<%BANNER%>

Race, Gender, and the Prisoner's Dilemma

Permanent Link: http://ufdc.ufl.edu/UFE0018961/00001

Material Information

Title: Race, Gender, and the Prisoner's Dilemma A Study in Social Dilemma Cooperation
Physical Description: 1 online resource (105 p.)
Language: english
Creator: Romano, Victor Eduardo
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: black, cooperation, dilemma, expectation, game, gender, prisoners, race, religiosity, social, trust, white
Sociology -- Dissertations, Academic -- UF
Genre: Sociology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Understanding the intricacies of social dilemmas in general and the Prisoner's Dilemma in particular is important due to the myriad of social issues that are manifestations of such dilemmas. Macro and micro issues ranging from global warming and nuclear arms treaties to recycling and gun ownership are at their crux, social dilemmas?situations where individual interests are in opposition to and contend with the interests of a larger social unit. Despite the importance of social dilemma cooperation in a world that is becoming increasingly interdependent as a result of globalization, there is a dearth of research examining the influence of race on intra- and inter-group cooperation. To identify if there is a link between race, gender, and levels of cooperation, trust, and expectations of a partner in social dilemmas, this study utilizes quantitative and open-ended data collected from undergraduate college students attending a major public university in the southeastern United States. Drawing on the theoretical perspectives of games theory, race relations, and gender studies, this study examines the relationship between race, gender and the decision of an individual to either cooperate or defect in a monetary Prisoner's Dilemma game. By placing participants in same-race and same-gender pairs as well as cross-race pairs the study shows that Black females have significantly higher same-race group cooperation rates than do White females, and that Blacks have significantly higher overall cooperation rates than do Whites. The study also shows that cross-race pairs have higher rates of cooperation than do exclusively White pairs, and that religiosity is positively correlated with cooperation. The reasons given for social dilemma cooperation and defection by different racial and gender groups were also explored. It was found that though reasons for cooperation and defection were mostly similar across groups, minority group members were more likely to take their racial and gender group status into consideration when making game decisions. Additionally, it was also found that Blacks were more likely to view game decisions from a moral perspective and as a reflection of individual character than were Whites.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Victor Eduardo Romano.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Vera, Hernan.
Local: Co-adviser: Koropeckyj-Cox, Tanya.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0018961:00001

Permanent Link: http://ufdc.ufl.edu/UFE0018961/00001

Material Information

Title: Race, Gender, and the Prisoner's Dilemma A Study in Social Dilemma Cooperation
Physical Description: 1 online resource (105 p.)
Language: english
Creator: Romano, Victor Eduardo
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: black, cooperation, dilemma, expectation, game, gender, prisoners, race, religiosity, social, trust, white
Sociology -- Dissertations, Academic -- UF
Genre: Sociology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Understanding the intricacies of social dilemmas in general and the Prisoner's Dilemma in particular is important due to the myriad of social issues that are manifestations of such dilemmas. Macro and micro issues ranging from global warming and nuclear arms treaties to recycling and gun ownership are at their crux, social dilemmas?situations where individual interests are in opposition to and contend with the interests of a larger social unit. Despite the importance of social dilemma cooperation in a world that is becoming increasingly interdependent as a result of globalization, there is a dearth of research examining the influence of race on intra- and inter-group cooperation. To identify if there is a link between race, gender, and levels of cooperation, trust, and expectations of a partner in social dilemmas, this study utilizes quantitative and open-ended data collected from undergraduate college students attending a major public university in the southeastern United States. Drawing on the theoretical perspectives of games theory, race relations, and gender studies, this study examines the relationship between race, gender and the decision of an individual to either cooperate or defect in a monetary Prisoner's Dilemma game. By placing participants in same-race and same-gender pairs as well as cross-race pairs the study shows that Black females have significantly higher same-race group cooperation rates than do White females, and that Blacks have significantly higher overall cooperation rates than do Whites. The study also shows that cross-race pairs have higher rates of cooperation than do exclusively White pairs, and that religiosity is positively correlated with cooperation. The reasons given for social dilemma cooperation and defection by different racial and gender groups were also explored. It was found that though reasons for cooperation and defection were mostly similar across groups, minority group members were more likely to take their racial and gender group status into consideration when making game decisions. Additionally, it was also found that Blacks were more likely to view game decisions from a moral perspective and as a reflection of individual character than were Whites.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Victor Eduardo Romano.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Vera, Hernan.
Local: Co-adviser: Koropeckyj-Cox, Tanya.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0018961:00001


This item has the following downloads:


Full Text





RACE, GENDER, AND THE PRISONER'S DILEMMA:
A STUDY IN SOCIAL DILEMMA COOPERATION
























By

VICTOR EDUARDO ROMANO


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2007




























2007 Victor Eduardo Romano

































To Kelli, for always being right and always being on my side









ACKNOWLEDGMENTS

I would foremost like to acknowledge my parents, Hugo and Patricia C. Romano, for

their unwavering love, faith, and support. Acknowledgments are also due to Dr. Hernan Vera

and Dr. Tanya Koropeckyj-Cox for always being generous with their time and encouragement.

The guidance and mentorship they have provided me over the years means more than they know.

Lastly, I would like to acknowledge my wife and proofreader, Kelli Romano, for always reading

anything at anytime. For her patience and care, I am forever grateful.









TABLE OF CONTENTS
page
pM.ge

ACKNOWLEDGMENTS ....................................................... .. ...........4

LIST OF TABLES......................................................7

ABSTRACT.......................8........... ......................8

CHAPTER

1 INTRODUCTION ............................................................... ........... 10

Specific Aim s ............................................................ .. ......... 13
M ajor Research Questions ... ................................................ .. ......... 15

2 LITERATURE REVIEW ...................................................... ...........17

Race and Cooperation ............................ .. .. .... .............................. 18
Critique of Race in Prisoner's Dilemma Literature............................... ............. 19
Theorizing Racial Cooperation.................................. ......... .............. 21
Theoretical Basis for Greater Black Cooperation........................ ............21
Theoretical Basis for Greater Cross-Racial Group Defection.......................23
Alternative Theoretical Arguments: Against Differing Cooperation Levels.......26
Gender and Cooperation................ .........................................28
Theorizing Gender Cooperation ...................... .. .. .... ............. 29
Gender as a Social Construction and Social Structure.............................. 29
Gender Cooperation and Social Decision Making........................ ...........31
The Intersection of Race and Gender.......... ............... .............35

3 RESEARCH METHODOLOGY.................. ......................37

Research Design .......................................................... .......... 37
Measures and Instruments................. ....... ..................... .......... 39
Experim ental D esign .................................... ........... ............. ............40
Recruitment of Participants ....................................................... .......... 41
Demographic Characteristics of Sample......................................... ..............41
Analyzing the Data ..................................................... ......... 43
R ole of the R researcher ............................................. .............. ............43

4 QUANTITATIVE RESULTS............. .......................45

Hypotheses ............................................................... .......... 45
Quantitative Findings ................................................... .......... 45
Cooperation .............................................. .... ... ... ................45
Quantitative Analysis of Trust and Expectations......................... ............48
The Role of Religiosity........... ....................................................49









D discussion of Quantitative Findings.................................... ...............51

5 OPEN-ENDED DATA RESULTS........... ................... ..............57

Open-ended Data Findings.............................. .............57
External Validity ............................................. ....................58
Reasons for Cooperation and Defection................................... ...........60
W hite male experiment group..................... ..... ..... ..............60
White female experiment group........ ........ ................61
Black male experiment group......................... .......... ............63
Black female experiment group................... ............65
Male cross-race experiment group............. ............. ...........66
Female cross-race experiment group.............................. ...........68
Open-ended Data Analysis of Trust and Expectations................... ............ 69
Gender and Race Differences .......................... .... ......... .............73
Discussion of Open-ended Data Findings................................. ............76
6 C O N CLU SIO N .......................................... ........... ............ ............81

APPENDIX

A MEASUREMENT INSTRUMENT.........................................................90

B INFORM ED CONSENT .......................................... ............. ............97

LIST OF REFERENCES..................................................... .. .......... 98

BIOGRAPHICAL SKETCH ............................................................105









LIST OF TABLES


Table page

4-1 Proportion of cooperation and defection by race/gender experiment group .......54

4-2 Distribution of Game Pair Outcomes by Race and Gender Experiment Group... .54

4-3 Proportion of cooperation and defection in exclusively White and exclusively
Black experiment groups.............................. .............54

4-4 Proportion of cooperation and defection by race......................... ...........54

4-5 Proportion of cooperation and defection by gender...................... ............55

4-6 Proportion of cooperation and defection by trust in game partner...................55

4-7 Proportion of cooperation and defection by expectation of game partner's
decision ................................................. .. ......... 55

4-8 Expectation of game partner's decision by race.......................... ...........55

4-9 Expectation of game partner's decision by race and gender......................55

4-10 Proportion of cooperation and defection by religiosity................. ............55

4-11 Level of religiosity by race ..................... ............. ................ ...56

4-12 Level of religiosity by race and gender............................56

5-1 Thematic coding categories with frequencies for responses to the question,
"Why did you cooperate or defect? Please explain in detail........................79

5-2 Reasons for cooperation and defection by experiment group with
frequency of response ........................................ .. .......... 79









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

RACE, GENDER, AND THE PRISONER'S DILEMMA: A STUDY IN SOCIAL DILEMMA
COOPERATION

By

Victor Eduardo Romano

December 2007

Chair: Hernan Vera
Cochair: Tanya Koropeckyj-Cox
Major: Sociology

Understanding the intricacies of social dilemmas in general and the Prisoner's Dilemma

in particular is important due to the myriad of social issues that are manifestations of such

dilemmas. Macro and micro issues ranging from global warming and nuclear arms treaties to

recycling and gun ownership are at their crux, social dilemmas-situations where individual

interests are in opposition to and contend with the interests of a larger social unit. Despite the

importance of social dilemma cooperation in a world that is becoming increasingly

interdependent as a result of globalization, there is a dearth of research examining the influence

of race on intra- and inter-group cooperation.

To identify if there is a link between race, gender, and levels of cooperation, trust, and

expectations of a partner in social dilemmas, this study utilizes quantitative and open-ended data

collected from undergraduate college students attending a major public university in the

southeastern United States. Drawing on the theoretical perspectives of games theory, race

relations, and gender studies, this study examines the relationship between race, gender and the

decision of an individual to either cooperate or defect in a monetary Prisoner's Dilemma game.

By placing participants in same-race and same-gender pairs as well as cross-race pairs the study









shows that Black females have significantly higher same-race group cooperation rates than do

White females, and that Blacks have significantly higher overall cooperation rates than do

Whites. The study also shows that cross-race pairs have higher rates of cooperation than do

exclusively White pairs, and that religiosity is positively correlated with cooperation.

The reasons given for social dilemma cooperation and defection by different racial and

gender groups were also explored. It was found that though reasons for cooperation and

defection were mostly similar across groups, minority group members were more likely to take

their racial and gender group status into consideration when making game decisions.

Additionally, it was also found that Blacks were more likely to view game decisions from a

moral perspective and as a reflection of individual character than were Whites.









CHAPTER 1
INTRODUCTION

In an era of global conflicts, nuclear proliferation, global diseases, and environmental

destruction and deterioration, the need to understand the macro and micro dynamics of racial and

gender cooperation is paramount. It is not an exaggeration to suggest that the survival of the

human species depends largely on our ability to cooperate with people who we see as different

from ourselves. Moreover, what Marx (1862) refers to as "the social productive power," which

arises from cooperation, is both society's best tool for the betterment of humankind and as with

exploitation, patriarchy, sexism and racism, the foundation of amongst the most oppressive of

social conditions (361). According to Marx, "the essence of simple cooperation remains

simultaneity of action, a simultaneity whose results can never be attained by the temporal

succession of the activities of the individual workers" (1862:215). In other words, many people

working together can achieve what one alone cannot. Cooperation, as utilized in this study, refers

to an individual's decision to attempt to act in solidarity with another person to achieve a

mutually beneficial outcome, instead of acting in competition.

In the social world, reasons for cooperation among individuals include economic

conceptions such as self-interest, reciprocity, obligation, and necessity. Yet, not all forms of

cooperation can be explained by economic theories of rational choice, which posit that

individuals are rational, self-interested, economic human beings (Hu & Liu 2003). Ideas such as

altruism and loyalty often cause individuals to act against their own so-called rational economic

interests in favor of pursuing the interests of others. Volunteerism, philanthropy, and community

activism are a few examples of individuals altruistically contributing to the welfare of others at

their own expense. Furthermore, conceptions of altruistic cooperation can extend beyond

individuals to groups (including racial, gender and religious groups) and nation-states. In this,









cooperation is not merely as Marx stated, "the general form on which all social arrangements for

increasing the productivity of social labour are based," but also a means of demonstrating

solidarity and empathy toward other group members (Marx 1862:209).

One of the largest stumbling blocks to cooperation, however, is the proclivity of humans

to defect (work against one another) in what are called social dilemmas. Social dilemmas "are

situations in which (a) [defection] yields the person the best payoff in at least one configuration

of choices made by others; (b) [defection] has a negative impact on the interests of other persons

involved; (c) the collective choice of [defection] results in a deficient outcome, that is, a result

that is less preferred by all persons than the result which would have occurred if all had

[cooperated] instead of [defected]" (Liebrand 1986:113-14). Simply stated, a social dilemma is

at its essence a situation that pits individual gain against the greater good of the dyad, group,

society, and/or humankind.

Among the most studied social dilemmas is the game Prisoner's Dilemma (Simpson 2003).

Bearing in mind that games with strategies, rules, and often winners are a universal social

activity, "game theorists analyze human behavior by arguing that many situations can be seen as

manifestations of one or other quite simple game" (Wallace & Wolf 1999:322). The Prisoner's

Dilemma is likely the most popular game of games theory; its popularity is largely due to the

inevitable dreadfulness of its most frequent outcome (Wallace & Wolf 1999). The premise of

the game is that two prisoners who have committed a crime together are under arrest and unable

to communicate with each other. In order to force a confession the interrogator offers each

prisoner, separately, the following deal:

1. If you confess and your companion does not, he will be sentenced to nine years in
prison and you will be let off scot-free.









2. If your companion confesses and you do not, you will be sentenced to nine years in
prison and he will be let off scot-free.

3. If you both confess, you will each be sentenced to five years in prison.

4. If neither of you confess you will each be sentenced to two years in prison.

More often than not, both prisoners end up confessing (defecting) and must each serve five-year

sentences. The possibility of getting off scot-free, coupled with the fear of being imprisoned for

nine years, persuades the prisoners to choose the third worst individual sentence and the worst of

all in terms of total number of years' imprisonment for both prisoners combined. "The paradox

that makes the Prisoner's Dilemma so intriguing is that both prisoners end up defecting even

though they both know they would be better-off cooperating" (Wallace & Wolf 1999:323).

Comprehending the intricacies of social dilemmas in general and the Prisoner's Dilemma

in particular is of critical societal importance due to the large number of social issues that can be

seen as manifestations of such dilemmas. As Barash (2004) notes, macro social issues such as

global warming, water shortages, abuse of public lands, public versus private transportation, and

even the Kyoto Treaty can be viewed through the theoretical framework of the Prisoner's

Dilemma. Additional social issues in which the Prisoner's Dilemma framework can be applied

include recycling, home upkeep, gun control, blood donation, littering, and even micro-level

social issues such as reciprocity in everyday exchanges between neighbors (Yamagishi & Cook

1993; Macy & Skovoretz 1998). Considering the myriad of issues that are at their crux social

dilemmas, understanding the factors that contribute to or hinder cooperation is of critical

importance because-as in all of the above mentioned social dilemmas-when defection

becomes the rule, everyone loses.









It is also for this reason that understanding the gender and to a larger extent the racial

dynamics that influence group cooperation and defection is important. As Kaplan and Kaplan

(2006) eloquently describe:

In real life, times of visible change-when populations increase, resources decline
or new spoils become available for distribution-create conditions where the
slightest germ of mistrust can rapidly generate a prisoner's dilemma. Protestant
and Catholic, Serb and Croat, Hutu and Tutsi; two populations in one space can
find, even without any great prior enmity between them, that the outcomes of
life's game are suddenly realigning. The majority in a mixed population may still
believe that peace and cooperation are best, but if a sufficiently large minority
come to think that its interests are served only by the victory of its own tribe or
creed, then this rapidly becomes a self-fulfilling assumption. You fear your
neighbor might burn down your house; will you wait until he comes with his
shadowy friends and their blazing brands? No, best call your friends, best find
matches and fuel... Civil society rapidly curdles: individuals lose the chance to
choose for themselves. Even the brave who stand up for peace lose everything,
betrayed by their fellow prisoners.

With the above scenario in mind, it is surprising that the particular effect of race and

culture on cooperation is an area of inquiry that has been largely neglected by past studies that

employed the Prisoner's Dilemma. Since different racial and gender groups place different types

and levels of expectations on their members, racial and gender considerations of group loyalty,

obligations, trust, and solidarity can complicate the individual rational choice process of

choosing social dilemma cooperation and defection. Furthermore, if social dilemma cooperation

or defection levels are higher among certain societal groups (as my research shows they are),

what factors) can this difference be attributed to? This is a question that qualitative and

theoretical Prisoner's Dilemma research on race, gender, and their intersectionality can help to

answer, and which I will explore further in this dissertation.

Specific Aims

To identify if there is a link between race, gender, and levels of cooperation, trust, and

expectations of a partner in social dilemmas, I collected and analyzed data from college students









attending a major university in the southeastern United States. The goal of the research was to

examine if and how race and gender are correlated to levels of cooperation, trust and

expectations of a partner through the use of the theoretical framework of games theory and the

social dilemma game Prisoner's Dilemma.

Above all, the specific aims of this study were to: (1) determine if Black1 college

students, as members of a subordinated group, are more likely to cooperate in a social dilemma

than are White college students who are members of the dominant group; (2) determine if female

college students are more likely to cooperate with each other in a social dilemma than are male

college students; and (3) to explore the reasons and rationale given by respondents for

cooperation or defection. By creating a monetary variant of the Prisoner's Dilemma game for

the participants of this study, I examined the role of race and gender in participants' decision to

cooperate or defect. I found that Black females have significantly higher same-race group

cooperation rates than do White females, and that Blacks have significantly higher overall

cooperation rates than do Whites. The study also shows that cross-race pairs have higher rates of

cooperation than do exclusively White pairs and that religiosity is positively correlated with

cooperation. Drawing on works such as Axelrod's (1984) The Evolution of Cooperation and

Gilligan's (1982) In a Different Voice, this study contributes to our knowledge of games theory,

social dilemma theory, social psychological theory, rational-choice theory, symbolic interaction

theory, as well as the fields of racial and ethnic relations and gender studies. Finally, this study

strives to provide insight into the ways in which men and women reason and cooperate and how

race may influence these processes.


1 I employ the racial classifications Black and White rather than ethnic classifications of African American and
Caucasian in acknowledgement of the diverse origins of my sample population. Additionally, the use of the racial









Major Research Questions

Though many scholars have employed Carol Gilligan's (1982) work on differences in the

moral development of women and men as the foundation from which to project differences in

cooperation in the Prisoner's Dilemma (including Stockard, van de Kragt & Dodge 1988; Brown

& Taylor 2000), inconsistent results and contradictory findings suggest that the question of

whether gender is associated with cooperation in the Prisoner's Dilemma and other social

dilemmas remains unanswered (Ledyard 1995). Additionally, the role played by race in an

actor's decision to cooperate or defect has been largely absent from the research literature and

needs to be further explored. In this study, I examine the relationship between race, gender, and

the tendency of an individual to either cooperate or defect by analyzing decisions made by same-

race and same-gender pairs of participants, as well as cross-race pairs of participants in a

monetary Prisoner's Dilemma game. As a student specializing in racial relations, I was surprised

at the lack of Prisoner's Dilemma studies employing race as a variable and exploring the effect

of race, gender, and their intersectionality on social dilemma cooperation. My knowledge of

groups led me to believe that racial and gender group membership (each with their own varying

expectations of group loyalty and obligation) would influence participants to be more

cooperative when paired with a member of their same race and gender group. Additionally, I felt

that minority group status would result in greater dyad solidarity and thus a higher probability of

social dilemma cooperation. These views, buttressed by social theory led me to my particular

research questions, which are:

1. Is there a correlation between race and likelihood of cooperation in a social
dilemma?


term "Black" places emphasis on race as referring to the physical, phenotypic, and visible differences that are given
social significance in American society. When discussing the research of others, however, I keep with the racial
classifications they employed.









