Bonus Article for Researchdesign explained
You may want to assign thefollowing article:
Liberman, V.,Samuels, S. M., & Ross, L. (2004). “The name of the game: Predictivepower of reputations versus situational labels in determining prisoner’sdilemma game moves. Personality and Social Psychology Bulletin
You can use the first part of this article withChapter 1 (to show that research, rather than merely verifying intuition,sometimes finds that intuition is wrong), with Chapter 2 (to introduce studentsto a research paradigm that was frequently used by social psychologists in the1960s and is frequently used by experimental economists these days: ThePrisoner’s Dilemma game), with Chapter 9 (to discuss simple experiments),or with Chapter 11 (to discuss factorial designs that have one experimentalfactor and one nonexperimental factor). Although the article is easy to read,you can make it even easier for students to read by giving them Table 1.
Guide to Understanding the Article
Tips, Comments, and Problem Areas
N-move Prisoner’s Dilemma: The Prisoner’s Dilemma gets its name from a problem sometimes faced by a criminal who, along with his partner, has been arrested. The police have a case, but their case would be a lot stronger if one or both criminals confess. Therefore, the criminal is asked to testify against the partner (defect) rather than maintaining the partner’s innocence (cooperating with the partner). The dilemma comes from the various payoffs of cooperating versus defecting. The criminal can get off free by ratting out the partner (defecting)—if the partner has decided not to rat out the criminal. If, however, both criminals defect, they both get long sentences. If, on the other hand, they both refuse to rat the other one out (they cooperate with each other), they both get light sentences. Thus, the situation is set up where (a) the group as a whole benefits most if people cooperate, but (b) an individual can be better off by not cooperating—if the partner cooperates, and (c) the worst outcome is to be a sucker: to cooperate when your partner defects. The N-move Prisoner’s Dilemma means that it is not just a one-round game: In Round 2 and other rounds, you will be able to adjust your strategy based on what your partner did on previous rounds.
Probably the best way to understand the Prisoner’s Dilemma is to experience it. Either of the two websites below will let you play the Prisoner’s Dilemma game online.
To learn more about the Prisoner’s Dilemma game, you can read this interesting encyclopedia entry by going to
“dialogue between social psychology and personality psychology”: refers to the discussion about the relative importance of situational, social factors (in this case, seeing if varying the name of the game from one that might encourage competition [“The Wall Street Game”] to one that might encourage cooperation [“The Community Game”]) would have more impact than differences in personality factors (in this case, comparing people judged to have highly competitive personalities with people judged to have cooperative personalities).
“subjectivist tradition of psychology and the objectivist spirit of game-theory economics”: Economists have focused only on the objective reality (e.g., how much money does the participant get for making a certain decision). In contrast, psychologists have looked at how people intepret reality (e.g.,
altruistic motives: unselfish reasons
strategic motives: selfish reasons
Payoff matrix: the rewards and costs of making certain decisions. The matrix can be expressed in a table. For an example of a payoff matrix, see the table in the Procedure section (page 1176).
“framed”: worded in a way that affects how people interpret the situation; presenting the problem in a way that certain values, beliefs, or norms seem more relevant
malleability of construal processes: the extent to which people’s interpretations of events can be changed; the degree to which interpretations can be influenced; we can manipulation how people will interpret an event
Connotes: implies, suggests
dispositionist expectations: beliefs that personality is what determines behavior
STUDY 1: REPUTATIONS VERSUS SITUATIONAL LABELS IN THE PD GAME
Participants were not told why they had been selected to be in the study.
Participants all played the same game. However, for half of them, the game was called “Wall Street Game” and, for the other half, the game was called “Community Game.”
Results and Discussion
Nomination status: whether a participant was viewed by the resident assistants as being cooperative or competitive.
Chi-square (collapsing across nomination status): chi-square is a simple statistical test. Collapsing across nomination status meant that they just compared all the community game players against all the Wall Street game players—ignoring how resident assistants had labeled each player.
You can do their analyses by going to the Chi-square calculator at http://www.georgetown.edu/faculty/ballc/webtools/web_chi.html
You will be using a 2 X 2 table. Therefore, in the "Table dimensions" line, enter "2" in the square before "rows" and "2" in the square before "columns" (as we have done below).
After pressing the “Generate table” button, you will fill in the blanks in the table. Note that the researchers had 48 participants—24 who played the “Wall Street” game and 24 who played the “Community game.” Thus, if 33% in the Wall Street game condition cooperated, 8 (.33 * 24 =8) participants cooperated. When you are finished, your table should look like the following:
When you press the “Calculate chi square” button, you should get the following:
p < .01 : means that, if the manipulation did not have an effect, we would get a difference this large in fewer than 1 in 100 cases. Usually, with such findings, researchers conclude that the manipulation did have an effect.
Fisher’s exact test: an alternative to the Chi-Square test. It can only be used when you have a 2 X 2 table. However, when you have fewer than 50 observations (as in this case), it is more accurate than the chi-square test.
p < .02 : means that, if the manipulation did not have an effect, we would get a difference this large in fewer than 2 in 100 cases. Usually, with such findings, researchers conclude that the manipulation did have an effect.
Nomination status: In this case, participants belong to one of two: (a) being chosen to be in the study because the participant’s RA thought the participant would use a competitive strategy or (b) being chosen to be in the study because the participant’s RA thought the participant would use a cooperative strategy.
Mean: average calculated like you did in grade school—by adding up the scores and dividing by the number of scores
Dyads: two-person groups; in this case, they are referring to the pairs of participants who played the game together.
F < 1.0 = The greater the difference between groups, the bigger F tends to be. The larger F is, the more likely it is that the differences will be statistically significant. Often, people consider significant differences to be reliable differences—ones that you would find again if you repeated the study again. If the manipulation does not have an effect, you would still expect an F of 1 or higher. Thus, an F of less than one means there was no evidence that the groups are reliably different.
“… because responses across successive rounds obviously are not independent”: Most statistically tests presume that responses are independent—one response does not affect another response. In this case, however, earlier responses affect later responses. For example, what happened in Round 1 (e.g., cooperation or not) will affect what happens in Round 2.
Note that participants knew that if they chose to defect on the seventh round, they did not have to fear retaliation on the next round because there was not going to be a next round.
Belied by the data: the evidence shows that it is wrong