yet to find the man, however exalted his station, who did not do better work and put forth greater effort under a spirit of approval than under a spirit of criticism.” – Charles Schwab


  • […] psychologist and Stanford professor Deborah Gruenfeld conducted a research study with some of her colleagues that tested the effects of power on behavior. Here’s a description of […]

  • This study was mentioned in a recent Dan Ariely blog, which triggered people to go looking for the source, and some (including me) came here. For anyone who may in the future be looking for the source of the cookie experiment, it’s from an unpublished 1998 paper by Ward and Keltner. That paper is summarized in Ward, Keltner & Anderson, Power, Approach and Inhibition, Psychological Review, 2003, Vol 110, No. 2, 265-284, with the summary appearing at page 277. This latter paper can be downloaded from Keltner’s website,

  • This experiment is absolutely brilliant. I find the story useful in seminars on happiness at work where I deliver the point, that people in a position with power over others tend to feel they are more right and more entitled to have things their way. When this spurs people to take action to actually get things their way, it can diminish morale and reduce happiness at work. So the clear advice to bosses and everyone in positions of power and autority: Hold back, just a little – you’re not always right, you know.

    Thank You Scott for linking to the paper.

  • […] leadership behaviors. Here’s an example of a study that lead to the above conclusions, called the cookie experiment by Keltner, Gruenfeld & Anderson […]

Leave a Reply