2. Is there a correlation between sameness of gender and likelihood of cooperation in a
social dilemma?

3. What are the reasons and rationale given by participants for choosing to cooperate or
defect?

4. Do levels of trust and expectations of a game partner's decision differ between race and
gender groups?

To explore these questions, Chapter 2 reviews the relevant previous Prisoner's Dilemma research

and discusses the theoretical support for varying levels of cooperation among different racial and

gender groups. Based on the theoretical perspectives discussed in Chapter 2, Chapter 3 describes

the design and methodological underpinnings of the experiment and poses five hypotheses.

Chapter 4 specifies and discusses the study's quantitative results, which yield support for the

claim that Blacks are more likely to cooperate in social dilemmas than are Whites. Chapter 5

examines the validity of the experiment design and then specifies and discusses the open-ended

questionnaire results, which describe distinctions in the ways the various race and gender groups

rationalize their game decisions. Chapter 6 comes to a conclusion and provides suggestions for

future research.









CHAPTER 2
LITERATURE REVIEW

The literature on the Prisoner's Dilemma is astonishingly extensive and varied.

Researchers in the fields of sociology, psychology, economics, political science, criminology,

mathematics, biology, zoology, and others have used various versions of the Prisoner's Dilemma

to study a plethora of issues. One intriguing example of the breadth of research conducted

incorporating the Prisoner's Dilemma included using functional magnetic resonance imaging to

scan the brain of women playing the Prisoner's Dilemma to gain insight into the decision making

process by looking at what brain regions were activated during game play (Spinney 2004).

Others have used the Prisoner's Dilemma to explain why certain foraging animals share food

without any overt aggression (Dubois & Giraldeau 2003) or how an attractive experimenter or

game partner affects a participant's decision on whether or not to defect (Morse, Reis, Gruzen &

Wolff 1974). There is even a Prisoner's Dilemma world-championship competition in which

researchers from many disciplines submit computer programs to a repetitive version of the

Prisoner's Dilemma in a round-robin tournament (Axelrod 1980). Interestingly, the most

successful strategy, tit-for-tat2, is also one of the simplest (Axelrod 1980).

Though both the variation and uses that the Prisoner's Dilemma has undergone are

exhaustive, the Prisoner's Dilemma literature most relevant to this study can be grouped into two

manageable categories. The first category, Race, is exceedingly small and contains research on

how race correlates with cooperation/defection in Prisoner's Dilemma games. The second and

larger category, Gender, contains research addressing how gender correlates with

cooperation/defection in the Prisoner's Dilemma. Below is a brief overview of the Race and



2 The tit-for-tat strategy in the Prisoner's Dilemma game consists of a player cooperating on their first move and
then copying the previous round's cooperation/defection decision of their opponent for the duration of the game.









Gender categories each followed by a discussion of the theoretical foundations for predicting

variance in cooperation rates between racial/gender groups.

Race and Cooperation

The research literature on how race affects cooperation in the Prisoner's Dilemma is

miniscule. Early research on race and the Prisoner's Dilemma shows that participants

demonstrate higher rates of cooperation when they play against a member of the same racial

group (Sibley, Senn & Epanchin 1968; Wilson & Kayatani 1968; Baxter 1973). Furthermore,

Cederblom and Diers (1970) showed a tendency among White students to behave in a less

cooperative manner toward cooperative Black game partner confederates3 than toward

cooperative White game partner confederates. Additionally, Cederblom and Dies showed that

Whites were more cooperative toward Black confederates who at first displayed competitiveness

and later cooperated, versus Black confederates who mostly cooperated throughout the game

play. These findings were taken by the researchers to mean that Black assertiveness is both

understandable and necessary in order to affect institutional change.

In a later study of 42 same-sex pairs of eighth-grade students of varied (Black and

White) racial composition, Downing and Bothwell (1979) discovered that females of either race

tended to cooperate in same-race pairs and to compete in mixed-race pairs. They also found that

White males tended to cooperate and Black males to compete independent of their partner's race.

Owens' (1998) designed a study "to explore several facets of cross-racial interactions of young

male African Americans and Caucasian Americans," finding evidence that for both African

Americans and Caucasian Americans cooperation was in certain cases correlated with the

participant's respective stage of racial identity development as measured by Racial Identity Scale


3 I use the term confederate to refer to a participant's game partner who was (unbeknownst to the participant) hired
by the researcher and typically employed a predetermined ratio of cooperation and defection responses.









(RIAS) or White Racial Attitude Identity Scale (WRIAS) (1988:2488). Although Owens' study

provides interesting insight into psychological aspects of cross-racial cooperation, it does not

measure or compare same-race group cooperation rates, which is one of the aims of this present

study.

More recently, Heider and Skowronski (2007) found that in two 50-shot4 Prisoner's

Dilemma games (designed based on procedures adapted from Baxter [1973]) Caucasian

participants unexpectedly cooperated at a higher rate when presented with a computer image and

name of an African American partner (M= 51.8%) than when presented with a computer image

and name of a Caucasian partner (M = 48.4%). They also found that measures of racial attitude

such as the Implicit Association Test (IAT) and the Pro-Black subscale of the Pro-Black/Anti-

Black Attitudes Questionnaire (PAAQ) predicted social behaviors of Caucasian participants

toward African American targets. In the following section, I offer a critique of past studies and

highlight some of their limitations and weaknesses.

Critique of Race and Prisoner's Dilemma Literature

A major problem with many of the past studies using the Prisoner's Dilemma to measure

racial attitudes and cooperation rates is the use of exclusively White participants. In doing so,

studies such as Baxter's (1973), Cederblom and Diers' (1970) and Heider and Skowronski's

(2007) privilege the views and decisions of Whites over those of Blacks. Moreover by recruiting

only White participants, these studies provide no insight into Black same-race and cross-race

cooperation.

Another problem with past studies is the exclusive use of multi-shot games. Minkler and

Miceli (2004) have noted that by employing a one-shot game version of Prisoner's Dilemma,

rather than a multi-shot game version, you can assure that any resulting cooperation or defection









does not come from reputation or punishment effects associated with repeated games.

Additionally, the one-shot version employed by this study protects against players cooperating or

defecting as the result of a refined strategy developed over multiple plays. In terms of validity

and generalizability, a one-shot game may also prove more similar to social dilemmas

encountered by people on a daily basis. The decision of an individual to litter, cut in front of a

line, or use an express lane of a supermarket with more than the maximum allowed items is a

decision that is typically made without consideration to past or future decisions in the same

social dilemma-style situation. Also, the law of diminishing marginal utility suggests that a

participant will place more significance on a one-shot Prisoner's Dilemma decision, than on, for

example, the 47th-shot decision of a Prisoner's Dilemma game. This suggestion is supported by

the work of Clark and Sefton (2001) who found that the most important variable influencing

cooperation in sequential multi-shot Prisoner's Dilemma games is the first-mover's choice.

Their study yields support for the argument that cooperative behavior in multi-shot social

dilemmas reflects reciprocation rather than unconditional altruism. Moreover, they found their

cooperation in Prisoner's Dilemma games decreases with repetition. Hence, single-shot

Prisoner's Dilemma games are also better barometers of altruistic behavior than are multi-shot

games.

As will become evident in the Methods Chapter (3), I seek to address these deficiencies

of past research in this area. In the next section, I will explore some of the theoretical

foundations for possible differences in cooperation rates between different racial groups in a

social dilemma, as well as the theoretical foundations for no variation by racial groups in a social

dilemma.



4 The word "shot" refers to how many iterations of the game were played.









Theorizing Racial Cooperation

Though much has been written on the rationale behind actors' decisions to cooperate or defect in

social dilemmas, the possible influence of race on these decisions is a topic that has not yet been

thoroughly explored. While race is a fluid and ever-changing social construct that should no

longer be used as a system of classification in either the biological or social sciences, it is a

concept that is nevertheless accepted as a real and meaningful distinction by most Americans

(Vera & Gordon 2003a). W.I. Thomas (1931) noted that, "situations that are defined as real, are

real in their consequences," hence, since most Americans define race and racial groups as real,

being categorized into a particular racial group has multitude of real consequences for an

individual that will affect their life outcomes and perspectives (145). This study seeks to

examine what influence, if any, racial group membership has on a participant's decision to

cooperate or defect with another participant of either the same or a different racial group. For

this purpose, the study employs Feagin and Feagin's (2002) definition of a racial group as "a

social group that persons inside or outside the group have decided is important to single out as

inferior or superior, typically on the basis of real or alleged physical characteristics subjectively

selected" (7). Despite the paucity of research in this area, there is a sufficient theoretical basis to

suggest that racial group affiliation may influence a participant's decision to cooperate or defect

in a social dilemma.

Theoretical Basis for Greater Black Cooperation

I begin with the hypothesis5 that Black pairs will have higher same-race group

cooperation rates than White pairs (referred to later as hypothesis 1). This hypothesis is

grounded in several historical and sociopolitical considerations, the most important of which is


5 All hypotheses are formally presented in Chapter 4 (page 45).









the Black subjugation and response to White racism. White racism as defined by Feagin, Vera,

and Batur (2001) can be viewed as the "socially organized set of practices, attitudes, and ideas

that deny blacks and other people of color the privileges, dignity, opportunities, freedoms, and

rewards that this nation offers to White Americans" (17).

As Simmel (1955) noted in his classic work Conflict and the Web of Group Affiliations,

threats to a group (such as White racism) have the function of increasing group solidarity and

causing group members to transcend individual interests in favor of the moral imperative to

defend the group. As the level of violence against a group escalates, so does the group's internal

solidarity. Assensoh (2000) sums up this view, stating "nothing is known to bind and promote

both collaboration and cooperation more than years and shared experience of fighting common

and easily identifiable enemies" (115). Examples of this increased solidarity and intra-group

cooperation among Blacks reacting to the social dilemma of White racism can be found in the

work of McAdams (1995) and Chong (1991). Both scholars note that various forms of social

protest employed by Blacks during the civil rights movement (such as demonstrations, sit-ins and

boycotts) were all extremely costly to the individuals who participated in them, though it was

known to all that any gains made by the protesters would benefit all Blacks in the community

whether they participated or not.

To explain this phenomenon, Chong contends that social incentives such as increased

social standing and avoidance of ridicule and ostracism for not participating worked as the

impetus for participation in social protests among small social groups like Black Churches.

Though social incentives also influence racial groups other than Black Americans, America's

racist past has provided Black Americans (compared to White Americans) much more

motivation to sacrifice individual gain for the good of the community. I contend that the Black









history of subjugation to slavery, Jim Crowism, and the continuing racial discrimination faced by

Blacks today, has created a culture of cooperation among Blacks that is rooted in what Dawson

(1994) argues is a collective historical memory of racial oppression and sense of shared racial

fate. He writes, "[T]he collective memory of the African American community continued to

transmit from generation to generation a sense that race was the defining interest in the

individual's life and that the well-being of blacks individually and as a group could be secured

only by social and political agitation... [that memory] has been reinforced historically" (Dawson

1994:58). Since individual participation in "social and political agitation" can itself be viewed as

a social dilemma situation, it is reasonable to suggest that this collective memory may also

increase intra-group cooperation among Blacks. This outcome is also supported by what

McCormick and Franklin (2000) have called the inclusionary dilemma, "a dialectical process

wherein emergent racial consciousness on the part of African Americans can be seen as response

to the racial impediments that African Americans encounter as they attempt to gain access to the

economic, social, and political opportunity structure of the larger society" (330).

In contrast to Blacks, Whites have no equivocal long-standing historical memory of

oppression and racism in the United States (Rodgers 2000). Moreover, in a nation where

Whiteness is taken to be the norm and therefore becomes invisible, it is no surprise that Blacks

tend to be more conscious and aware of their racial group status than are Whites (Doane 2003).

What past research has failed to examine is whether or not that awareness will translate into

increased racial group cooperation in social dilemmas.

Theoretical Basis for Greater Cross-Racial Group Defection

In addition to the obvious and very pressing problems of racism, prejudice, and

discrimination that plague American society and that may manifest themselves as increased









defection in a cross-racial pairing during a social dilemma game, there is also significant

historical and sociological support for greater cross-racial group defection between Black and

White actors (referred to later as hypothesis 5). American labor history has an array of examples

where when faced with a social dilemma, White workers chose defection (and its benefits) over

cooperation with Blacks and/or other minority groups (Feagin 2000). Though cooperation by

White workers would have significantly increased the likelihood of higher wages for all workers

White and Black, defection maintained what Du Bois (1935) and later Feagin (2001) describe as

"the psychological wage of whiteness" (Du Bois:700; Feagin:27).

The psychological wage of whiteness as presented by Feagin (2001) is a complicit

support for a White dominated racially hierarchical society in which working class Whites

receive superior privileges, opportunities and cultural resources than do minorities, at the cost of

lower wages from the White elites who employ both. In other words, White workers choose the

benefits of being White, over "the greater economic wages they might have had if they had

joined in strong organizations with black workers" (Feagin 2001:30). Historical examples of this

social dilemma can be found at numerous times in American history: in Iron Cages, Takaki,

(2000) recounts how the 1870's strike by The Secret Order of the Knights of Crispin, an all-

White labor union founded in 1867, failed because of the importation of Chinese laborers to

replace the White union members. By excluding minorities from their labor union and

workplace, white laborers did not improve their bargaining position, but detrimentally

undermined it. By the time the Crispins tried to recruit Chinese workers into their union it was

too late; the strike would soon end with the union having lost (Takaki 2000:239-40). In this

example, racial stratification was the primary stratification because it "decentered class

oppression in the thinking and orientation of most White workers" (Feagin 2001:31). This









emphasis on race was of course the paramount goal of the White industrial elite who wanted

White workers to buy into a racist ideology and its psychological wages, so as to divert attention

from the fact that these elites were exploiting both White and Black workers (Feagin 2001).

Labor movements, however, are not alone in their choices to exclude minority

participation in favor of psychological wages and privilege. The U.S. women's rights movement

of the late 19th century "dismissed the concerns of black women and developed segregated

interests and organizations" (Feagin 2000:31). Instead of working for the betterment of women

of all racial groups, this movement's negation of minority women's issues undermined its

objective of equality, fair pay and women's rights. Though since the civil rights movement, the

psychological wage of Whiteness and White privilege have become less dejure and overt, they

nevertheless still exist, albeit in more de facto and hidden forms (Feagin 2000; Bonilla-Silva

2003). The existence of both the psychological wage of Whiteness and of White privilege may

make Whites more likely to cooperate with each other and less likely to cooperate with Blacks in

a social dilemma.

Also relevant to the hypothesis that there will be greater cross-racial group defection in a

social dilemma is the statistical theory of discrimination, which posits "that people use race as a

proxy to evaluate others by assuming in the absence of more specific information that the person

has the average qualities of her race" (McAdams 1995:1021). Considering that negative

stereotypes of Blacks are broadcast to American television and movie viewers on a daily basis, it

would not be surprising to find that in our highly segregated society, many White people reify

these stereotypes and take them to be true, therefore making them less likely to cooperate with a

Black person about whom little is known (Hacker 1995).









Conversely, among White dyads facing a social dilemma, what Vera and Gordon (2003a)

have identified as sincere fictions of the white self may lead to increased cooperation (15).

Sincere fictions of the white self are "deliberately constructed images of what it means to be

white" (Vera & Gordon 2003a:15). Many television programs and films are "sincere" fictions in

so far as "they are rooted in the self-concept that we seldom examine, that we take for granted"

(Vera & Gordon 2003a: 16). In other words, these fictions are embraced without considering real

or potential alternatives to them. The "white self' is the "concept of white Americans proposed

by white American moviemakers," who present Whites as morally and intellectually superior

persons who are "powerful, brave, cordial, kind, firm, good-looking, generous [and] natural born

leaders" (Vera & Gordon 2003b: 114). Since people define themselves in relation to others and

considering that many Whites internalize these sincere fictions as true, it would not be surprising

to see Whites more likely to cooperate with one another in a social dilemma, than with a Black

person. As social science research continues to show, "groups matter... people have a loyalty to

groups that goes beyond what serves their narrow pecuniary self-interest" (McAdams

1995:1084).

Alternative Theoretical Arguments: Against Differing Cooperation Levels

Despite the theoretical and historical support for differing levels of social dilemma

cooperation by race, there is also theoretical support to the contrary. The argument for different

levels of cooperation hinges on a group interest versus individual interest argument. Simply put,

for reasons described above, group interest is more likely to affect the social decision making

process in favor of cooperation for Blacks than it is for Whites. That being said, it is difficult to

create experimental designs in which pairs of participants take racial-group identity and









allegiance into consideration without overtly or covertly revealing to them that their decisions

will be compared to that of participants from different racial groups.

As McAdams (1995) reports, in same-race pairs in which participants know they are

being compared to pairs of another racial group, the proxy effect encourages cooperation since

participants may speculate that they will be evaluated in pairs and thus go for the best pair

outcome cooperation. Furthermore, Blacks may experience a stronger desire than Whites to

achieve the best pair outcome in order to combat racial stereotypes and racist theories of

inferiority. Not revealing that they are being compared to other racial groups protects against

this bias, however, it also poses the risk that racial-group interest will not be seen as relevant to

the decision making process. It is for this reason that the social location of the study and its

participants is of critical importance. Factors such as age, education, income, geographic

location, and experiences of racism can all influence the extent to which group interest is seen as

important or relevant to the decision to cooperate/defect. The role of social desirability should

also not be discounted. Participants desire to be perceived by both their game partner and the

researcher as good, nice, and non-racist people may also increase their likelihood of cooperation.

Another theoretical consideration that could negate any racial-group differences is the

idea that White racism is also a form of cooperation. So while I have proposed that Blacks are

more likely to have higher cooperation rates as a result of responding to White racism, one could

contend that Whites have forged an identity and cooperative spirit based on centuries of

practicing white racism in both its dejure and de facto forms.

Also pertinent is what critical theorist Derrick Bell (1980) calls the Interest-Convergence

Hypothesis. According to Bell, "whites support the cause of equality and justice for blacks only

when it is in i1heir interest to do so" (Bell 1980: 252). In a social dilemma such as the Prisoner's









Dilemma, White participants could easily see it as in their best interest to cooperate with Black

participants for a mutually beneficial outcome. If this is the case, then different levels of

cooperation will be unlikely.

Finally, theoretical support against lower cross-racial group cooperation levels and larger

cross-racial group defection rates can be found in Wilson's (1978) declining significance of race

argument. In his classic work The Declining Significance ofRace, Wilson contends that in the

modern industrial period, Black subordination and social advancement are more associated with

economic class, than with race. Following this argument, cooperation/defection in a social

dilemma could be more closely associated with class status than with race, and any measures of

cooperation should closely control for the class status of the participants.

One final concept that is germane to the issue of social dilemma cooperation/defection is

that of White racism as societal waste. Feagin (2001) notes that the tremendous amount of

talent, energy, and resources that are spent in perpetuating White privilege are an extravagant

waste that ultimately benefit no one and which have the potential to destroy society. Similarly,

individual and mutual defection in a social dilemma ultimately results in a waste of resources

that in the long term benefits no one.

In the plethora of social issues that can be viewed through the lens of a social dilemma,

when defection becomes the rule, everyone loses. It is for this reason that understanding if and

why different racial groups are more likely to defect or cooperate is important.

Gender and Cooperation

In contrast to the literature that examines racial cooperation in the Prisoner's Dilemma,

the literature that examines gender cooperation in the Prisoner's Dilemma is vast; it is also often

times contradictory. Prisoner's Dilemma gender cooperation results vary according to the time









the study took place, the geographic location of the study, as well as the study's design. Hence,

certain studies have found evidence that females have higher cooperation rates (see Sibley, Senn

& Epanchin 1968; Fisher & Smith 1969; Bonacich 1972; Smith, Vernon & Tarte 1975; Dawes,

McTavish & Shaklee 1977; James, Soroka & Benjafield 2001; Hu & Liu 2003), while others

have found support for the converse (see Kahn, Hottes & William 1971; Kahn, Hottes & Davies

1971; Sell & Wilson 1991; Brown-Kruse & Hummels 1993). Most studies, however, have found

no significant differences for the cooperation rates of women versus those of men (Caldwell

1976; Goering & Kahn 1976; Orbell, Dawes & Schwartz-Shea 1994; Sell 1997).

