Question

The Dictator Game

The strongest experimental evidence against the idea that people are only interested in what they can get for themselves comes from the dictator game.

The Rules:

In the game, two people interact anonymously. One of them is randomly designated as the “dictator.” It is his job to split an amount of money that is put up for that purpose by the researcher running the game. A typical amount is $10. The defining feature of the game is that the dictator can dictate whatever split he prefers. It could be to keep all the money for himself. It could be to give all the money to the other player. It could be to split it in any other possible way, such as $8.67 for himself and $1.33 for the other person.

Because the game is fully anonymous, the dictator doesn’t have to worry about retaliation by the other person. He can get away with being as selfish as he wants.

a. Do people playing the dictator game show only self-interested behavior?

b. How much divergence is there in the splits given by dictators to the other player?

So what actually happens when people play the dictator game? After running the experiment many thousands of times in many different countries, experimenters have found that only one-third of dictators keep all of the money for themselves. The other two-thirds show substantial generosity, allocating an average of 42 percent of the money to the other player. In addition, 17 percent of all dictators split the money perfectly evenly and a little over five percent of all dictators give the other player everything.

c. Evaluate the following statement. “We shouldn’t generalize from what people do in the ultimatum game because $10 is a trivial amount of money. When larger amounts of money are on the line, people will act differently.”

Answer 1

Qa) The answer to the question, as worded here, is indeterminable because it is not specified whether the dictator is anonymous only to the other participant or to the research group also.

1. Put another way, can the one or more of the Research Group (RG) members map (link) the Dictator's identity to the share ratio?

2. Does the Dictator know this?

3. Even if (1) isn't true, does the Dictator believe it to be 'yes' (i.e. knowledge of his choice will "leak") based on everyday worldly experience? Or does the Dictator not necessarily believe it but still continue with that as a rational presupposition?

These questions are relevant because the individual's preference profile (ranking of alternative choices) is relevant. The preference profiles govern individual motivation, and motivation governs action.

-----

The question of motivation is one of the basic questions of social psychology and sociology, so we cannot expect a simple answer, but an interesting theory is based on the term FACE, which is used in much the same way as in the expressions to lose face and to save face, meaning something like "self-respect" or "dignity". The theory was developed by Erving Goffman, an American sociologist.

The basic idea is this: we lead unavoidably social lives, and "Face" is something people give to us (e.g. me, you), and which may be more or less worthy of dignity in our (my, your) own eyes as by cultural standards.

Our face is a very fragile thing which other people can easily damage, so we can lead our social lives according to the Golden Rule - do to others as you would like them to do to you - by looking after other people's faces in the hope that they will look after ours.

I provided this background to deliberately draw a distinction between social image and face: "face" is image and worth in our own eyes, "social image" is image and worth in the community's eyes. They may not overlap. You may feel more or less bad about a bad action depending upon how certain you are the information doesn't "leak" outside of your personal bubble.

And this is because humans, individuals and/or collectively, as adaptive organisms, objectively but perhaps unknowingly on a micro level, optimize their own social image to maximize survival / utility.

They do this not only by momemt to moment microadjustment, but also by long-term training of baseline behavior using internal feedback mechanism of "feelings". Habits associated with "good" feelings are strengthened.

Having the information leak means that at least one person (and by unconstrainted transmission, possibly the world) knows that the Dictator (or rather the person momentarily playing Dictator) doesn't follow the Golden Rule when playing the Social Game of cooperation. I capitalized G in Game to stimulate a game theoretic perspective of the situation. When the situation arises an unknown number of people, armed with this knowledge, are likelier to choose to forego following the Golden Rule when playing a 1v1 with Dictator in a future encounter (like in a repeated game).

Suppose the Dictator doesn't follow the logic I clarified above. Then Dictator will make selfish decisions in public eyes, harming the symbiotic relationship with the environment around them and receiving negative stimuli in the process. This negative stimuli goes on to form a feedback loop for behavioral adjustment, where Dictator continues to adjust their behavior until the negative stimuli vanishes or is minimized - i.e. until present and future expected behavior falls in line with implicit coordination rules of the social matrix.

On future occasions, the soon-to-be-Dictator would carry this product of operant conditioning with them into the experiment. They'd know that IF the information leaks out then the monetary gain isn't worth the risk - but only if.

There is not enough data to resolve this IF for the given context.

^ As a concluding addendum, when I say things like the Dictator "knows" in "They'd know that IF the information leaks out . . . ", I don't mean they necessarily know on the conscious level. The bulk of the computation informing decisions and behavior happens below consciousness, so what we're really talking about is emprically observed behavior rather than consciously premeditated behavior.

-----------

Qc) Of course changing the money sum affects the expected payoff of each alternative strategy, so in general yes people will behave differently with a larger sum of money.

But with the logic I demonstrated above, it stands to reason that the change can be in either direction - more selfish or less selfish, given that a representative person's preference for free money isn't monotonic.

^ You may feel uncomfortable accepting more than 2 free drinks from somebody lest you look like a penniless hobo. But if that same person had offered you a coupon for an year of free drinks from the get-go, no T&C, would you have denied? Even if the coupon doesn't have your name on it i.e. it's regiftable?

-----

Qb) I assume what is asked for here is Standard Deviation. Calculate the weighted SD with the following data:

Weight w1= (0.33) of people ---> X1 = share ratio 0

Weight w2= (0.62) of people ---> X2 = share ratio 0.42

Weight w3= (0.05) of people ---> X3 = share ratio 1

N = 3 || Mean of x = 0.47