Monday, July 4, 2011

Incentives from the Group Crowd Out Joy

The "overjustification effect" is the cognitive process whereby getting a reward for doing something makes people do it less.

In experiments demonstrating the overjustification effect (such as the classic felt-tip marker study), the rewards were never (that I know of) things the subjects found or discovered themselves, but were presented by the researchers. I emphasize: the rewards were perceived by subjects to come from people.

The process of overjustification is often explained by the idea that a reward acts to shift salience from the inherent enjoyability of the activity to the external-reward-generating aspect of the activity. In short, with the introduction of an incentive, fun becomes work.

Why should this be? Why should the shift to "reward" thinking make the activity less fun?

There is a perfectly rational reason: because the reward is a signal given by interested others, who are one's competitors.

When one receives a promise of an incentive for some behavior, one might rationally ask the promissor, "what's in it for you?" Of course, words can lie, but costly incentives cannot. By providing a valuable incentive (or the reliable promise of one) to someone, the promissor assures the recipient in very credible language that the action of the recipient is valuable to the promissor, and in fact he is willing to pay for it. The recipient should rationally slow down and think about whether, instead of wasting his time doing the activity for free, he should be getting compensated for it - and, relatedly, whether this activity has a cost to him that he hadn't perceived.

One way for this abstract strategy to be implemented would be to cause the actor to feel less joy in an activity another indicated a willingness to pay to promote.

This strategy also applies to disincentives. From a group's perspective, why waste resources prohibiting an activity unless individuals have some reason to engage in the activity? A group prohibition is a reliable signal to individuals that there is some individual benefit - an advantage over the group - to be gained from violating the prohibition. One way to instantiate this strategy would be to build into your creatures curiosity about forbidden things and a little thrill from rule-violation.

The overjustification effect is our rational, built-in stubbornness. We recognize that incentives are signals, and immediately adapt to best take advantage of those signals. Joy is the casualty. The thrill of illicitness is the consolation.


  1. If you are interested in the issue of competition and competitive rewards, I would recommend Alfie Kohn's "No Contest: The Case Against Competition." Someone recommended it to me and I couldn't be more grateful.

  2. The considerations adduced here make me wonder if people undervalue things being given away for free. (The giving away could be anything, whether physical stuff or something intangible, like art published under Creative Commons licenses.) After all, if someone is giving something away, rather than trying to get you pay for it in some way, doesn't that seem like a likely signal that whatever it is isn't worth much?

    If so, sad.

    1. Conversely, it should be possible to create joy from false value by introducing artificial scarcity that can be overcome with effort - many computer games use this principle to give value to virtual goods.

  3. Yes, I believe that's a strong effect. And giving something away for free does function, to some degree, as a signal that it's not of high value.

    In terms of animal training, if you reward an animal every time it exhibits a behavior, it will only exhibit the behavior when it wants the reward. If you reward unpredictably, the animal will exhibit the behavior more consistently.

    A major part of our project as social organisms is figuring out what our behaviors are worth to others. Any signal relevant to this project will be taken into consideration by savvy organisms. This may be a signal of a particular magnitude of value (even zero), or a signal that something can be measured in a particular type of "market" or "currency" (as with Richard Titmuss' work on the American versus British blood supply).


Tweets by @TheViewFromHell