The following section will review some of the theoretical support for the hypothesis that

women will have higher cooperation rates than will men in a monetary Prisoner's dilemma.

Theorizing Gender Cooperation

Gender as a Social Construction and Social Structure

In The Social Construction of Gender, Lorber (200 la) notes that "gender is so pervasive

in our society that we assume it is bred into our genes" (as cited in Shaw 2001:121). Gender,

however, is not a genetic attribute; it refers to "the personal traits and social positions that

members of a society attach to being female or male" (Macionis 2005:325). Furthermore,

conceptions of gender norms are fluid, varying, and changing across cultures and over time;

gender is constantly created and reproduced through human interaction a process commonly

referred to as "doing gender" (West and Zimmerman 1987). At the individual level, gender

construction begins at birth (and possibly even before, thanks to medical technologies such as

sonograms) when newborns are assigned a sex based on the appearance of their genitalia. The

newborn's sexual classification quickly becomes a gender status through naming, dressing, and

the use of other gender indicators (Lorber 2001a). People respond to these indicators according









to gendered norms and the child soon learns and internalizes their assigned gender identity and

begins routinely doing (performing) gender for themselves (Butler 1990; Messner 1990). At an

early age gender identity-construction is accomplished by the individual who then, according to

West and Zimmerman (1987), becomes hostage to its production.

In addition viewing gender as a social construction, it can also be seen as a social

structure. Risman (2004) makes the case for conceptualizing gender as a social structure6 (others

such as Patricia Martin (2004), prefer the term institution) and using this conceptualization as a

tool for analyzing the ways in which gender is grounded in the individual, interactional, and

institutional dimensions of society. Focusing on gender as a social structure also suggests that

gender can be a constraint that opposes individual motivation (Blau 1977). Gender is

structurally buttressed through gendered norms and expectations that are "enforced through

informal sanctions of gender-inappropriate behavior by peers and by formal punishment by those

in authority should behavior deviate too far from socially imposed standards for men and

women" (Lorber 2001a:123).

The function of gender categories is (according to some feminist scholars) to keep

women subordinated through a system of stratification so that they can be exploited in both the

marketplace and the household (Lorber 200 1b). Despite the glaring inequality gender

classifications produce, they persist because they are deeply embedded as a basis for

stratification in our individual personalities, cultural norms, and societal institutions (Risman

2004). Thus, the status inequalities that gender divisions create can potentially manifest

themselves in an individual's social decision making.




6 Social structure is defined as "any relatively stable pattern of social behavior" (Macionis 2005:655).









Gender Cooperation and Social Decision Making

Many scholars have employed social learning theories, such as Carol Gilligan's (1982)

work on differences in the moral development of women and men, as the foundation from which

to project differences in cooperation in social decision making games such as Prisoner's

Dilemma (Brown & Taylor 2000; Stockard, van de Kragt & Dodge 1988). According to

Gilligan, boys are taught to hold a "justice view" of morality that emphasizes formal rules and

the rights of individuals under those rules, while girls (because of socialization) tend to favor a

"care view" moral perspective that emphasizes solidarity, community, and close relationships.

Additional theoretical support for increased cooperation rates among women can be found in the

work of Janet Lever (1978), who after examining single-sex play groups found boys favor games

that nearly always have winners and losers; these games reinforce masculine traits of aggression

and competition. Conversely, girls are more likely to play games (such as jump rope, hopscotch,

dancing, and singing) where victory is not the ultimate goal and interpersonal skills of

communication and cooperation are fostered. This finding is important considering the

construction of the Prisoner's Dilemma game and it's flexibility to end in either a draw or with a

clear winner and loser. Applying Lever's conclusion to Prisoner's Dilemma suggests that

women are more likely to cooperate and play for the draw that most benefits both players, and

that men are more likely to attempt to out-compete one another by defecting.

As theoretical foundations for hypothesizing gender differences in social decision making

scenarios the above explanations fit nicely with the social construction and social structure

theories of gender described earlier. Though gender is a social construction, it is nonetheless

considered as a real and meaningful distinction by most people. Since, as discussed earlier,

"situations that are defined as real, are real in their consequences," the societal belief that women









place more value on interpersonal relationships, friendships, solidarity, and cooperation than do

men who in turn are believed to favor aggression, domination, and competition becomes a self-

fulfilling prophecy (Thomas 1931:145). In other words, since as women and men we are

expected to act in certain gendered ways, we internalize that expectation and act accordingly.

Similarly, gender as a social structure can both act as a constraint that tempers individual

motivations to cooperate/defect according to socialized gender roles or as a prescribed response

to the cooperate/defect decision.

Brent Simpson (2003), however, argues that previous research theoretically grounded on

work by Gilligan (1982) and other feminist scholars often fails to find gender differences in

cooperation levels because researchers have consistently used the Prisoner's Dilemma game to

investigate if a difference exists. Simpson contends that the Prisoner's Dilemma is problematic

because several theories of gender/sex related behavior support the notion that "females are more

likely to defect out of fear (the prospect that one's cooperation may be exploited) while males are

more likely to defect out of greed (the temptation to free-ride on others' cooperation)" (Simpson

2003:36). Since the Prisoner's Dilemma contains both fear and greed, Simpson argues that one

should not expect to observe sex differences in this classic greed and fear version (his study

supports this hypothesis). Instead, Simpson predicted that in a dilemma that emphasized the fear

component, fewer females than males would cooperate. In Simpson's study, this was not

supported by the evidence since there where equal levels of cooperation among men and women.

Lastly, in a dilemma that emphasized greed, Simpson predicted males would be less likely than

females to cooperate. In this greed dilemma study, he found that 57% of females versus 33% of

males cooperated, thus supporting his hypothesis. In a repeat of this study using similar methods,

Simpson found similar results.









Simpson brings to light some interesting considerations; his work shows that for all but

the original (greed and fear) version of the Prisoner's Dilemma, women in his study cooperated

more than men despite the fact that their cooperation levels often did not reach the level of

statistical significance. This begs the question of whether a larger and broader sample would

have yielded differing results. Also questionable is whether Simpson's tweaking of the

Prisoner's Dilemma to create a fear dilemma and a greed dilemma is valid and is actually

measuring fear and greed influences on cooperation levels. Since he does not ask participants

the reason for their defection, we can only theorize as to their real motives. In contrast, the open-

ended questionnaire component of the present study allows for an examination of the reasons and

rationale participants give for cooperating and defecting.

A final important factor that is present in the work of Simpson and many others

employing the Prisoner's Dilemma in their research, is that study participants are not allowed to

see or meet one another throughout the experiment. As a result, participants have no ability to

take racial and gender group membership into account when playing against a person they have

never seen. By allowing participants to see the person they will play the monetary Prisoner's

Dilemma game with before the game instructions are explained to each separately, this study

gives insight into whether race and/or gender are somehow being used as a proxy to gauge trust

and altruism.

It should also be noted that additional theoretical grounding for the hypothesis (referred

to later as hypothesis 2) that women pairs will have higher same-gender group cooperation rates

than male pairs in a Prisoner's Dilemma game, can be based on several historical and

sociopolitical considerations, the most important of which is the existence of patriarchy and

women's resistance to this social-structural system of domination. Hence, Simmel's (1955)









work on Conflict and the Web of Group Affiliations, can also provide theoretical grounding for

the view that patriarchy and sexism, if viewed as a group threat to women, may have the function

of increasing female group solidarity and fostering group members to transcend individual

interest in favor of the moral imperative to support the group. From this perspective, women's

subordinate social position will result in higher female cooperation levels rather than higher

female defection levels that are based on fear and self-protection, as is Simpson's contention.

Thus, in the case of the monetary Prisoner's Dilemma game I employ, increased female

cooperation theoretically hinges on whether women see themselves as a subjugated and/or

oppressed group.

In summation, the primary theoretical rationale past researchers have employed to

predict gender differences in the Prisoner's Dilemma has typically been that women's

socialization would result in higher cooperation rates for women relative to men (Stockard, van

de Kragt & Dodge 1988; Brown & Taylor 2000). I propose that in addition to socialization,

female experiences of subjugation under a patriarchal system may also motivate greater female

cooperation in the Prisoner's Dilemma. Inconsistent results and contradictory findings, however,

suggest that the question of whether gender affects cooperation in Prisoner's Dilemma and other

social decision making games remains unanswered (Ledyard 1995). By employing traditional

experimental and quantitative methods along with more qualitative methods (that allowed for the

analysis of open-ended responses to the question of why participants chose to cooperate or

defect) this study was able to further explore Prisoner's Dilemma gender differences and better

elucidate how gender influences the decision-making processes of participants.









The Intersection of Race and Gender

Various scholars (particularly Women of color) have pointed out that when considering

the system of gender stratification, it is valuable to consider various other dimensions of social

stratification such as race and class since all of these systems of oppression operate

simultaneously, intersect, and can have cumulative effects (Anderson and Collins 2001; Carby

2000; Collins 2000; Williams 2000; Baca Zinn 1996; Delgado 1995; Baca Zinn and Thorton

1994; Frankenberg 1993; Rothenberg 1992; hooks 1991; King 1988). In light of this, some have

noted that the theoretical divisions among scholars who specialize in race or class or gender or

queer theory have undermined a more sophisticated analysis of inequality (Reskin, 2002; Tilly,

1999). Others such as Collins (1998), Calhoon (2000) and Risman (2004) advocate a "both/and"

strategy that accents the need to "understand gender structure, race structure, and other structures

of inequality as they currently operate while also systematically paying attention to how these

axes of domination intersect" (Risman 2004:429). Though gender and race are social constructs

they are nevertheless universally used to rationalize inequality.

The implications of intersectionality research for Prisoner's Dilemma studies are that

racial and gender considerations are intertwined, and that participant's race and gender status

cannot be divorced from one another. In other words, Black women cannot respond separately

as Black and as women, but rather these two statuses work in tandem in shaping social

perceptions and responses. The race and gender status of a participant's game partner may also

influence social perceptions and expectations in ways that vary across race and gender

categories. As a result, there is a need to shift from research that generally examines race or

gender but not both together, to research that examines both race and gender and the intersection









of these statuses. By doing so we can learn if one status dominates or if the intersection of race

and gender creates unique perceptions and responses for different race and gender groupings.

In the following chapter, I present the methodology employed in this study and describe a

research design that was created to measure both race and gender group differences in Prisoner's

Dilemma cooperation/defection rates, while also addressing the need for intersectional

exploration.







































7 The intersection of race and gender compose the theoretical basis for hypotheses 3 and 4 presented in Chapter 4
(page 45).









CHAPTER 3
RESEARCH METHODOLOGY

Research Design

In order to examine differences in levels of cooperation between Blacks and Whites and

women and men in a social dilemma, an experimental design based on a monetary version of the

Prisoner's Dilemma game was created. This specific research design makes it possible to test the

following key questions:

1. Is there a correlation between race and likelihood of cooperation in a social dilemma?

2. Is there a correlation between gender and likelihood of cooperation in a social dilemma?

3. What are the reasons and rationale given by participants for choosing to cooperate or defect?

4. Do levels of trust and expectations of a game partner's decision differ between race and
gender groups?


To investigate these questions a sample with the following race and gender composition was

recruited:

36 White males
36 White females
36 Black males
36 Black females

This created a total sample size of 144 participants, split into the following six experiment

groups:

1. White male group (n = 24)
2. White female group (n = 24)
3. Black male group (n = 24)
4. Black female group (n = 24)
5. Black and White male cross-race group (n = 24)
6. Black and White female cross-race group (n = 24)









The purpose of creating these experiment groups and dividing participants accordingly was to

allow the researcher to compare and contrast the cooperation/defection decisions and rationale of

participants in a given group with those of participants in different groups. Participants in the

study were scheduled in dyads and played a one-shot game of the Prisoner's Dilemma with a

member of their same-gender group. For race, both same-group and cross-race group differences

were measured and analyzed. Cross-gender group differences were not examined due to

financial constraints and the multitude of previous studies that have explored this question.

At the beginning of the experiment (after participants read and signed an informed

consent form8), participants were briefly allowed to see one another and told that they would

both soon be participating in a social decision making game. They were then separated and told

that the person they would be playing the game with was the same person they just saw and that

they would leave the experiment in isolation (and not encounter the person they played) once the

game was complete. Next, the participants were escorted to separate rooms where they received

an extensive briefing on a monetary Prisoner's Dilemma game until the researcher was confident

that they understood the various possible choices and the consequences and payoffs of each

combination of choices. During this time they were told that if both players cooperated they

would each receive $6. If they both defected, they would each receive $3, and if one cooperated

and the other defected, the person who cooperated would receive $0 while the person who

defected would receive $9. Like the classic version of the game, which potentially punishes

cooperation9 with a nine-year prison term, the prospect of not receiving compensation for your

time in this monetary Prisoner's Dilemma game served to temper players desire to cooperate.




8 The informed consent form is located in Appendix B
9 As mentioned in Chapter 1, cooperation in the classic Prisoner's Dilemma scenario refers to not confessing to the
crime and thus cooperating with the co-criminal.









Following the briefing, participants indicated their cooperation or defection by circling

either "cooperate" or "defect" on a game sheet, which was located on the first page of a

questionnaire that was given to them during the briefing period (see appendix A). After

indicating their cooperation/defection game decision, participants then filled out the

questionnaire by responding (in writing) to a series of open-ended questions about their game

decision and providing demographic information. Upon learning the cooperation/defection

decision of both players the researcher compensated each player accordingly. The researcher

then debriefed both players individually to prevent any possible animosity resulting from

defection during game play.

Measures and Instruments

The data for this study came from two sources: (1) participants recorded decision to

cooperate or defect and (2) a questionnaire that asked participants about their demographic

characteristics and the following open ended questions10:

Why did you cooperate or defect? Please explain in detail.
Do you feel you can trust the other player? Why or why not?
Do you believe the other player will cooperate? Why or why not?

The participants recorded decisions to cooperate or defect were entered into a statistical

analysis computer program, Statistical Package for the Social Sciences (SPSS), which was used

to calculate whether differences between groups were statistically significant (greater than zero)

and whether the study's hypotheses (presented in Chapter 4) were supported by the data. The

researcher coded the open-ended questionnaire responses according to any common themes that

emerged. Of particular interest were the reasons and rationale put forward by participants for

cooperation/defection, and how these differed between and within the four categorical gender

and racial groups and the two cross-race groups.









The strengths of this study are that: (1) it examines if there is a correlation between race

and cooperation. This is major gap in the social dilemma literature and this study is a step

toward filling it; (2) Unlike many studies incorporating the Prisoner's Dilemma, this study uses

a one-shot version of the game that guarded against revenge, reputation and changes in strategy

affecting cooperation rates; (3) Incorporating an open-ended survey component allows for the

examination of why participants cooperate or defect in addition to the data on their actual

cooperation and defection rates; and (4) Unlike most previous research that has tended to

exclusively measure the likelihood of cooperation, irrespective of the gender and/or racial group

membership of each individual and pair of participants, this study specifically examines intra-

group and cross-group cooperation rates.

Experimental Design

According to Bernard (2000), the experimental method "is not one single technique but

an approach to the development of knowledge" (139). Experiments are often used by social

science researchers to control the effects of confounds to understand the effects of a particular

independent variable. However, since this experiment (like most) entailed investigating the

effects of several independent variables at once, a factorial design that allows for the systematic

analysis of each independent variable was most appropriate. Moreover, factorial design made it

possible to measure interaction effects on dependant variables (such as cooperation/defection)

that occur as a result of interaction between two or more independent variables (such as race and

gender).

The strength of using this method is that it enabled the researcher to see if there are any

statistically different levels of cooperation/defection both among and between racial and gender

group pairings. Additionally, a general benefit of factorial experimental designs is that they can


10 The questionnaire can be viewed in Appendix A.









be replicated, thus augmenting the confidence that can be placed in the findings' internal validity

and generalizability (Babbie 2001). Perhaps the greatest strength of this method, however, is

that a controlled experiment can measure social dilemma cooperation in an isolated environment

and under such conditions that would be difficult, if not impossible, to observe otherwise.

A general limitation of the experimental method used is its artificiality (Babbie 2001). In

other words, what happens in a controlled setting may not be reflective of what happens in more

natural social encounters. Any study concerning social dilemma cooperation/defection should

take this into account. In the case of social dilemma cooperation/defection the possibility of

spurious correlations were also of concern. Factors other than race or gender, such as income,

class, and/or religiosity may be at the root of greater group cooperation/defection and these were

statistically controlled for when appropriate.

Recruitment of Participants

Participants for this study were Black and White undergraduate students attending a

major public southeastern university. They were recruited through the use of flyers that were

posted on campus and distributed in general elective introductory level sociology classes.

Recruitment flyers stated that participants could earn from $0-$9 dollars for playing a social

decision making game that would take between 15 and 20 minutes to complete. The recruitment

flyers also stated that the study was seeking African-American and White non-Hispanic

participants.

Demographic Characteristics of the Sample

Participation in the study was restricted to self-described Black and White students to

examine if race (or perhaps racism) influenced a participant's decision to cooperate or defect.

Due to the systemic form of White-on-Black racism that has persisted in the U.S. from the 17th









century until the present, any racial group differences in social dilemma cooperation resulting

from subordination are most likely to be evident in same-race and cross-race pairings of Blacks

and Whites. Latinos were not recruited 11 for the sample because they trace their origins to a

multitude of countries including Mexico, Puerto Rico, Cuba, Dominican Republic, as well as all

the Spanish speaking countries of Central and South America, each of which have their own

differing systems of racial hierarchy and domination. Moreover, racially "Latinos are white,

black, indigenous, and every possible combination thereof' (Suarez-Orozco & Paez 2002:3).

A general weakness of this study is that the sample was limited to undergraduate college

students. Participants for the study ranged from 18-28 years of age. The mean age was 20 with

a standard deviation of 1.63. Also, the majority of participants reported coming from a middle-

class, upper-middle class, or upper-class background with 80.5 percent reporting a family

income of $41,000 or higher. Fifty-five college majors were represented with sociology being

the most common major (n = 25). In these characteristics, the sample is reflective of the typical

ages, social class positions, and variety of educational majors of traditional college students and

therefore is susceptible to any biases present in this population. For example, studies on aging

employing the Prisoner's Dilemma provide evidence that cooperation rates may be higher for

children than for adults (Sjoberg, Bokander, Denick, & Lindbom 1969). Social class may also

influence cooperation and researchers have shown that college major can be correlated to

increased or decreased cooperation rates (James, Soroka, & Benj afield 2001; Hu & Liu 2003).

By only studying college undergraduates, this study did not address the question of whether

having a broader and more diverse sample population would have yielded different results.



" A few participants who considered themselves White non-Hispanic, but who indicated they had a Latino ancestor
on the questionnaire were permitted to participate. Similarly, participants who considered themselves African
American, but indicated that they had ancestors from the Caribbean were also permitted to participate.









More specific weaknesses of this study's sample include the large contingent of

sociology majors (n = 25) who participated, the use of sociology facilities to conduct

experiments, and the participation of approximately 40 students who were enrolled in a course

taught by this researcher. To combat possible influences on responses that a participant's

connection to the sociology department and/or to this researcher may have, it was explained to

participants (and explicitly stated in the informed consent form) that their responses would be

anonymous to all but the researcher and that the responses would have no bearing on any course

grade.

Analyzing the Data

Quantitative procedures for analyzing the data in this study included cross-tabulations,

bivariate and multivariate analysis, analysis of variance and logistic regression. Qualitative

procedures included a content analysis of the open-ended responses to the question "Why did

you cooperate or defect? Please explain in detail." Questionnaire responses were coded

according to various thematic categories and compared within and between the six experiment

race and gender group categories. In addition to being qualitatively coded, responses to the

questions "Do you feel you can trust the other player?" and "Do you believe the other player will

cooperate?" were quantitatively coded and assigned into one of the three following categories:

(1) mostly yes; (2) mostly no; and (3) uncertain. Statistical procedures to analyze race and

gender group differences to the responses to these questions were then conducted using SPSS.

