Wednesday 2 November 2011

What a waste

Norwegian experimental economists ran some social trust experiments on a sample of prisoners. In prison.

You'd think the first thing they'd do would be to try the Prisoner's Dilemma, right?

Alas, they run a dictator game and some variants on a trust game.

Is there no poetry left in the world?

HT: Chris Dillow

Oh, they found that prisoners weren't much different from non-prisoners. But what a let-down. I'd have wanted to see two treatments on the Prisoner's Dilemma game. One with neutral framing, the other with the complete "Two prisoners, separated..." framing.

1 comment:

  1. Not quite what you're looking for, but the post reminded me of the following abstract:

    A new inmate, or `rookie', who enters a total institution usually faces `tests' and `games' organized by the `old crew'. In this article I argue that such initiation rituals are often designed by inmates in order to uncover a rookie's personal characteristics, such as toughness and cleverness. While such rituals may seem violent, they usually involve more skillful deception and tricks than pain and suffering. The basic idea is to persuade the rookie that he or she faces some tough choices - and watch his or her reaction to adverse or unusual circumstances. The mock character of a typical test creates a fundamental problem for its validity, since an informed rookie can simulate both toughness and cleverness. Thus, an informed rookie cannot be distinguished from one with the desired characteristics. This problem is well recognized by most knowledgeable inmates and motivates them to search for new games and tests. The result is a wide variety of competing tests, frequent changes of argot and the secret code of behavior. The initiation rituals are modeled as simple,games and decision problems. The ethnographic material was collected by the author as a political prisoner in Poland in 1985.

    Source: http://www.mendeley.com/research/games-prisoners-play-allocation-social-roles-total-institution-14/

    ReplyDelete