Role of the Researcher

The researcher/social dilemma game facilitator attempted to remain as neutral to the

process as possible. He employed the same verbal script to explain the rules of the game to each

participant and provided clarification when needed. Furthermore, he made sure that participants









were familiar with all of the possible outcome scenarios and the corresponding pay-out for each

one.

At the conclusion of the game he paid participants accordingly and debriefed each one

individually. He also asked participants not to reveal the dynamics of the game to any of their

peers who might participate in the study. At the games conclusion he made sure participants left

their separate rooms at staggered times so as not to encounter one another on the way out.

Finally, he made every effort to maintain the same professional demeanor throughout each game

and debriefing period.

For the purpose of reflexivity, the researcher would like to disclose that he is a Latino,

male, graduate student, 28-years of age and that his presence in itself may have influenced the

outcome of the study in unforeseen and unknown ways.









CHAPTER 4
QUANTITATIVE RESULTS

Hypotheses

In consideration of the first two research questions, (1) Is there a correlation between race

and likelihood of cooperation in a social dilemma? and (2) Is there a correlation between gender

and likelihood of cooperation in a social dilemma? outlined in Chapter 1 (page 12), I posit the

following five hypotheses which are theoretically grounded in the works of Gilligan (1982),

Chong (1991), McAdams (1995), James, Soroka and Benjafield (2001), and Hu and Liu (2003)

among others.

Hypothesis 1: Black participants will have higher intra-group cooperation rates than
White participants.

Hypothesis 2: Women will have higher intra-group cooperation rates than men.

Hypothesis 3: Black women will have the highest cooperation rates of any of the four
same-race and same-gender experiment groups.

Hypothesis 4: White men will have the lowest cooperation rates of any of the four same
race and same-gender experiment groups.

Hypothesis 5: same-gender cross-racial group pairings will have lower cooperation rates
than the intra-group pairings of both Blacks and Whites.

The following sections discuss the quantitative results of the study, which yield support for

hypotheses numbers one and three. The findings from a quantitative analysis of trust and

expectations are also discussed, as is the role of religiosity in social dilemma decision making.

Quantitative Findings

Cooperation

Tables 4-1 and 4-2 display the cooperation rates and distribution of outcomes for each of

the six experimental groups. Hypothesis 1 predicts Blacks will have higher intra-group

cooperation rates than Whites in the monetary Prisoner's Dilemma. Concordant with this









hypothesis, 69% of Blacks and 50% of Whites in same-race groups cooperated. This difference

in same-racial group cooperation approached statistical significance (p = .061) (Table 4-3).

Additionally, if one also takes into account the participants in cross-racial group experiments, the

overall cooperation rate is 72% for Blacks and 54% for Whites. This difference in Black and

White cooperation is significant (p = .025) (Table 4-4).

Hypothesis 2 predicts that women will have higher intra-group cooperation rates than

men will have in the monetary Prisoner's Dilemma. Contrary to this hypothesis, 61% of women

and 65% of men cooperated (Table 4-5). This difference was not significant (p = .604) and

indicates virtually no difference between female and male cooperation rates. Furthermore, there

were also no statistically significant differences between females and males on the variables trust

in partner and expectation ofpartner 's decision to cooperate or defect. These findings are

concurrent with the majority of previous studies that have found no significant difference in the

cooperation rates of women compared to those of men (see Caldwell 1976; Goering & Kahn

1976; Orbell, Dawes & Schwartz-Shea 1994; Sell 1997).

Hypothesis 3 predicts that Black women will have the highest cooperation rates of any of

the four same-race and same-gender categorical groups. Supporting this hypothesis, the

cooperation rates for same-race and same-gender group pairings are as follow: Black women

75% cooperation; Black men 62% cooperation; White men 54% cooperation; White women 46%

cooperation (Table 4-1). Though the differences between the cooperation levels of all the same-

race and same-gender groups is not statistically significant (p = .202), the cooperation level

difference between the Black female group and White female group is significant (p = .039).

Hypothesis 4 suggests White men will have the lowest cooperation rates of any of the

four same-race and same-gender categorical groups. This hypothesis was not supported.









Although the White male group had the lowest cooperation rate of any of the categorical groups

besides White women, the differences between the White male group's cooperation rate and

most of the other groups was not statistically significant.

The final hypothesis predicts that same-gender cross-racial group pairings will have

lower cooperation rates than the same-gender same-race group pairings of both Blacks and

Whites. Surprisingly, the exact converse proved true in that the male cross-racial group pairings

had the highest cooperation rate (79%) of any group and the female cross-racial group pairings

cooperation rate (63%) tied for third (out of the six groups) with the Black male group (Table 4-

1). This finding contradicts the findings of earlier research on how race affects cooperation in

the Prisoner's Dilemma (see Sibley, Senn & Epanchin 1968; Wilson & Kayatani 1968;

Cederblom & Diers 1970; Baxter 1973). On the other hand, this study corroborates the more

recent cross-racial group cooperation findings ofHieder and Skowronski (2007).

Additionally, this study demonstrated that male cross-racial group pairings had

cooperation rates that approached significantly greater levels than White male group pairings

(p=.066) and that were statistically greater than the White female group pairings (p = .017).

These findings run counter to those of Dowing and Bothwell (1979) and suggest that either racial

discrimination in social dilemmas is subsiding or that the desire to appear non-racist is leading to

increased cross-racial cooperation.

Finally, it should be noted that overall Black and White differences in cooperation (that

indicate that Blacks have higher cooperation levels than Whites irrespective of race of game

partner) remain significant even in logistic regression models that control for the influence of

gender and family income (p = .044).









Quantitative Analysis of Trust and Expectations

The quantitative component of this study also revealed strong relationships between

whether a participant trusted their game partner, their expectation that their game partner would

cooperate, and their decision to cooperate or defect (Table 4-6). Of the participants who

indicated in the questionnaire that they mostly trusted their game partner, 86% cooperated

whereas 37% of those who mostly did not trust their partner cooperated. Of the participants who

indicated they where uncertain as to whether or not they trusted their partner, 64% cooperated. In

this, cooperation levels for participants who identified themselves as uncertain more closely

resembled that of participants who trusted their partner than those who did not trust their partner.

The relationship between trust of game partner and cooperation is significant at thep < .001

level.

Expectation of game partner's decision was also a significant predictor of cooperation

(Table 4-7). Participant's who mostly believed their game partner would cooperate had a

cooperation rate of 82%. Conversely, participants who mostly believed their game partner

would defect had a cooperation rate of 25%. Of the participants who were mostly uncertain as to

whether their partner would cooperate or defect, 68% choose to cooperate. The relationship

between expectation of partner's decision and cooperation is significant at the p < .001 level.

Additionally, there is a significant (p < .001) relationship between whether a participant trusts

his/her game partner and his/her expectation that the game partner will cooperate.

Also of note, when excluding participants who indicated they were uncertain as to

whether their game partner would cooperate or defect, statistically significant (p = .051)

differences between Black and White participants became visible (Table 4-8). Seventy-two

percent of Blacks believed their game partner would cooperate compared to only 56% of Whites.









A breakdown of participant's expectation of game partner's decision by race and gender is

shown on Table 4-9.

The Role of Religiosity

Of the variables measured, the only one other than racial variables to correlate with

cooperation was religiosity. The religiosity variable was originally coded with three categorical

options for the response to the question "How religious are you?" These options were (1) below

average: (2) average; and (3) above average. Later this data was collapsed into the categories

"less than average religiosity," and "average or higher than average religiosity" in order to better

explore how the presence and absence of religiosity is correlated with social dilemma

cooperation and defection. In this, participants who identified as having less than average

religiosity had a cooperation rate of 50%, whereas those who identified as having average or

higher than average religiosity had a cooperation rate of 69% (p = .029) (Table 4-10).

Religiosity was also associated with trust in partner. Although participants who

identified as having less than average religiosity were just as likely (39%) to trust their partner as

participants who identified as having average or above average religiosity, the former group was

more likely (45% versus 30%) to indicate that they did not trust their partner. Furthermore,

participants with below average religiosity were half as likely to indicate uncertainty as to

whether they trusted their partner (16%) than where participants who identified as average or

above average religiosity (32%). These measures approached statistical significance at thep =

.069 level.

While religiosity and race are clearly associated to cooperation levels in the monetary

Prisoner's Dilemma, the quantitative results of this study are complicated by an intersection of

race and religiosity (see tables 4-11 and 4-12). Eighty-nine percent of Black participants









indicated having average or above average religiosity compared to 50% of White participants.

This difference in religiosity was significant at thep < .001 level. The higher level of Black

participant religiosity was also evident in the six experiment racial/gender groups with the Black

female group and Black male group tied for the greatest level (88%) of group participants in the

average and above average religiosity category. Differences in the six racial/gender groups'

religiosity levels are significant at thep < .002 level. In logistic regressions that incorporate

religiosity and race, race as a predictor of cooperation is no longer a significant measure (p =

.136). Therefore, it is possible that cooperation level differences by race are a reflection

differences in religiosity and vice-versa; however, further exploration of this intersection is

needed in order to draw a more definitive conclusion. It should also be noted that reported levels

of religiosity correlated with cooperation rates in a similar manner for both Black and White

participants.

In terms of the religion of participants, 76% of Blacks identified as either Protestant,

Christian or both compared to 38% of Whites. Differences in religious affiliation between

Blacks and Whites was significant at the p < .001 level. While religious affiliation was not

significantly associated with cooperation rates or expectation of partners' decision rates,

participants who indicated they were either Protestant or Christian were more likely to trust their

partner (47%) than were non-Protestant and non-Christians (29%). This difference approached

statistical significance at thep = .081 level. Also of note, in logistic regressions that controlled

for gender and family income of participants, religiosity as a variable remained significant at the

p =.042 level.









Discussion of Quantitative Findings

The quantitative findings of this study provide support for the hypothesis that Black

participants will have higher same-race group cooperation rates than White participants in a

social dilemma. They also show that in this particular social dilemma, Black participants are

generally more likely to cooperate with a game partner (irrespective of that game partner's race)

than are White participants. In Chapter 2, I proposed that hypothesized racial differences in

cooperation rates would be based on Black subjugation and responses to White racism, as well as

greater Black racial identity awareness. However, this study's quantitative component measures

only if differences in cooperation exist, not what the reasons for these differences are. With that

said, the quantitative findings are in line with the presented theoretical arguments for greater

Black cooperation, although further analysis is needed (and will be presented in Chapter 5) to

verify the validity of these claims. It should also be noted, that racial differences in participants'

levels of religiosity may be the underlying cause of greater Black cooperation, thus presenting

the possibility that associations between race group and cooperation levels are the result of a

spurious correlation.

On the question of whether women have higher same-gender group cooperation rates

than men (presented as hypothesis 2), the quantitative findings indicate no significant evidence

of gender difference and thus are concordant with the majority of previous Prisoner's Dilemma

studies (see Caldwell 1976; Goering & Kahn 1976; Orbell, Dawes & Schwartz-Shea 1994; Sell

1997). This finding contradicts the view that gender socialization and the status inequalities that

gender divisions create (in the United States) will by themselves lead to greater female social

dilemma cooperation levels.









However, the finding (supporting hypothesis 3) that the Black female group had the

highest cooperation levels of any same-race experiment group while the White female group had

the lowest, suggests that the intersection of race and gender creates unique perceptions and

responses for different race and gender groupings. These differences in cooperation levels of

Black and White females may be a reflection of the different social positions and power these

women hold. As Lengermann and Wallace (1985) point out, White women are less likely to see

themselves as an oppressed minority group than are Black women because the former are well

represented on every level of U.S. social class structure, whereas the latter are found primarily on

the lower and middle levels. With regard to social dilemma cooperation, this suggests that if

White women do not see themselves as a minority group it is not likely they will have high

cooperation rates that are due to their shared experiences of minority group subjugation. This is

not to imply that gender group membership does play an important role in the

cooperation/defection decisions of some White women, but rather that the influence of gender

status on social dilemma cooperation may be different for White women than it is for Black

women.

The negation of hypotheses 4 and 5 further speaks to the complicated nature of the results

that the intersection of race and gender can create. In particular, the findings of higher than

expected levels of both male and female cross-racial group cooperation can be viewed in many

ways. These quantitative findings can lend support for Bell's (1980) Interest Convergence

Hypothesis, Wilson's (1978) declining significance of race argument, and a host of other

theoretical perspectives. In Chapter 5, I report and discuss the open-ended data findings of this

study, which help to assess and interpret the reasons behind the relatively high levels of cross-

racial group cooperation.









In light of the above findings it is important to note that one of the biggest challenges

faced by gaming studies, and more particularly the Prisoner's Dilemma, is questions about the

external validity of laboratory experiments and their relevance to real-life social dilemmas

(Nemeth 1972; Apfelbaum 1974; Pruitt & Kimmel 1977). As Pruitt and Kimmel (1997) note,

"most researchers who use games simply report their results with no attempt to speculate about

the real-life implications" (367). This is in part due to the view of certain researchers such as

Rapoport (1964) who feel laboratory results are chiefly valuable for the perspectives they bring

forth and questions they give rise to. Thus, they feel no need to generalize beyond the

laboratory. Other's, such as Pruitt and Kimmel (1997), feel it is preferable for researchers to try

to generalize their findings to settings with similar background conditions, however, side against

those, such as Rubin and Brown (1975), who attempt to generalize their game results to all

settings involving interpersonal dependence.

The current study utilizes race and gender theory as the basis for testable hypotheses and

employs content analysis in an effort to show the relevance of the findings to social dilemma

research without over generalizing. Hence, what can be said about the quantitative findings

presented above is that minority racial group membership and religiosity are positively correlated

with cooperation in the social dilemma game Prisoner's Dilemma. Additionally, it is reasonable

to suggest that these variables may also be correlated in settings involving interpersonal

dependence with similar background conditions. However, as some of the open-ended responses

that will be discussed in the next section show, the ideas of what qualifies as a similar condition

can vary greatly from person to person. In the following chapter, a content analysis of

questionnaire responses will help address some of the validity concerns about employing games

to study social dilemma cooperation.









Table 4-1. Proportion of cooperation and defection by race/gender experiment group


Participants race/gender experiment group
White male group(n = 24)
White female group (n = 24)
Black male group (n = 24)
Black female group (n = 24)
Cross-race male group (n = 24)
Cross-race female group (n = 24)
Total


% Cooperated
54.2
45.8
62.5
75.0
79.2
62.5
63.2


% Defected
45.8
54.2
37.5
25.0
20.8
37.5
36.8


Table 4-2. Distribution of game pair outcomes by race and gender experiment group
Experiment Group Cooperate/Cooperate Cooperate/Defect Defect/Defect


White Males

White Females

Black Males

Black Females

Cross-race Males


3 pairs

1 pair

5 pairs

7 pairs

8 pairs


7 pairs

9 pairs

5 pairs

4 pairs

3 pairs*


2 pairs

2 pairs

2 pairs

1 pair

1 pair


Cross-race Females 4 pairs 7 pairs* 1 pair
*In cooperate/defect category of the cross-race male group all defectors were White participants. In cooperate/defect
category of the cross-race female group 4 defectors were White participants and 3 were Black participants.

Table 4-3. Proportion of cooperation and defection in exclusively White and exclusively Black
experiment groups
Experiment group % Cooperated % Defected
Exclusively White groups 50.0 50.0
Exclusively Black groups 68.8 31.3
Chi-square = 3.498 p = .061 n = 96

Table 4-4. Proportion of cooperation and defection by race
Race of participants % Cooperated % Defected
Whites 54.2 45.8
Blacks 72.2 27.8
Chi-square = 5.046 p =.025 n = 144









Table 4-5. Proportion of cooperation and defection by gender
Gender of participants % Cooperated
Women 61.1
Men 65.3


% Defected
38.9
34.7


Chi-square = .269 p =.604 n = 144

Table 4-6. Proportion of cooperation and defection by trust in game partner
Participant trusts game partner % Cooperated % Defected
Mostly yes 85.7 14.3
Mostly no 36.7 63.3
Uncertain 64.1 35.9
Chi-square = 26.974 p < 0.001 n = 144

Table 4-7. Proportion of cooperation and defection by expectation of game partner's decision
Expects game partner to cooperate % Cooperated % Defected
Mostly yes 82.3 17.7
Mostly no 25.0 75.0
Uncertain 68.4 31.6
Chi-square = 39.944 p <0.001 n = 142

Table 4-8. Expectation of game partner's decision by race
Expects game partner to cooperate % White % Black
Mostly yes 55.7 72.6
Mostly no 44.3 27.4
Chi-square = 3.796 p = 0.051 n = 142

Table 4-9. Expectation of game partner's decision by race and gender
Expects game partner to cooperate % White % Black
Mostly Yes
Women 56.7 71.0
Men 54.8 74.2
Mostly No
Women 43.3 29.0
Men 45.2 25.8
Chi-square = 3.889 p = 0.274 n = 142

Table 4-10. Proportion of cooperation and defection by religiosity
Level of religiosity % Cooperated % Defected
Average and above 69.0 31.0
Below average 50.0 50.0
Chi-square = 4.742 p = 0.029 n = 144











Table 4-11. Level of religiosity by race
Level of religiosity
Average and above
Below average


% White
50.0
50.0


% Black
88.9
11.1


Chi-square = 25.658 p < 0.001 n = 144

Table 4-12. Level of religiosity by race and gender
Level of religiosity % White % Black
Average and above
Women 52.8 91.7
Men 47.2 86.1
Below average
Women 47.2 8.3
Men 52.8 13.9
Chi-square = 26.182 p < 0.001 n = 144









CHAPTER 5
OPEN-ENDED DATA RESULTS

Open-ended Data Findings

As stated earlier, the aim of the open-ended questionnaire component of this study is to

explore the reasons and rationale participants give for their decisions to cooperate or defect and

to see if there are general differences in the way different gender and/or racial groups explain

their choices. It is also to see if the common explanations given by participants differ when their

game partner is of a different racial group. To examine the written responses to the question,

"Why did you cooperate or defect? Please explain in detail," a coding system was developed

with 27 categories that captured the most typical responses (see table 5-1). These responses were

then compared within and between the six racial and gender experimental groups. Responses

that were atypical were also noted and compared in the same fashion. Responses to questions on

the variables of trust and expectations though mostly utilized for quantitative analysis were also

examined and are discussed here. It should be noted that women generally wrote longer answers

than did men and perhaps concomitantly gave more reasons for their decisions. Unexpectedly,

open-ended responses also revealed information about the validity of the game and how certain

participants equated this laboratory social dilemma to other social dilemmas they might

encounter in natural settings. In this chapter, I will first discuss the findings on validity and then

present a summary of the general reasons each of the six experimental racial and gender groups

gave for both cooperation and defection. Next, I will discuss experiment group differences to

questions three and four of the questionnaire. I will conclude the chapter with a discussion of

what these findings mean for social dilemma research.









External Validity

Although none of the questions on the questionnaire were specifically designed to test the

validity of the study, in their written responses to the open-ended questions some participants

made direct comparisons between the social dilemma game they were playing and other social

dilemmas that they have encountered in the past or may encounter in the future. One participant

equated trusting someone to cooperate in the game to asking a stranger or associate to "save my

place in line please" (Participant 77). Others compared game cooperation to trusting someone to

return a pencil that they had borrowed or not "tattling" on someone (Participant 37 and

Participant 5). An interesting analogy for cooperation in the game came from a participant who

wrote: "In life sometimes you will be confronted with life making decisions; therefore, you have

to be able to cooperate with your peers even if you don't know their name. For example, you

might be stuck in an elevator with a pregnant women. What do you do? There is another person

inside, do you and him cooperate or do you argue? So, you see it is very important to cooperate"

(Participant 114). These types of responses were equally present in all of the racial/gender

groups and serve to demonstrate that at least some participants (n = 6) equate the social dilemma

present in the game Prisoner's Dilemma with social dilemmas that are encountered in natural

settings.

Additionally, five participants referenced metaphors popular in other settings to explain

their decisions in the laboratory. One participant compared his decision to cooperate as takingn]

one for the team" (Participant 61). Another said "although society is all about survival of the

fittest, I do believe that we should work together to better survive [rather] than try to kill each

other" (Participant 135). Two participants alluded to the "win-win" philosophy popular in the

corporate sector, one writing that, "In the long run, it is better for us to cooperate and not be









greedy. Same idea as organizations or businesses. If they all helped each other out instead of

working against one another, everyone would win" (Participant 10).

Support for the external validity of the study is also evidenced in responses such as "I

equate this game to life and I felt to cooperate is the risk I take everyday, so why not transfer that

same value to a game" (Participant 131). Similar responses like "In my opinion in a life

perspective cooperating is the only way to succeed" demonstrate the parallels that some

participants draw between their decision in the game and decisions they make in their daily life

(Participant 33). On the other hand, a handful of participants make reference to the artificiality

of the experiment with comments such as, "I know I'll feel bad for her if she gets no money, but

I kind of just convinced myself that it's a game and the whole point of playing a game is to have

a winner. This now as I'm reading it sounds really terrible though because we should have fun

together not compete" (Participant 44). Overall, however, most references to the external

validity of the study support the assertion that participants draw on their real world experience in

their game decision making process. Furthermore, some participants equate their game decision

to decisions they would make in other social dilemmas. What these other social dilemmas are,

however, can vary greatly between participants.

An issue that negatively affects the external validity of the study is the amount of

monetary pay-out given to participants depending on their game decision and the decision of

their game partner. As noted by Simpson (2003), changes in the monetary reward and

punishment pay-out matrix produce different types of social dilemmas with possibly different

results. With regard to this, one to two cooperative participants in every experiment group except

the Black female group, expressed that had the pay-out been larger they would have defected.









Therefore any generalization of these findings to other contexts should consider the relatively

minor stakes the social dilemma in this study presented participants with.

In the following sections, I summarize, compare, and contrast the general reasons each

of the six experiment race and gender groups gave for both cooperation and defection (see Table

5-2).

Reasons for Cooperation and Defection

White male experiment group

Among participants in this group there were five frequently cited reasons for cooperation.

As with all of the six experimental racial/gender groups, the most popular reason given by

participants in the White male group for cooperating is that cooperation offers the most mutually

beneficial outcome for themselves and their game partner. As one participant wrote "I

cooperated because that's the best outcome for the both of us. It's not the most important to care

about my well-being but I would like to see others benefit as well" (Participant 20).

The next most common explanation for cooperation amongst the White male group

centered on the theme that they cooperated because of their personal faith in humanity. Of the

13 participants in this group who cooperated, five expressed sentiments such as "I chose to

cooperate because I believe that humans are inherently good and unfortunately we have been

given all this incorrect, negative propaganda that it is us against the world. But in reality it is

not; we are all in this together and the more we cooperate (Pun intended!) the better off we are

all going to be" (Participant 2). And, "I chose to cooperate because I tend to give people the

benefit of the doubt and [believe] that people are good-hearted" (Participant 13). It should be

noted that faith in humanity as a reason for cooperation was either not cited at all or not cited by

more than one participant in any of the other five experimental groups.









After faith in humanity, the next major reason for cooperation stated amongst the White

male group was that his game partner seemed nice/friendly. "Though I never knew that student

[before today], I get the impression that he's a nice, genuine person from the first impression"

(Participant 10). Another writes, "I cooperated because the other participant appeared that he

would be more willing to cooperate. By that I mean when I saw him he did not look like the type

of person that would make a decision to go against a complete stranger. He looked like a

genuinely good-mannered person." Another reason given by multiple participants to explain

their cooperation was simply that "cooperation is the better choice" (Participant 13). To

participants employing this reasoning it seems that cooperation is the most rational alternative.

The primary explanation for defection given by the White male experimental group

participants was that defection guarantees money. Ten out of the eleven participants in this

group who defected attributed their choice, at least in part, to the fact that irrespective of their

game partner's decision, they would receive either $3 or $9." Statements such as "I choose to

defect because there is a guarantee I will receive money" were frequently observed in this

experimental group (Participant 1).

A second, less major reason for defection cited by members of this experimental group is

that defection made the maximum individual pay-out possible. "I defected because it... offers

me the chance to win the most [money] possible" (Participant 4). Finally, the sentiment that a

person's game partner "will likely think and/or act like me" was equally given as a rationale for

both defection and cooperation decisions.

White female experiment group

Participants in the White female experiment group primarily cited "cooperation is the

most mutually beneficial outcome" as the reason for their cooperation. Slightly more than half









of the women who cooperated expressed this logic in their written response to the question of

why they cooperated of defected. Another popular response for this group was, "I did not want

to sucker my game partner." One participant who cooperated noted that "by choosing to defect I

feel as if I am being unjust to the other participant" (Participant 31). A participant with a similar

view wrote, "if I chose to defect and she chose to cooperate, I would feel guilty about the money.

Choosing to cooperate means that I am still electing to win money, but not hurting the other

person with my motives" (Participant 27). Unlike the White male group, "faith in humanity"

was not a popular response in the White female group and was mentioned by only one person.

Interestingly, three respondents in this group wrote that they cooperated because they

"did not need the money." "Even if she chose defect, losing the money would not be an issue for

me" (Participant 47). This sentiment was also expressed multiple times in the Black female

experiment group and in the female cross-racial group; however, only one male (White) in the

entire study mentioned not caring about the money (Participant 105).

One final cooperation theme that became evident in the White female experiment group

is that the female gender of the game partner played a role in some participant's decision to

cooperate. As one participant put it, "There was the possibility... that I would cooperate and she

would defect to receive the full money, but I felt that isn't something a girl would do, more of a

sabotage act a male would commit (sexist, I agree)" (Participant 25). This view was shared by

another participant who wrote, "My first instinct is to cooperate to be nice but I would figure

most people would defect because they will make money and they can't lose, so if she chooses as

I think she will we both will win. It's just an initial judgment to go by. If it were a male I would

have no question but to defect, because she's female she might cooperate" (Participant 38). A

third expressed "I think the other participant, as a girl, is more likely to cooperate than defect"









(Participant 41). Notably, White women were the most likely to list gender as the reason or

partial reason for their decision. None of the men and only two Black women in the study

mentioned gender in describing their decision-making rationale. Despite the view by some

White women in this group, as well as in the cross-race group, that females are more likely to

cooperate as the quantitative part of this study demonstrated, White women had the overall

lowest cooperation rates of any of the six racial/gender experimental groups.

Like the White male group, the large majority of defectors in the White female group

cited the guarantee of money as the principal reason for their defection decision. One White

women stated "I chose to defect because I figured I cannot leave with $0" (Participant 30). Also,

similar to the White male group, the White female group's second most commonly given reason

for defecting is that defecting makes the maximum pay-out possible. Due in part to their higher

defection rates, however, White women listed this reason twice as often as White men did.

Three women also mentioned that they defected (among other reasons) because they did not

know their game partner. "I did not know the stranger," wrote one participant of her decision to

defect (Participant 28). This reason was given either one or two times in each of the other five

racial/gender experiment groups. Finally, the belief that a person's game partner "will likely

think and/or act like me," was listed slightly more often as a rationale for cooperation than for

defection.

Black male experiment group

Participants in the Black male group (with a few notable exceptions) generally gave the

briefest written answers to the questionnaire of all the six racial/gender experiment groups.

Concomitantly, they provided fewer types of reasons for their cooperation or defection. Like the

previous groups discussed, the most common reason for cooperation given by participants in the









Black male group centered on the theme that "cooperation is the most mutually beneficial

outcome." Ten out of the 15 players in this group who cooperated gave this as their entire or

partial response. The only other rationale for cooperation given multiple times by participants in

this group is variations on the sentiment, "I cooperated because I am a good/nice/cooperative

person." Statements such as "that is just the type of person I am, if we can both get paid why not

cooperate," and, "seriously, I am a nice person and I do not see any reason to defect" were used

to communicate that participants viewed cooperation as a positive character trait (Participants 53

and 61). One participant in this group explicitly wrote, "I am a cooperative person" (Participant

63). Curiously, this response theme, although fairly common among Blacks in the study, was

solely mentioned by one White male and one White female. On the other hand, the popular

White male group cooperation response, "faith in humanity," was only mentioned by a single

Black (male) participant in the entire study.

Unlike the majority of defectors in the White male and White female groups who listed

guaranteed money as the most common reason for their defection, the most common response

participants in the Black male group gave for defection is that defection made the maximum

game pay-out possible. Six of the nine defectors in this group mentioned the possibility of

achieving the maximum pay-out as a motivating factor. One participant wrote, "I chose to defect

because I am a very competitive person. This is a game and defecting gives me the chance to

win the most" (Participant 54). Another stated, "[I] defected in hopes that my partner will pick

cooperate and I will get nine dollars. Sounds selfish but oh well," thus again highlighting how

this social dilemma decision can be viewed by some as a reflection of individual character

(Participant, 59). The second most typical rationale for defection among participants in the









Black male group was that "defection guaranteed money." Four respondents gave this as part or

all of their answer. Members of this experiment group gave few other reasons for defection.

Black female experiment group

Participants in the Black female group made frequent reference to the "mutually

beneficial outcome" of cooperation and like all other racial/gender groups listed it more often

than they did any other reason for cooperation. Other, moderately common responses to the

question "Why did you cooperate or defect?" given by participants in the Black female group

are: (1) I cooperated because I am good/nice/cooperative person; and (2) I think my game partner

"will likely think and/or act like me." In an illustration of the latter reason a respondent wrote

"[I cooperated] because I like to think positively and I believe that she will do the same"

(Participant 95).

The other major theme to emerge from this group's responses (more clearly than from

any others) was that of a moral/altruistic imperative to cooperate. Examples of altruistic and

morally laced statements made by participants in this group include responses such as [I

cooperated] because I was thinking about helping the other person out and we could both leave

with money" (Participant 74) and "I cooperated because I like to share and am not a very selfish

person so, it was OK to cooperate; the more the better. Why would I be greedy and not share the

wealth" (Participant 75). Another respondent stated, "I cooperated because, why would I want to

work against someone for seemingly no reason, it seems a bit wrong and selfish... I'd feel a tiny

bit bad if I somehow robbed her of an easy, free $6" (Participant 77).

Related to the moral/altruistic theme was that of "I cooperated because I do not need the

money." A handful of Black female respondents expressed statements such as "if she gains all

the money and I gain nothing, it's OK because I would feel that she needs the money. Whereas I









really am not in need" (Participant 87). Additionally, assertions such as "the other person seems

caring enough to want to help someone with a survey, so she seems trustworthy; also, I figure if

she decides to defect I'm not at a great loss" are also examples of this line of reasoning

(Participant 92). An additional type of response that loosely corresponds to the moral/altruistic

theme is that of choosing cooperation to "avoid feelings of guilt" which was cited by three

participants in the group.

Though this group had relatively few defections, the two most common reasons given for

defection by participants in this group were that "defection guarantees money," and that,

"defection made the maximum pay-out possible. Participants in this group gave no other high-

frequency responses for why they chose defection.

Male cross-race experiment group

The male cross-race group had the highest cooperation rate of any of the six experiment

groups. The most common reason given among both Black and White respondents was that

"cooperation provided the most mutually beneficial outcome." The next most common reasons

for cooperation given by Black respondents centered on the "I cooperated because I am a

good/nice/cooperative person" theme and the moral/altruistic theme. Both of these response

types typically expressed that the participant felt a strong connection between their character and

values and their decision to cooperate. One Black participant in this group wrote:

I chose to cooperate because the money being issued collectively is maximized if
both parties decide to cooperate ($12 in total), along with the idea that I don't
want to seem like a selfish jerk if I happened to choose defect and he chooses
cooperate. Sure, I'd receive $9, but the fact that he received no money because of
my selfish action I believe significantly outweighs the monetary value of $9. Of
course, he could choose to defect, leaving me with no money; that would simply
convey the type of character he has. In this case, it would bother me much more
that he received no money because of my selfishness, than vice-versa. Money
isn't something that could deter my feelings and values (Participant 117).









Another Black participant in this group wrote, "maybe I chose not to defect because of my

parents and upbringing" (Participant 112).

White males in this group who cooperated mostly wrote about cooperation as "mutually

beneficial," but were less likely to invoke moral or character arguments in their response. One

White participant wrote, "I chose to cooperate because the net benefit to both people cooperating

($12) is greater than the net benefit of either both defecting ($6) or one defecting and one

cooperating ($9)" (Participant 101). Another wrote, "I cooperated because I felt making $6

would be a decent gain for me and the other, and another $3 to defect isn't worth losing more

[total] dollars. However if the game was for millions, I would probably defect" (Participant

120).

Interestingly, no other cooperation themes among White male participants in the cross-

race group emerged. The theme "faith in humanity" that was the second most popular response

among the exclusively White male group was not mentioned by a single White male in the cross-

race group. Similarly, the response, "I cooperated because my game partner looks

nice/friendly," which was the third most popular response in the White male group, was not

mentioned by a single White male in the cross-race group (although one Black respondent listed

it).

Notably the cross-race group had only one Black participant who defected (compared to

four White participants who chose defection). Also notable, is the rationale given by the Black

participant for his defection. He wrote, "I really wanted to cooperate but I don't know the other

person I'm playing with... my theory: with me being who I am, if he chooses to cooperate, then I

would have much respect for him and I would give him half of the money I received when I see

him again. However, I couldn't put cooperate because I'm pretty sure he wouldn't do the same









for me" (Participant 116). This participant was one of only two players in the entire study to

suggest he would split the maximum defection money (though this task would have been

difficult because participants' departure time was intentionally staggered and they did not know

one and other).

The most frequent response White participants in this group who chose defection gave

was based on the view that defection was the most "economically rational" choice. The other

most common responses for defection mentioned by Whites in this group were that defection

"guaranteed money," and that "defection was the safest option." One participant employing all

of these views wrote, "I think choosing cooperate is a risky and unwise choice since you are not

guaranteed money and even if you do win, you cannot win the highest amount possible"

(Participant 110). Notably, every White who defected in the male cross-race group cited

economic logic as the primary reason for their defection.

Female cross-race experiment group

The female cross-race group had the most varied responses for why they cooperated than

any of the six racial/gender experiment groups. Still, the most common reason for cooperation

given by both Black and White women in this group was that cooperation offered "the must

mutually beneficial outcome." The second most common response among Black women was

that their "game partner will likely think and/or act like me." It should be noted that Black

women in this group were twice as likely as White women in this group to employ this response.

The second most common types of responses for White women was "I cooperated because I

have nothing to lose," and, "defecting is greedy," which were each mentioned three times.

Employing the former type of reasoning, one White women wrote, "If, by chance the other









player decides not to cooperate, I am no better or worse off than when I entered this room,

having received $0" (Participant 127).

The most common response for defection given by both Black and White women in this

experiment group was that "defection guaranteed money." A couple of White women also

utilized the "defection is the safer option" theme in describing why they defected. One White

women wrote of her defection, "I didn't want to end up with $0... and I also prevented myself

from being left in a vulnerable position" (Participant 137). This emphasis on safety was a theme

that was generally more common among women than among men and which I further discuss in

the section titled, Gender and Race Influences. In the next section, I explore and discuss the

responses to the questions dealing with trust and expectations.

Open-ended Data Analysis of Trust and Expectations

Though responses to the questions "Do you feel you can trust the other player? Why or

why not?" and, "Do you believe the other player will cooperate? Why or why not?" were

primarily incorporated into the questionnaire for quantitative coding and analysis (see Chapter

4), a content analysis of responses to both these questions uncovered a difference in experiment

group responses for the White male and Black male experiment groups. Before discussion of

this difference, however, it should be stated that responses to these questions were generally very

short (one to three sentences) and mostly similar across all groups. Typical responses read, "no,

because she probably wants to receive the larger amount also," or, "I have faith that he would

pick cooperate" (Participants 31 and 63). Moreover, quantitative analyses found no significant

race or gender differences on the trust variable and no gender differences on the variable

measuring expectation of game partner's decision. However, some overall race group

differences were found on the latter variable (see Chapter 4).









Despite the limited racial and gender group quantitative variation in the overall coding of

responses, content analysis of experiment group responses found a major difference in the

language used by different experiment groups to explain the decision to cooperate. This

difference centered on the use of the physical appearance of a game partner as a factor in

determining that person's trustworthiness and predicting their game decision to cooperate or

defect. Labeled as the "halo effect," this phenomenon causes others to assume that those who

are physically attractive are also more generous, trustworthy, and sociable than those who are not

attractive (Katz, 2003). Additionally, several studies have also established a positive correlation

between a person's level of physical attractiveness and their chances of being the recipient of

helping behavior (see Benson, Karabenick & Lerner 1976; Mims, Hartnett & Nay 1975; West &

Brown 1975). With regard to the Prisoner's Dilemma, Morse et al. (1974) showed that physical

attractiveness plays an important role in the cooperation/defection decision. In their study,

females playing the Prisoner's Dilemma made more cooperative responses when playing with an

attractive male partner than with an unattractive male partner. They also found that males work

harder for an attractive female experimenter than for an unattractive one, and that males allow

attractive females to exert more social influence over them than they do an unattractive female.

Even among players of the same gender, perceived physical attractiveness could possibly

influence cooperation and defection rates.

In this study, over one-third of participants in the White male group cited the physical

appearance of their game partner as a factor in their cooperation. One respondent wrote, "I feel

like I can trust him because of his disposition and clothing style. I believe judging a book by it's

cover is completely necessary, often useful, and yes sometimes results in false assumptions,

however with his clean-cut looks, I don't think he's got too many radical tricks up his sleeve"









(Participant 21). ) Other participants wrote, "he looked like a trustworthy person. He appeared

well maintained and intellectual. He looked like someone that would trust me as well," and,

"simply put he looks like a nice dude" (Participants 13 and 19). Interestingly, all participants in

the White male group who referenced physical appearance as a factor in their decision to

cooperate used the term "looks" to describe the assumed attributes of their game partner,

whereas most women in the study employed the term "seemed" to describe assumed attributes.

The Black male group, in contrast, had only one participant who employed the physical

appearance of his game partner to explain his cooperation. This participant wrote, "he looks like

he needs money just like I know I need money, so I feel we both will make the same decision

and cooperate so we can have a few extra dollars in our pockets" (Participant 65). Black and

White women, irrespective of their experiment group, were about equally likely to cite the

physical appearance of their game partner in their responses to questions on trust and

expectations, although White women showed a clear preference for the word "seems" in

describing their game partner. Statements such as "she seems nice" were very common in the

White female group.

An interesting finding is that no White males in the cross-race group (in contrast to one-

third of the males in the exclusively White male group) referred to the physical appearance of

their game partner in their responses to the questions regarding trust and expectations. Instead, a

common response for White males in this group was that they could not say if they trusted their

game partner or if he would cooperate because they did not know him. One White male

respondent wrote, "I don't know him well enough to make that judgment" (Participant 105).

Although this was a common style of response for many participants irrespective of their

experiment group, unlike White males in the cross-race group, one third of the Black males in









the cross-race group cited the appearance of their White game partner as a factor in why they

trusted him or expected him to cooperate. One participant wrote, "I think he will cooperate

because of the way he looks. His demeanor seems to be laid back" (Participant 107). Another

stated, "He seemed like a cool person" (Participant 97). The absence of references to the

physical appearance of Black game partners by White males and greater mention of perceived

visible attributes of White game partners by Black males in the Cross-race group poses some

interesting questions about who is allowed to talk about physical appearance and in what context.

Though the use of physical appearance as a proxy to measure assumed social and

personality characteristics of strangers is well documented (see Zebrowitz, Waltham &

Montepare 2005; Katz 2003; Mulford, Orbell, Shatto & Stockard 1998; Kahn, Hottes & Davis

1971), whether physical appearance matters more for certain racial and gender groups is a

question that needs further exploration. This study's open-ended data findings show that many

participants rely on stereotypes in their judgments of the trustworthiness and cooperative

potential of others. Yet while many participants freely admit to the use of certain stereotypes

(such as stereotypes about socioeconomic status, educational level, or social group membership),

the mentioning of race and/or racial stereotypes is clearly taboo. For example, differences in

response styles according to the race of a participant and the race of their game partner suggest

that judgments based on physical appearance are permissible so long as you are a racial-minority

group member or if your game partner is of your same racial group. Thus, a White male can list

the "look" of a White game partner as his reason for trusting him or expecting him to cooperate,

but is socially prohibited (lest he be considered racist) from listing the "look" of a Black game

partner as a reason for his cooperation/defection decision. Additionally, it is more common and









appears to be more acceptable to list a game partner's physical appearance as a reason for

trusting or cooperating with them rather than as a reason for not trusting them or defecting.

The fact that the male cross-race experiment group had the highest cooperation rates of

any in the study may mean that participants in this group are taking racial considerations into

account in their decision making and are cooperating at a higher rate because of it. In the next

section, I analyze the references made to gender and race by participants and explore the ways

these demographic factors may influence cooperation/defection decisions.

Gender and Race Differences

Only one male participant in the entire study cited gender in his written responses to the

open-ended questionnaire. He was in the Black male experiment group and wrote, "I generally

would trust someone that never gave me a reason not to trust them, but there can also be

underlying trust because the other player was the same gender and race as me" (Participant 51).

Significantly, no other males in the study cited gender in their written responses. As stated

earlier in the chapter, White women (in both the White women and cross-race experiment

groups) were most likely to cite gender as factor in their cooperation/defection decision. One

White female participant expressed, "I believe she is more likely to cooperate than to defect

because she is a woman. Women are more prone to cooperation, and would be more likely to try

to guarantee the maximum total payout between two players. I think she will comply with this

stereotype" (Participant 126). Another stated, "I feel that I can to some degree, trust her. I

suppose it's because she is female as well as soft-spoken. She does not appear as a threat"

(Participant 31).

A theme of being more willing to trust another woman rather than a man also emerged

among both White and Black female participants. Explaining why she trusts her female game









partner, a White women wrote, "[her] being a female instead of a male influences my safety

decisions" (Participants 28). Similarly, another participant wrote, "the other player is female like

me and I more willingly trust a female I don't know than I would a male" (Participant 124). In

one case, a woman who was perceived as portraying masculine qualities was discriminated

against because of them. Her game partner (after defecting) wrote, "I don't think I trust her

because she seemed more tom-boyish. But maybe she is more nice than her exterior predicts. I

don't really need to trust her one way or another" (Participant 38). Interestingly, this greater

trust of females by certain participants did not manifest itself in high cooperation rates for the

White female group that was most likely to employ this rationale in explaining their game

decision. Instead, a stated emphasis on issues of threat, risk, and safety resulted in the more self-

serving and competitive choice, defection.

There are several explanations for White women having the highest defection rates. The

most common is that White women did not want to be taken advantage of (suckered) or be seen

as weak. A second, less obvious reason is that they are using a gender stereotype to predict and

take advantage of their female partner. In other words, certain female participants may perceive

cooperation as a gendered weakness (but one they need not adhere too) that they can exploit for

their own gain. This is evident in defection statements such as "I choose to defect because I

think [she] would cooperate" or "normally I cooperate in my day-to-day life, choosing defect

breaks me out of my normal routine" and "with my personality, I am usually one to take an

opposite and independent stand on issues" (Participants 46, 40, and 28). Rather than cooperating

and thus conforming to gendered expectations, many White women may be doing the opposite

for either personal gain or as a result of the high personal risk that cooperation entails.









Men's responses also reflect gendered patterns in that they are less concerned about the

personal risk of cooperation and more concerned about being perceived as bad, greedy, self-

interested, and/or biased. Masculinity places them in a position of strength where from they can

take perspectives similar to that of the male participant who wrote, "I could defect and be almost

positive I'll get $9 but I'll feel like an ass and $3 is not worth the social punishment... I don't

think too many would 'risk' the normal cordial social lives they have with strangers for a few

bucks" (Participant 19). Thus, for men (and perhaps Black women) it may be that cooperation

yields more social rewards than potential economic risks, and that the social punishments

defection may entail take precedence over the economic advantages defection provides.

In addition to, and sometimes coinciding with, gender, a game partner's race was also

listed by a handful of participants as a factor in their decision making process. However, it

should be noted that no participants in the White male, White female, or male cross-race

experiment groups mentioned race in any of their written responses. Instead, discussion of race

as a decision making factor appeared exclusively in the Black male, Black female, and female

cross-race groups.

One White woman in the female cross-race experiment group wrote, "I feel that because

she is female, the other player stands a greater chance of cooperating than defecting. However, I

also feel that an African-American woman may be more likely to stand up for what she wants by

defecting, than to take the chance of cooperating. Based on these facts I feel that I have a greater

chance of winning more money, and I have guaranteed that I will not leave empty-handed"

(Participant 126).

Another example of racial influences on a participant's decision to defect were evident in

this response by a Black participant:









I feel that in this world, the one person we must first look out for is you. This by
no means is suggesting that people should be selfish with one another. I just feel
as though in order to help the people around us, we must be at our most optimal
level (financially, psychologically, socially, emotionally, etc.). I was always
taught not to trust anyone, but still treat others the way you want to be treated. I
can say that I believe most people have a more selfish (for the lack of a better
word) mindset than an altruistic one. I briefly met with the girl in the other room.
She seemed very outgoing and willing to talk. I would assume she is of
Caucasian descent, and although I would never automatically assume that she has
any negative feelings toward me, my experiences growing up have taught me that
whites secretly do not like blacks on a standard equal to them. I've even been
told they have 'a secret agenda' toward blacks (Participant 129).

Asked whether he believed his game partner would cooperate, a Black male wrote, "I would

hope so because we are both Black and Greekl2. I think that he'll cooperate" (Participant 53).

Though the quantitative component of this study showed that certain racial differences in

cooperation levels exist, the societal taboos against openly discussing race make the precise

nature of and reasons for these differences difficult to uncover. In the following section I will

summarize and discuss the open-ended data findings of this study.

Discussion of Open-ended Data Findings

The open-ended data findings on the external validity of the study support the view that

the social dilemma present in a monetary Prisoner's Dilemma game is to a certain extent relevant

to other social dilemma situations. Defining exactly which social dilemmas Prisoner's Dilemma

findings can be used to generalize about, however, is a question whose answer seems to vary

largely from person to person based on subjective frames of reference and personal histories. It

is also a question that needs more in depth examination.

The analysis of open-ended questions also demonstrates that the most common reasons

given for both cooperation and defection are very similar across various racial and gender group


12 By "Greek" the participant presumably means affiliated with a fraternity. Since participants in the study were not
allowed to participate with someone they knew previously, measures were adopted to ensure members of the same









pairings. Secondary response themes for cooperation, on the other hand, have greater variance

by racial and gender groups. Blacks were more likely to view the decision to cooperate or defect

as a moral choice and reflection of an individual's character than were Whites. This helps to

explain the greater cooperation levels of Blacks and the statistically significant correlation of the

religiosity, race, and cooperation variables.

Differences in responses by gender lend limited support for the view advocated by

Simpson (2003) that women are more likely than men (particularly Black men in this study) to

defect out of fear rather than out of greed. Nonetheless, a significant difference between Black

female and White female cooperation rates point to the important role that cultural factors such

as racial and religious group membership can play in social dilemma decisions. In light of this,

the findings of past studies that rely exclusively or overwhelmingly on White participants should

be revisited and duplicated using broader, more diverse samples. Furthermore, it should be

noted that among the various race and gender experiment groups, participants belonging to

minority groups are much more likely to take racial and/or gender group membership into

account in their game decisions. This is supported by the finding that not a single White male

participant mentioned either the race or gender of a game partner as a factor in their

cooperation/defection decision. Similarly, only one Black male mentioned gender as a factor in

his cooperation/defection decision. White women in the study were most likely to mention the

gender of their game partner as a factor in their decision making, yet only one White women

cited race. In line with intersection theory, Black women who may have faced subjugation and

discrimination due to both their race and gender, were the only racial/gender group participants

in the study to list both race and gender as factors in their cooperation/defection decision


fraternity or sorority group were not partnered with one another. In this case, the participants knew of each other's
Greek affiliation because they were both wearing shirts identifying their respective fraternities.









multiple times. In this group the intersection of race and gender coupled with the influence of

religiosity resulted in significantly higher cooperation rates than were present in the White

female experiment group.

The complete absence of race talk in the cross-race male experiment group, despite clear

differences in the language participants in this group versus same race groups used to describe

their game partners, suggests that race, as a basis for cooperation/defection and as a topic of

discussion, remains taboo. Additionally evidence of this taboo is apparent in the lack of

references to the physical appearance of Black game partners by White males in the cross-race

group. Though Bonilla-Silva (2003) has documented how many times the alleged colorblindness

of most Whites serves as a mechanism for maintaining White supremacy, the greater cooperation

rates of White participants in cross-race groups is likely not the result of covert racism, but

instead may be evidence of Bell's (1980) Interest Convergence H)pv,/wh\i\ In other words,

White participants may be viewing the experiment as a chance to show they are non-racist by

cooperating with Black participants for a mutually beneficial outcome. There is also modest

evidence suggesting that Black male participants at this predominantly White university may

have bought into the sincere fictions of Whiteness presented by our society and discussed earlier

(Chapter 2), and that they are more likely to cooperate with White males as a result. Due to the

limitations of the open-ended data gathered, more in-depth research is needed to substantiate

these theoretical conclusions. In the final section of this paper, I will conclude and present

suggestions for future research.










Table 5-1. Thematic coding categories with frequencies for responses to the question,
"Why did you cooperate or defect? Please explain in detail.
Cooperation Themes Defection Themes

1) Mutually beneficial (56) 1) Defection guaranteed money (38)
2) Person will think or act like me (26) 2) Made maximum pay-out possible (22)
3) I am a good/cooperative person (14) 3) I do not know the other person (11)
4) Morally superior choice (11) 4) Defection is the safer option (9)
5) Wanted to avoid feelings of guilt (10) 5) Defection is the logical choice (4)
6) Person seemed nice/friendly (9) 6) I think other person will cooperate (3)
6) Did not want to sucker the player (9) 7) Defected because I am greedy (2)
8) I do not need the money (8) 7) Defected because the stakes are low (2)
8) Faith in humanity (8) 9) Defected due to person's appearance (1)
8) Defecting is greedy (8) 9) Cooperation could result in zero gain (1)
11) Cooperating is fair (7)
12) Cooperation is the better choice (6)
13) May have defected if more at stake (5)
14) Gender influenced my decision (4)
14) I1 have nothing to lose (4)
16) Makes me feel better about myself (3)
17) Makes me the better person/man (1)
*Frequency distribution of responses is in parenthesis after each reason for cooperation and defection.
*Multiple responses were counted for participants who provided them.

Table 5-2. Reasons for cooperation and defection by experiment group with frequency
of response
Experiment Group Reasons for cooperation Reasons for defection


White Males
n = 24






White Females
n = 24






Black Males
n = 24


n = 13
1. Mutually beneficial (9)
2. Faith in humanity (5)
3. Nice/friendly partner (4)
4. Better choice (3)
5. Partner will act like me (3)

n = 11
1. Mutually beneficial (6)
2. Against suckering partner (3)
3. Do not need money (3)
4. Gender of partner (3)
5. Partner will act like me (3)

n =15
1. Mutually beneficial (10)
2. I am a good person (4)


n = 11
1. Guaranteed pay-out (10)
2. Maximum pay-out possible (3)
3. Partner will act like me (3)



n = 13
1. Guaranteed pay-out (10)
2. Maximum pay-out possible (6)
3. Partner is a stranger (3)



n=9
1. Maximum pay-out possible (6)
2. Guaranteed pay-out (4)











Table 5-2 continued
Experiment Group Reasons for cooperation Reasons for defection


Black Females
n = 24







Cross-race
White Males
n= 12




Cross-race
Black Males
n= 12




Cross-race
Black Females
n= 12


Cross-race
White Females


n = 18
1. Mutually beneficial (7)
2. I am a good person (4)
3. Partner will act like me (4)
4. Moral/altruistic choice (3)
5. Do not need money (3)
6. Guilt avoidance (3)


n = 8
1. Mutually beneficial (5)





n = 11
1. Mutually beneficial (8)
2. I am a good person (3)
3. Moral/altruistic choice (3)


n = 8
1. Mutually beneficial (7)
2. Partner will ace like me (4)


n=7


n=6
1. Guaranteed pay-out (3)
2. Maximum pay-out possible (3)







n=4
1. Economically rational (3)
2. Guaranteed pay-out (2)
3. Safest option (2)


n=l
1. Partner is a stranger (1)





n=4
1. Guaranteed pay-out (3)


n=5


n = 12 1. Mutually beneficial (4) 1. Guaranteed pay-out (5)
2. Nothing to lose (3) 2. Safest option (2)
3. Defecting is greedy (3)
*Frequency distribution of responses is in parenthesis after each reason for cooperation and defection.
*Multiple responses were counted for participants who provided them.









CHAPTER 6
CONCLUSION

Though experimental games such as the Prisoner's Dilemma have long been popular,

questions concerning their relevance to real-life settings and their general lack of theoretically

grounded hypotheses have contributed to the results of gaming studies being largely ignored by

social scientists outside the field of experimental gaming (Pruitt & Kimmel 1977). Despite this

disregard, experimental games are valuable in studying social dilemmas in part because they

permit precise measurements of elusive variables such as extent of cooperation, and because they

allow hostile and competitive behavior to transpire without injury to participants or their

relationships (Pruitt & Kimmel 1977).

The need to comprehend social dilemmas is of critical importance since as Van Lange

and Messick note, "the functioning of societies, groups, and relationships is perhaps most

strongly challenged by social dilemmas... because the well-being of these larger units is

threatened when most or all individuals pursue their own well-being rather than the collective

well-being" (1996:93). In the modem era, technological advancements in nuclear and biological

weaponry, communications systems, and industry have created with them new social dilemmas

in which defection's mutually deficient outcomes (and the human suffering those outcomes may

entail) take place on a global scale. Since the well-being of much of humanity lies in our ability

to cooperate on major social dilemmas, we must understand and be capable of overcoming

socially constructed differences in race, gender, ethnicity, nationality, ideology, and religion. As

stated by civil rights leader Martin Luther King Jr., "We must learn to live together as brothers or

perish together as fools" (1964).

Keeping in mind the need to employ theoretically grounded hypotheses to uncover the

racial and gender intricacies of social dilemma cooperation, the aims of this study consisted of:









(1) determining if Black college students, as members of a subordinated group, are more likely to

cooperate in a social dilemma than are White college students who are members of the dominant

group; (2) determining if female college students are more likely to cooperate with each other in

a social dilemma than are male college students; (3) exploring the reasons and rationale given by

respondents for cooperation or defection and (4) testing if levels of trust and expectations of a

game partner's decision differ between race and gender groups.

In regard to the first aim, this study provides support for the hypothesis that members of

subordinated racial groups are more likely to cooperate in a social dilemma than are members of

a racially dominant group. In the constructed social dilemma employed by this study,

quantitative analyses demonstrated that Blacks had both higher intra-group cooperation rates

than did Whites (69% intra-group cooperation for Blacks compared to 50% for Whites) and

significantly higher overall cooperation rates than did Whites (72% overall cooperation for

Blacks compared to 54% for Whites). In other words, Black participants were more likely to

cooperate with their game partner (irrespective of that game partner's race) than were White

participants. This finding lends limited support to the theory that Blacks have developed a

culture of cooperation as a response to White racism, and additionally suggests that Blacks

(especially Black males) are willing to extend that culture of cooperation to include Whites.

With regard to the second aim, overall gender cooperation levels did not significantly

differ between males (65% cooperation) and females (61% cooperation), although the

cooperation rates in the Black female experiment group (75% cooperation) and White female

experiment group (46% cooperation) did. These differing levels of female cooperation by race

suggest that in the United States cultural factors-such as race and religiosity-play a more

important role than gender in a participant's decision to cooperate or defect. These differences









also potentially reflect an important intersectional difference in the way women of different

racial groups are influenced by gender subordination. In this, subordination may be leading

Black women to cooperate at higher rates and White women to defect (mostly out of fear) at

higher rates when presented with a monetary social dilemma.

With regard to the third aim, it was found that although the reasons and rationale given

by participants in different racial and gender experiment groups for cooperation and defection

were mostly similar, minority racial and gender group members were more likely to account for

their minority group status in their cooperation/defection decision. This greater awareness of

race and/or gender group status by minority group members, coupled with the higher cooperation

rates of Black participants, lends support to the claim that experiences of subjugation (such as

McCormick and Franklin's [2000] inclusionary dilemma) and the increased group awareness that

they generate can manifest themselves as increased minority group cooperation in social

dilemmas.

Furthermore, the finding that White women had the lowest cooperation rates of any race

and gender experiment group can be interpreted to mean that most White women do not see

themselves as subjugated. This interpretation of the data is consistent with the work of

Lengermann and Wallace (1985) who argue that White women are less likely to see themselves

as an oppressed minority group than are Black women. Additionally, it was found that Black

participants were more likely to view the decision to cooperate or defect as a moral issue and as a

reflection of character, than were Whites. This finding is likely correlated with the significantly

higher levels of religiosity reported by Black participants compared to White participants.

With regard to the fourth aim, quantitative analyses revealed no significant race or gender

differences on the trust variable and no gender differences on the variable measuring expectation









of game partner's cooperation decision. There was, however, a statistically significant difference

between Black and White participants' expectation of game partner's cooperation decision when

participants who indicated they were uncertain as to whether their game partner would cooperate

or defect were excluded from the cross-tabulation. With participants who reported being

"uncertain" filtered out, 72% of the remaining Black participants believed their game partner

would cooperate compared to only 56% of White participants. Like the findings on cooperation

rates, this finding also suggests that subjugation increases a group's cohesiveness and ability to

work together in social dilemmas.

Lastly, content analyses on the trust and expectation variables revealed that among males

in the study, Whites were more inclined to report making judgments based on physical

appearance when playing a member of their same racial group, but refrained from mentioning

physical appearance when paired with a Black game partner. Possible explanations for White

males omitting mention of the physical appearance of Black game partners is that race is

considered by Whites to be a taboo topic, and that references to the physical appearance of a

Black game partner may be interpreted as racist.

Unlike most previous studies employing the Prisoner's Dilemma, the present study did

not rely on a sample composed almost exclusively of White participants and was therefore able

to generate insights into Black same-race and cross-race social dilemma cooperation and

defection. Additionally, the methodological decision to assign participants to one of six race and

gender groups allowed for cross-race group comparisons and for analyses of how the intersection

of race and gender influences social dilemma cooperation and defection decisions. Other

methodological advantages of the present study included the use of a single-shot rather than

multi-shot game, and the utilization of theoretically based hypotheses. Both of these









methodological design elements served to improve the validity and generalizability of the study's

findings.

Compared to previous Prisoner's Dilemma research on cross-race pairs, this study's

findings on racial cooperation contradict many of the earlier studies conducted in the 1960's and

1970's that showed lower levels of cross-race cooperation compared to same-race cooperation.

Nonetheless, this study's findings coincide with the most recent study of Heider and Skowronski

(2007), which produced results that were the exact converse of the earlier studies. This reversal,

coupled with the finding that White males refrained from referring to the physical appearance of

Black game partners, may be evidence of a societal shift toward less race-based discrimination.

However, an equally plausible and less optimistic interpretation of the cross-race findings is that

they merely reflect a desire among participants to appear non-racist. As Heider and Skowronki

have noted, this latter interpretation suggests that White participants in the study "were aware of

their potential prejudice and were overtly trying not to treat the African American partner

poorly" (2007:61).

Additionally, these findings may also be evidence of Bell's (1980) Interest Convergence

Hypothesis whereby Whites realize it is in their interest to cooperate with Blacks for a mutually

beneficial economic outcome. This study's findings on racial cooperation add to the racial

relations and social dilemma literature by showing both that (1) racial minority group status

positively correlates with social dilemma cooperation and (2) that anti-racism may be influencing

Whites' cooperation/defection decisions-in favor of increased cooperation-when they are

paired with a Black game partner.

On the question of gender cooperation in the Prisoner's Dilemma, this study like most

before it, suggests no overall difference in cooperation levels by gender. This, however, does not









necessarily imply that gender does not play a role in social dilemma cooperation since many

studies conducted at other times and places have shown that it does. What the lack of overall

gender difference implies, rather, is that cultural difference, including differences in gender

socialization and levels of patriarchy and subjugation, are more important to social dilemma

cooperation than is the biological classification of female or male.

Considering that the sample was composed entirely of undergraduate college students,

this study's findings may also be indicative that U.S. anti-discrimination educational

curriculums, programs, and policies are reducing overt discrimination against Blacks.

Furthermore, the findings imply that minorities are more likely to account for their minority

group membership in social dilemma decision-making, and that among Whites, concerns about

political correctness may suppress competitive behavior toward Blacks and make the mentioning

of race in formal settings taboo.

The finding that Blacks have significantly higher same-race and overall cooperation rates

than do Whites, may be evidence that Black students at predominantly White universities are

collectively resisting experiences of subjugation and/or alienation through cooperative strategies

that emphasize group interests over individual interests. In this, greater Black cooperation in this

social dilemma may in part be the result of coping strategies developed by Black students to

overcome social-structural and institutional hurdles faced by students of color.

The policy implications of the above findings are limited by the usual issues that arise

when trying to generalize from an experimental setting to a target setting and by the variation in

behavior that different types of social dilemmas may elicit. With this in mind, any attempts at

generalization should be made cautiously and should be limited to social dilemmas that involve

the allocation of finite resources such as those that often occur in the economic, social, political,









and interpersonal arenas. In these arenas, the most practical implication of this study centers on

the finding that racial domination, be it in the form of racism, discrimination, subjugation, and/or

exploitation, increases a groups awareness of its minority status and concomitantly also increases

its cohesiveness and likelihood of social dilemma cooperation. Researchers from the various

fields that employ games theory, social dilemma theory, social psychological theory, rational

choice theory, and symbolic interaction theory may find the results of this study relevant to their

work. Additionally, policy makers and activists in education, government, the military, the labor

movement, non-governmental organizations and in other sectors that require an understanding of

the intricacies of social dilemma cooperation and defection may also find this study useful.

Considering this is one of the few studies to utilize Black participants in the examination of

social dilemma behavior and to provide cross-race comparisons of social dilemma cooperation

and defection rates, these findings should not be overlooked.

Suggestions for future Prisoner's Dilemma and social dilemma research include the need

for researchers to account for and study the religiosity levels of participants in their sample. As

discussed in Chapter 4, the intersection of race and religiosity complicate this study's findings.

Based on the theoretical rationale that grounds this study's hypotheses, the observed differences

in racial cooperation are likely the result of shared experiences of minority group subjugation.

However, the significant differences in levels of religiosity by race make it feasible that observed

racial differences in cooperation are in actuality the product of differences in levels of religiosity.

In addition to the need to account for the influence of religiosity, the need to incorporate

significantly more racial minority group participants into future Prisoner's Dilemma studies is

also great. Latinos and Asian-Americans in particular are two large and growing US minority

groups that have been habitually ignored by Prisoner's Dilemma researchers. Corresponding









with a greater emphasis on racial minority groups' participants, more research needs to be

conducted to corroborate the findings of this present study. In the written responses to this

study's questionnaire, much about the racial factors that influence cooperation/defection was left

unsaid. Qualitative methods such as in-depth interviews and focus groups should be utilized to

uncover richer data on the more covert factors that influence a participant's

cooperation/defection decisions.

Quantitatively, games other than the Prisoner's Dilemma, such as ones based on

monetary ultimatums, should be employed to measure racial social dilemma

cooperation/defection. Doing so could not only reinforce, but also speak to the external validity

of the findings presented here. Also, employing a sample composed exclusively of

undergraduate university students is a limitation in this study that future research should address

by drawing from a more diverse pool of participants. The uniqueness of the university

environment may have influenced this study's outcome in ways that limit broader

generalizability.

The focus of the present research has been to use the Prisoner's Dilemma to identify the

links between race, gender, and levels of cooperation, trust, and expectations of a partner in

social dilemmas. However, due to a limited amount of research employing games theory to

study race relations, many questions still remain unanswered. By employing games similar to the

one adapted for this study, future researchers could gauge racial discrimination by comparing

intra- and inter-race cooperation rates for a given school, organization, town, or city.

Additionally, these results could be compared across time and place, providing hard data on the

relative levels of societal racial discrimination. In this, the application of social decision making

games to measure racial discrimination would be useful in that they permit hostile and aggressive









behavior to transpire without harm to participants or as Pruitt and Kimmel put it, "these games

permit conflict without tears" (1977:366). Furthermore, these games have the added advantage

of providing researchers with a way to measure racial discrimination without revealing what is

being measured to the participants of the study (who may otherwise inadvertently taint the

results). Finally, the potential for games in general and the Prisoner's Dilemma in particular to

be utilized as a barometer of institutional and societal levels of racial discrimination is great and

should be further explored.









APPENDIX A
MEASUREMENT INSTRUMENT

The following is a game in which you will have the opportunity to earn money for your
participation in this social decision making study. You will be playing the game alongside
another player (who is the person you just met). The game is simple; each of you can decide to
cooperate with each other or to defect and work against one another. Be aware that there are
benefits and risks attached to both cooperation and defection. The game's payout is a follows:

If you both cooperate you each receive $6.
If you both defect you each receive $3.
If one player cooperates and the other defects, the person who cooperated will
receive $0, and the person who defected will receive $9.


Game


Other Player


Cooperates


Cooperate







You





Defect


Defects


1. Please circle one of the following choices:


Cooperate


Defect









2. Why did you cooperate or defect? Please explain in detail.









3. Do you feel like you can trust the other player? Why or why not?









4. Do you believe the other player will cooperate? Why or why not?









Demographic Background

1. Gender: (circle one) Male Female



2. Age (in years)



3. How would you describe your racial or ethnic background? (circle all that apply)

Black (please specify)

African American

Asian (please specify)

White (please specify)

Caucasian (please specify)

Hispanic (please specify)

Latino (please specify)

Other (please specify)



4. College classification (please circle one)

a. freshman b. sophomore c. junior d. senior e. graduate student



5. College Major?



6. How many siblings do you have?










7. How many years of education do you expect to complete? (circle one)

a. B.A.

b. M.A./M.S./M.B.A./M.S.W.

c. J.D. (law)

d. Med./Vet/Dent.

e. Other Doctoral



8. What is your religious preference? (circle one)
a. Protestant -- what denomination?

b. Catholic

c. Jewish

d. Other please specify

e. Christian what denomination?

e. No religion



9. How religious are you?

A. very religious B. average C. below average



10. What is your family's yearly household income? (circle one)

a. $25,000 and below

b. $26,000 $40,000

c. $41,000 $80,000

d. $81,000 160,000

e. Above $160,000









11. Are you affiliated with a fraternity or sorority? (circle one) Yes or No

If you answered yes, what fraternity or sorority



12. Which of the following political orientations best describes you? (circle one)

A. Conservative

B. Liberal

C. Centrist

D. Other (please specify)



13. What political party do you identify with?

A. Republican Party

B. Democratic Party

C. Green Party

D. I am independent

E. Other (please specify)










APPENDIX B
INFORMED CONSENT

Informed Consent
Protocol Title: Social Decision Making Study

Please read the consent form carefully before you decide to participate in this study

Purpose of the research study:
The purpose of this study is to examine how people respond to social dilemmas.

What you will be asked to do in the study:
You will be asked to meet another student and afterward play a quick social decision making game. You will also
be asked to fill out a short survey.

Time required: 10-15 minutes

Risks and Benefits:
You may or may not experience minor feelings of anger or guilt after playing the game. We do not anticipate any
other risks to you by participating in the study. The benefit of the study is the opportunity to earn $0, $3, $6, or $9
for completing the game and survey.

Confidentiality:
Your identity will be kept confidential to the extent provided by law and your name will not be used in any report.

Voluntary Participation:
Your participation in this study is completely voluntary. There is no penalty for not participating.

Right to withdraw from the study:
You have the right to withdraw from the study at anytime without consequence.

Whom to contact if you have questions about the study:
Victor Romano M.A., Sociology Department, 3357 Turlington Hall, 786-338-1027
Supervisor: Hernan Vera Ph.D., Professor of Sociology, 3229 Turlington Hall, 392-0265 ext. 232
For questions about your rights as a research participant, please contact the UFIRB office at 352-392-0433
or PO Box 112250, Gainesville, Florida 32611.

Agreement:
I have read the procedure described above. I voluntarily agree to participate in the procedure and I have received a
copy of this description.


Participant: Date:

Principal Investigator: Date:









LIST OF REFERENCES


Anderson, Margaret L., and Patricia H. Collins. 2000. Race, Class, and Gender: An Anthology
(4 Edition). Wadsworth.

Apfelbaum, Erika. 1974 "On Conflicts and Bargaining." Advanced Experimental Social
Psychology 7:103-56.

Assensoh, Akwasi B. 2000. "Conflict or Cooperation?" Pp. 113-39 in Black and Multiracial
Politics in America, edited by Y.M. Alex-Assensoh & L. J. Hanks. New York University
Press.

Axelrod, Robert. 1980. "Effective Choice in the Prisoner's Dilemma." The Journal of Conflict
Resolution 24:3-25.

Axelrod, Robert. (1984). The Evolution of Cooperation. Basic Books,

Babbie, Earl. 2001. The Practice of Social Research (9th Edition). Wadsworth/Thomson
Learning.

Baca Zinn, Maxine. 1996. "What is Multicultural Feminism?" Feminist Studies 22:321-31.

Baca Zinn, Maxine, and Bonnie T. Dill. 1994. Women of Color in USA Society. Temple
University Press.

Barash, David.P. 2004. "Caught Between Choices: Personal Gain vs. Public Good." Chronicle of
Higher Education 50:B 13-16.

Baxter, George W. Jr. 1973. "Prejudiced Liberals? Race and Information Effects in a Two-
Person Game." The Journal of Conflict Resolution 17:131-161.

Bell, Derek. 1980. "Brown v. Board of Education and the Interest-Convergence Dilemma."
Harvard Law Review 518:93.

Benson, Peter., Stuart A. Karabenick, and Richard M. Lerner. 1976. "Pretty Pleases: The Effects
of Physical Attractiveness, Race, and Sex on Receiving Help." Journal ofExperimental
Social Psychology 12:409-15.

Bernard, H.Russell. 2000. Social Research Methods: Qualitative and Quantitative Approaches.
Sage Publications.

Blau, Peter. 1997. Inequality and Heterogeneity. Free Press.

Bonilla-Silva, Eduardo. 2003. Racism Without Racists: Color-Blind Racism and the Persistence
ofRacial Inequality in the United States. Rowman and Littlefield.









Brown, Kelly M. and Laura 0. Taylor. 2000. "Do As You Say, Say As You Do: Evidence on
Gender Differences in Actual and Stated Contributions to Public Goods." Journal of
Economic Behavior and Organization 22:255-67.

Brown, Kendrick T. 1999. "Skin Tone Bias and African-American Well-Being: A Dual
Influence Model Approach." Dissertation Abstracts International: Section B: The
Sciences and Engineering, 59:5616.

Butler, Judith. 1990. Gender Trouble: Feminism and the Subversion ofIdentity. Routledge.

Calhoun, Cheshire. 2000. Feminism, the Family, and the Politics of the Closet: Lesbian and Gay
Displacement. Oxford University Press.

Carby, Hazel V. 2000. "White Women Listen!" Pp. 389-403 in Theories ofRace
andRacism: A Reader, edited by Les Back and John Solomos. Routledge.

Cederblom, Douglas, and Carol J. Diers. 1970. "Effects of Race and Strategy in the Prisoner's
Dilemma." The Journal of Social Psychology 81:275-286.

Chong, Dennis 1991. Collective Action and the Civil Rights Movement. Chicago University
Press.

Clark, Kenneth, and Martin Sefton. 2001. "The Sequential Prisoner's Dilemma: Evidence on
Reciprocation." Economic Journal 111:51-68.

Collins, Patricia. H. 1997. "Comment on Hekman's 'Truth and Method: Feminist
Standpoint Theory Revisited: Where's the Power?" Signs 22:375-79.

Dawson, Michael. 1994. Behind the Mule: Race and Class in African American Politics.
Princeton University Press

Delgado, Richard. 1995. "Race, Sex, Class, and their Intersections. Introduction to part VI." P.
249 in Critical Race Theory: The Cutting Edge, edited by Richard Delgado. Temple
University Press.

Doane, Ashley W., and Eduardo Bonilla-Silva. 2003. White Out: The Continuing Significance of
Racism. Routledge.

Downing, Leslie L., and Kenneth H. Bothwell. 1979. "Open-space Schools: Anticipation of Peer
Interaction and Development of Cooperative Interdependence." Journal of Educational
Psychology 71:478-484.

Du Bois, W.E.B. [1903] 2004. The Souls ofBlack Folk. Paradigm Publishers.

Du Bois, WE.B. [1935] 1992. Black Reconstruction in America, 1860-1880. Atheneum.









Dubois, Frederique, and Luc-Alian Giraldeau. 2003. "The Forager's Dilemma: Food Sharing
and Food Defense as Risk-Sensitive Foraging Options." The American Naturalist
162:768-79

Feagin, Joe R. 2000. Racist America: Roots, Current Realities and Reparations. Routledge.

Feagin, Joe R. 2006. Systemic Racism: A Theory of Oppression. Routledge.

Feagin, Joe R., and Clarice B. Feagin. 2002. Racial and Ethnic Relations (7th Edition). Prentice
Hall.

Feagin, Joe R., Hernan Vera, and Pinar Batur. 2001. White Racism: The Basics. Routledge.

Frankenberg, Ruth. 1993. White Women, Race Matters: The Social Construction of Whiteness.
University of Minnesota Press.

Gilligan, Carol. 1982. In a Different Voice: Psychology and Women's Development. Harvard
University Press.

Hacker, Andrew. 1995. Two Nations: Black and White, Separate, Hostel, Unequal. Ballentine
Books.

Heider, Jeremy D., and John J. Skowronski. 2007. "Improving the Predictive Validity of the
Implicit Association Test." North American Journal of Psychology 9:53-76.

hooks, bell. 1991. Yearning: Race, Gender, and Cultural Politics. South End Press.

Hu, Yung-An, and Liu Day-Yang. 2003. "Altruism Versus Egoism in Human Behavior or
Mixed Motives." American Journal of Economics and Sociology 62:677-704.

James,Tammy., Lewis Soroka, and John G. Benjafield. 2001. "Are Economists Rational, or
Just Different?" Social Behavior and Personality 29:359-64.

Kaplan, Michael, & Ellen Kaplan. 2006. Chances Are...: Adventures in Probability. Penguin
Books.

King, Debra W. 1988. "Multiple Jeopardy, Multiple Consciousness: The Context of a Black
Feminist Ideology." Journal of Women in Culture and Society 14:42-72.

King, Martin L. 1964. Speech in St. Louis, Missouri, March 22.

Ledyard, John 0. 1995. "Public Goods: A Survey of Experimental Results." Pp. 111-94 in
Handbook ofExperimental Economics, edited by J.H. Kagel and A.E. Roth. Princeton
University Press.









Liebrand, Wim B.G. 1986. "The Ubiquity of Social Values in Social Dilemmas." Pp. 113-33 in
Experimental Studies of Social Dilemmas, edited by H. Wilke, C. Rutte, and D.M.
Messick. Frankfurt, Ger.: PeterLang.

Lengermann, Patricia M., and Ruth A Wallace. 1985. Gender in America: Social Control and
Social Change. Prentice-Hall

Lever, Janet. 1978. "Sex Differences in the Complexity of Children's Play and Games."
American Sociological Review 43: 471-83.

Lorber, Judith. 1994. Paradoxes of Gender. Yale University Press.

Lorber, Judith. 2001a. "The Social Construction of Gender.." Pp. 121-24 in Women's Voices,
Feminist Visions, edited by S. M. Shaw and J. Lee. Mayfield Publishing Company.

Lorber, Judith. 200 lb. Gender Inequality (2nd Edition). Roxbury Publishing Company.

Macionis, John J. 2005. Sociology (10th Edition). Pearson Prentice Hall.

Macy, Michael W., and John Skovoretz. 1998 "The Evolution of Trust and Cooperation between
Strangers: A Computational Model." American Sociological Review 63:638-60

Martin, Patricia Y. 2004. "Gender As Social Institution." Social Forces 82:1249-73.

Marx, Karl. [1862] 2006. "Economic Manuscripts of 1861-1863: Part 3. Relative Surplus
Value." Retrieved November 4, 2006 from the World Wide Web
http://www.marxists. org/archive/marx/works/1861 /economic/ch3 1. htm

McAdams, Richard H. 1995. "Cooperation and Conflict: The Economics of Group Status
Production and Race Discrimination." Harvard Law Review 108:1005-85.

McCormick, Joseph P., and Sekou Franklin. 2000. "Expressions of Racial Consciousness in the
African American Community." Pp. 315-36 in Black andMultiracial Politics in
America, edited by Y.M. Alex-Assensoh and L. J. Hanks. New York University Press.

Messner, Michael A. 1990. "Boyhood, Organized Sports and the Construction of Masculinities."
Journal of Contemporary Ethnography 27:197-211.

Mims, Patricia R., John J. Hartnett, and Robert W. Nay. 1975. "Interpersonal Attraction and help
volunteering as a Function of Physical Attractiveness." Journal of Psychology 89:125-31.

Minkler, Lanse.P. and Thomas J. Miceli. 2004. "Lying, Integrity, and Cooperation." Review of
Social Economy 62:28-50.









Morse, Stanley .J., Harry T. Reis, Joan Gruzen, and Ellen Wolff. 1974. "The 'Eye of the
Beholder': Determinants of Physical Attractiveness Judgments in the U.S. and South
Africa." Journal ofPersonality 42:528-42.

Nemeth, Charlan. 1972. "A Critical Analysis of Research Utilizing the Prisoner's Dilemma
paradigm for the study of bargaining. Advanced Experimental Social Psychology 6:203-
34.

Nowak, Martin. 2004. "Prisoners of the Dilemma." Nature 427:491.

Orbell, John M., Robyn M. Dawes, and Peregrine Schwartz-Shea. 1994. "Trust, Social
Categories, and Individuals: The Case of Gender." Motivation and Emotion 18:109-28.

Owens, Timothy K. 1998. "Effects of Fear and Bias Perceptions of Interracial Cooperation in
Young Males." Dissertation Abstracts International: Section B: The Sciences and
Engineering 59:2488.

Pruitt, Dean G., and Melvin J. Kimmel. 1977. "Twenty Years of Experimental Gaming: Critique,
Synthesis, and Suggestion for the Future." Annual Review of Psychology 28:363-92.

Reskin, Barbra. 2002. "The Proximate Causes of Employment Discrimination." Contemporary
Sociology 29:319-28.

Risman, Barbara. J. 2004. "Gender as a Social Structure: Theory Wrestling with Action." Gender
and Society, 18:429-50.

Rodgers, Reuel. 2000. "Afro-Caribbean Immigrants, African Americans, and the Politics of
Group Identity. Pp. 15-59 in Black and Multiracial Politics in America, edited by Y. M.
Alex-Assensoh and L. J. Hanks. New York University Press.

Rothenberg, Paula S. 1992. Race, Class, and Gender in the United States: An Integrated Study.
St. Martin Press.

Russell, Kathy, Midge Wilson and Ronald Hall. 1993. The Color Complex: The Politics of.\kin
Color among African Americans. Anchor Books.

Sibley, Sally A., Sara K. Senn, and Alexis Epanchin. 1968. "Race and Sex of Adolescents and
Cooperation in a Mixed-motive Game." Psychonomic Science 13:123-124.

Simmel, Georg. 1955. Conflict and the Web of Group Affiliations. The Free Press.









Simpson, Brent. 2003. "Sex, Fear, and Greed: A Social Dilemma Analysis of Gender and
Cooperation." Social Forces 82:35-52.

Sjoberg, Lennart., Ingvar Bokander, Lars Deneck, and Kerstin Lindbom. 1969. "A Comparison
between Children and Adults in the Prisoner's Dilemma Game." Scandinavian Journal
ofPsychology 10:158-66.

Spinney, Laura. 2004. "Why We Do What We Do." New Scientist 183:32-36.

Stockard, Jean, Alphons J.C. van de Kragt, and Patricia Dodge. 1988. "Gender Roles and
Behavior in Social Dilemmas: Are There Sex Differences in Cooperation and in Its
Justification?" Social Psychology Quarterly 51:154-63

Suarez-Orozco, Marcelo M., and Mariela M. Paez. 2002. "The Research Agenda. Pp. 1 in
Latinos: Remaking America, edited by M.M. Suarez-Orozco and M.M. Paez. University
of California Press.

Takaki, Ronald. 2000. Iron Cages: Race and Culture in 19th-Century America. Oxford
University Press.

Thomas, William I. 1931. "The Relation of Research to the Social Process. Pp. 175-94 in Essays
on Research in the Social Sciences: Papers Presented in a General Seminar Conducted
by the Committee on Training of The Brookings Institution, 1930-31. Brookings
Institution.

Tilly, Charles. 1999. Durable Inequality. University of California Press.

Van Lange, Paul A., and David M. Messick. 1996. "Part 1: Social Dilemmas: Psychological
Processes Underlying Cooperation in Social Dilemmas." Pp. 93-112 in SocialAgency:
Dilemmas in Education -Praxiology, edited by W. Gasparski, M. Mlicki & B. Banathy.
Transaction Pub.

Vera, Hernan, and Andrew M. Gordon. 2003 a. Screen Saviors: Hollywood Fictions of Whiteness.
Rowman and Littlefield

Vera, Hernan, and Andrew M. Gordon. 2003b. "The Beautiful American: Sincere Fictions of the
White Messiah in Hollywood Movies." Pp. 113-25 in White out, edited by A. Doane &
E. Bonilla-Silva. Routledge.

Wallace, Ruth A., and Alison Wolf. 1999. Contemporary Sociological Theory: Expanding the
Classical Tradition. Prentice Hal.

West, Candace, and Don H. Zimmerman. 1987. "Doing Gender." Gender and Society 1:125-51.

West, Stephen G., and Jan T. Brown. 1975. "Physical Attractiveness, the Severity of the
Emergency and Helping: A Field Experiment and Interpersonal Simulation. Journal of
Experimental Social Psychology 11:531-38.









Williams, Patricia J. 2000. "Race and Rights." Pp. 421-30 in Theories of Race and Racism: A
Reader, edited by Les Back and John Solomos. Routledge

Wilson, Warner U., and Myra Kayatani. 1968. "Intergroup Attitudes and Strategies in Game
between Opponents of the Same or a Different Race." Journal of Personality and Social
Psychology 9:24-30.

Wilson, William J. 1978. The Declining Significance ofRace: Blacks and Changing American
Institutions (2nd Edition). University of Chicago Press

Yamagishi, Toshio, and Karen S. Cook. 1993. "Generalized Exchange and Social Dilemmas."
Social Psychology Quarterly 56:235-48









BIOGRAPHICAL SKETCH

Victor Eduardo Romano was born in 1979, in Queens, New York. The only son of

Argentine immigrants, he grew up mostly in Miami, Florida, which is where he currently resides

with his wife Kelli.





PAGE 1

1 RACE, GENDER, AND THE PRISONERS DILEMMA: A STUDY IN SOCIAL DILEMMA COOPERATION By VICTOR EDUARDO ROMANO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Victor Eduardo Romano

PAGE 3

3 To Kelli, for always being right and always being on my side

PAGE 4

4 ACKNOWLEDGMENTS I would foremost like to acknowledge my pa rents, Hugo and Patricia C. Romano, for their unwavering love, faith, and support. Ackno wledgments are also due to Dr. Hernn Vera and Dr. Tanya Koropeckyj-Cox for always being ge nerous with their time and encouragement. The guidance and mentorship they have provided me over the years means more than they know. Lastly, I would like to acknowledge my wife an d proofreader, Kelli Romano, for always reading anything at anytime. For her patience a nd care, I am forever grateful.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS... LIST OF TABLES... ABSTRACT. CHAPTER 1 INTRODUCTION ....... 10 Specific Aims Major Research Questions ... 2 LITERATURE REVIEW..17 Race and Cooperation ... Critique of Race in Prisoners Dilemma Literature...19 Theorizing Racial Cooperation..21 Theoretical Basis for Gr eater Black Cooperation..21 Theoretical Basis for Greater Cr oss-Racial Group Defection...23 Alternative Theoretical Arguments: Against Differing Cooperation Levels....26 Gender and Cooperation.... Theorizing Gender Cooperation ...... 29 Gender as a Social Constructi on and Social Structure.. 29 Gender Cooperation and So cial Decision Making. The Intersection of Race and Gender.35 3 RESEARCH METHODOLOGY...37 Research Design.37 Measures and Instruments......39 Experimental Design..40 Recruitment of Participants Demographic Characteristics of Sample....41 Analyzing the Data Role of the Researcher...43 4 QUANTITATIVE RESULTS...45 Hypotheses.45 Quantitative Findings.45 Cooperation Quantitative Analysis of Trust and Expectations...48 The Role of Religiosity..49

PAGE 6

6 Discussion of Quantitative Findings.. 5 OPEN-ENDED DATA RESULTS....57 Open-ended Data Findings....57 External Validity Reasons for Cooperation and Defection White male experiment group White female experiment group.....61 Black male experiment group....63 Black female experiment group.....65 Male cross-race experiment group.66 Female cross-race experiment group.68 Open-ended Data Analysis of Trust and Expectations.. Gender and Race Differences....73 Discussion of Open-ended Data Findings.. 6 CONCLUSION..81 APPENDIX A MEASUREMENT INSTRUMENT..90 B INFORMED CONSENT... LIST OF REFERENCES...98 BIOGRAPHICAL SKETCH...105

PAGE 7

7 LIST OF TABLES Table page 4-1 Proportion of cooperation and de fection by race/gender experiment group .54 4-2 Distribution of Game Pair Ou tcomes by Race and Gender Experiment Group.54 4-3 Proportion of cooperation and defection in exclusively White and exclusively Black experiment groups...54 4-4 Proportion of cooperation and defection by race..54 4-5 Proportion of cooperation and defection by gender..55 4-6 Proportion of cooperation and defect ion by trust in game partner....55 4-7 Proportion of cooperation and defecti on by expectation of game partners decision.55 4-8 Expectation of game partners decision by race...55 4-9 Expectation of game partners decision by race and gender 4-10 Proportion of coopera tion and defection by religiosity....55 4-11 Level of religiosity by race...56 4-12 Level of religiosity by race and gender.... 5-1 Thematic coding categories with frequencies for responses to the question, Why did you cooperate or defect ? Please explain in detail...79 5-2 Reasons for cooperation and de fection by experiment group with frequency of response..79

PAGE 8

8 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy RACE, GENDER, AND THE PRISONER'S DI LEMMA: A STUDY IN SOCIAL DILEMMA COOPERATION By Victor Eduardo Romano December 2007 Chair: Hernn Vera Cochair: Tanya Koropeckyj-Cox Major: Sociology Understanding the intricacies of social dilemmas in general and the Prisoners Dilemma in particular is important due to the myriad of social issues that are manifestations of such dilemmas. Macro and micro issues ranging from global warming and nuclear arms treaties to recycling and gun ownership are at their crux, social dilemmas situations where individual interests are in opposition to and contend with the interests of a larger social unit. Despite the importance of social dilemma cooperation in a world that is becoming increasingly interdependent as a result of gl obalization, there is a dearth of research examining the influence of race on intraand inter-group cooperation. To identify if there is a link between race, gender, and leve ls of cooperation, trust, and expectations of a partner in so cial dilemmas, this study utilizes quantitative and open-ended data collected from undergraduate college student s attending a major public university in the southeastern United States. Drawing on the th eoretical perspectives of games theory, race relations, and gender studies, this study examines the relationship between race, gender and the decision of an individual to eith er cooperate or defect in a mone tary Prisoners Dilemma game. By placing participants in same-race and same-ge nder pairs as well as cross-race pairs the study

PAGE 9

9 shows that Black females have significantly higher same-race group cooperation rates than do White females, and that Black s have significantly higher overa ll cooperation rates than do Whites. The study also shows that cross-race pa irs have higher rates of cooperation than do exclusively White pairs, and that religiosity is positively correlated with cooperation. The reasons given for social dilemma coopera tion and defection by different racial and gender groups were also explored. It wa s found that though reasons for cooperation and defection were mostly similar across groups, mino rity group members were more likely to take their racial and gender group st atus into consideration wh en making game decisions. Additionally, it was also found that Blacks were more likely to view game decisions from a moral perspective and as a reflec tion of individual character th an were Whites.

PAGE 10

10 CHAPTER 1 INTRODUCTION In an era of global conflicts, nuclear prolif eration, global diseases, and environmental destruction and deterioration, th e need to understand the macro and micro dynamics of racial and gender cooperation is paramount. It is not an ex aggeration to suggest that the survival of the human species depends largely on our ability to cooperate with people w ho we see as different from ourselves. Moreover, what Marx (1862) refe rs to as the social productive power, which arises from cooperation, is both societys best tool for the betterment of humankind and as with exploitation, patriarchy, sexism and racism, th e foundation of amongst the most oppressive of social conditions (361). Acco rding to Marx, the essence of simple cooperation remains simultaneity of action, a simultaneity whose results can never be attained by the temporal succession of the activities of the individual workers (1862:215). In other words, many people working together can achieve what one alone can not. Cooperation, as utilized in this study, refers to an individuals decision to attempt to act in solidarity with another person to achieve a mutually beneficial outcome, instead of acting in competition. In the social world, reasons for coopera tion among individual s include economic conceptions such as self-interes t, reciprocity, obligation, and nece ssity. Yet, not all forms of cooperation can be explained by economic theo ries of rational choi ce, which posit that individuals are rational, self-i nterested, economic human beings (Hu & Liu 2003). Ideas such as altruism and loyalty often cause individuals to act against thei r own so-called rational economic interests in favor of pursuing th e interests of others. Volunt eerism, philanthropy, and community activism are a few examples of individuals altruis tically contributing to the welfare of others at their own expense. Furthermore, conceptions of altruistic coope ration can extend beyond individuals to groups (including racial, gender and religious groups ) and nation-states. In this,

PAGE 11

11 cooperation is not merely as Marx stated, the general form on which all social arrangements for increasing the productivity of so cial labour are base d, but also a means of demonstrating solidarity and empathy toward ot her group members (Marx 1862:209). One of the largest stumbling blocks to c ooperation, however, is the proclivity of humans to defect (work against one anot her) in what are called social dilemmas. Social dilemmas are situations in which (a) [defection] yields the person the best payo ff in at least one configuration of choices made by others; (b) [defection] has a negative impact on the inte rests of other persons involved; (c) the collective choice of [defection] results in a defici ent outcome, that is, a result that is less preferred by all pe rsons than the result which would have occurred if all had [cooperated] instead of [defected] (Liebrand 19 86:113-14). Simply state d, a social dilemma is at its essence a situation that pits individual gain against the greater good of the dyad, group, society, and/or humankind. Among the most studied social dilemmas is the game Prisoners Dilemma (Simpson 2003). Bearing in mind that games with strategies, ru les, and often winners are a universal social activity, game theorists analyze human behavior by arguing that many situations can be seen as manifestations of one or other quite simple game (Wallace & Wolf 1999:322). The Prisoners Dilemma is likely the most popular game of game s theory; its popularity is largely due to the inevitable dreadfulness of its most frequent outcome (Wallace & Wolf 1999). The premise of the game is that two prisoners who have comm itted a crime together are under arrest and unable to communicate with each other. In order to force a confession the interrogator offers each prisoner, separately, the following deal: 1. If you confess and your companion does not, he will be sentenced to nine years in prison and you will be let off scot-free.

PAGE 12

12 2. If your companion confesses and you do not, you will be sentenced to nine years in prison and he will be let off scot-free. 3. If you both confess, you will each be sentenced to five years in prison. 4. If neither of you confess you will each be sentenced to two years in prison. More often than not, both prisoners end up confe ssing (defecting) and must each serve five-year sentences. The possibility of ge tting off scot-free, coupled with the fear of being imprisoned for nine years, persuades the prisoners to choose th e third worst individual se ntence and the worst of all in terms of total num ber of years' imprisonment for both prisoners combined. The paradox that makes the Prisoners Dilemma so intriguing is that both prisoners end up defecting even though they both know they would be better-off cooperating (Wallace & Wolf 1999:323). Comprehending the intricacies of social dilemmas in genera l and the Prisoners Dilemma in particular is of critical societal importance due to the large number of so cial issues that can be seen as manifestations of such dilemmas. As Barash (2004) notes, macro social issues such as global warming, water shortages, abuse of public lands, public versus private transportation, and even the Kyoto Treaty can be viewed through the theoretical framewor k of the Prisoners Dilemma. Additional social issues in which the Prisoners Dilemma framework can be applied include recycling, home upkeep, gun control, bl ood donation, littering, an d even micro-level social issues such as recipr ocity in everyday exchanges betw een neighbors (Yamagishi & Cook 1993; Macy & Skovoretz 1998). Considering the myriad of issues that are at their crux social dilemmas, understanding the factor s that contribute to or hind er cooperation is of critical importance becauseas in all of the above mentioned social dilemmaswhen defection becomes the rule, everyone loses.

PAGE 13

13 It is also for this reason that understanding the gender and to a larger exte nt the racial dynamics that influence group cooperation and def ection is important. As Kaplan and Kaplan (2006) eloquently describe: In real life, times of visible changewhe n populations increase, resources decline or new spoils become available for distributioncreate c onditions where the slightest germ of mistrust can rapidly ge nerate a prisoners dilemma. Protestant and Catholic, Serb and Croat, Hutu and Tutsi; two populations in one space can find, even without any great prior enmity between them, that the outcomes of lifes game are suddenly realigning. The majority in a mixed population may still believe that peace and cooperation are best but if a sufficie ntly large minority come to think that its interests are serv ed only by the victory of its own tribe or creed, then this rapidly becomes a se lf-fulfilling assumption. You fear your neighbor might burn down your house; will you wait until he comes with his shadowy friends and their blaz ing brands? No, best call your friends, best find matches and fuel Civil society rapidly cu rdles: individuals lose the chance to choose for themselves. Even the brav e who stand up for peace lose everything, betrayed by their fellow prisoners. With the above scenario in mind, it is surpri sing that the particular effect of race and culture on cooperation is an area of inquiry that has been largely neglected by past studies that employed the Prisoners Dilemma. Since different racial and gender groups place different types and levels of expectations on their members, r acial and gender considerations of group loyalty, obligations, trust, and solidarity can complicat e the individual rationa l choice process of choosing social dilemma cooperatio n and defection. Furthermore, if social dilemma cooperation or defection levels are higher among certain soci etal groups (as my research shows they are), what factor(s) can this difference be attributed to? This is a questi on that qualitative and theoretical Prisoners Dilemma research on race, gender, and their intersectionality can help to answer, and which I will explore furt her in this dissertation. Specific Aims To identify if there is a li nk between race, gender, and leve ls of cooperation, trust, and expectations of a partner in soci al dilemmas, I collected and anal yzed data from college students

PAGE 14

14 attending a major university in th e southeastern United States. Th e goal of the research was to examine if and how race and gender are correlated to leve ls of cooperation, trust and expectations of a partner through the use of the theore tical framework of games theory and the social dilemma game Prisoners Dilemma. Above all, the specific aims of this study were to: (1) determine if Black1 college students, as members of a subor dinated group, are more likely to cooperate in a social dilemma than are White college students who are members of the dominant group; (2) determine if female college students are more likely to cooperate with each other in a social dilemma than are male college students; and (3) to explore the r easons and rationale gi ven by respondents for cooperation or defection. By cr eating a monetary variant of th e Prisoners Dilemma game for the participants of this study, I examined the role of race and gender in participants decision to cooperate or defect. I found that Black fema les have significantly higher same-race group cooperation rates than do White females, and that Blacks ha ve significantly higher overall cooperation rates than do Whites. The study also s hows that cross-race pairs have higher rates of cooperation than do exclusively White pairs and th at religiosity is posit ively correlated with cooperation. Drawing on works such as Axelrods (1984) The Evolution of Cooperation and Gilligans (1982) In a Different Voice this study contributes to our knowledge of games theory, social dilemma theory, social psychological theo ry, rational-choice theory symbolic interaction theory, as well as the fields of racial and ethni c relations and gender stud ies. Finally, this study strives to provide insight into the ways in which men and women reason and cooperate and how race may influence these processes. 1 I employ the racial classi fications Black and White rather than ethnic classifications of African American and Caucasian in acknowledgement of the diverse origins of my sample population. Additionally, the use of the racial

PAGE 15

15 Major Research Questions Though many scholars have employed Carol Gill igans (1982) work on differences in the moral development of women and men as the fou ndation from which to project differences in cooperation in the Prisoners Dilemma (including Stockard, va n de Kragt & Dodge 1988; Brown & Taylor 2000), inconsistent re sults and contradictory findings suggest that the question of whether gender is associated with cooperation in the Prisoners Dilemma and other social dilemmas remains unanswered (Ledyard 1995). Additionally, the role played by race in an actors decision to cooperate or defect has been largely absent from the research literature and needs to be further explored. In this study, I examine the relationship between race, gender, and the tendency of an individual to either cooperate or defect by analyzing decisions made by samerace and same-gender pairs of pa rticipants, as well as cross-r ace pairs of participants in a monetary Prisoners Dilemma game. As a student specializing in racial relations, I was surprised at the lack of Prisoners Dilemma studies employi ng race as a variable and exploring the effect of race, gender, and their intersectionality on social dilemma cooperation. My knowledge of groups led me to believe that racial and gende r group membership (each with their own varying expectations of group loyalty and obligation) would influenc e participants to be more cooperative when paired with a member of thei r same race and gender group. Additionally, I felt that minority group status would result in greater dyad solidarity and thus a higher probability of social dilemma cooperation. These views, buttressed by social theo ry led me to my particular research questions, which are: 1. Is there a correlation between race and like lihood of cooperation in a social dilemma? term Black places emphasis on race as referring to the phys ical, phenotypic, and visible differences that are given social significance in American society. When discussing the research of ot hers, however, I keep with the racial classifications they employed.

PAGE 16

16 2. Is there a correlation between sameness of gender and likelihood of cooperation in a social dilemma? 3. What are the reasons and rationale given by participants for choosing to cooperate or defect? 4. Do levels of trust and expect ations of a game partners decision differ between race and gender groups? To explore these questions, Chapte r 2 reviews the relevant previous Prisoners Dilemma research and discusses the theoretical suppor t for varying levels of cooper ation among different racial and gender groups. Based on the theoretical perspectiv es discussed in Chapter 2, Chapter 3 describes the design and methodological under pinnings of the experiment and poses five hypotheses. Chapter 4 specifies and discusses the studys qua ntitative results, whic h yield support for the claim that Blacks are more likely to cooperate in social dilemmas than are Whites. Chapter 5 examines the validity of the experiment design and then specifies and discusses the open-ended questionnaire results, which describe distinctions in the ways the various race and gender groups rationalize their game decisions. Chapter 6 come s to a conclusion and provides suggestions for future research.

PAGE 17

17 CHAPTER 2 LITERATURE REVIEW The literature on the Prisoners Dilemma is astonishingly exte nsive and varied. Researchers in the fields of sociology, psychology, economics, political science, criminology, mathematics, biology, zoology, and others have used various versions of the Prisoners Dilemma to study a plethora of issues. One intriguing ex ample of the breadth of research conducted incorporating the Prisoners Dilemma included using functional magnetic resonance imaging to scan the brain of women playing the Prisoners Dilemma to gain insight into the decision making process by looking at what brain regions were activated during game play (Spinney 2004). Others have used the Prisoners Dilemma to explain why certain foraging animals share food without any overt aggression (D ubois & Giraldeau 2003) or how an attractive experimenter or game partner affects a participan ts decision on whether or not to defect (Morse, Reis, Gruzen & Wolff 1974). There is even a Prisoners Di lemma world-championshi p competition in which researchers from many disciplines submit comput er programs to a repe titive version of the Prisoners Dilemma in a round-robin tournament (Axelrod 1980). Interestingly, the most successful strategy, tit-for-tat2, is also one of the simplest (Axelrod 1980). Though both the variation and uses that the Prisoners Dilemma has undergone are exhaustive, the Prisoners Dilemma literature most relevant to this study can be grouped into two manageable categories. The first category, Race, is exceedingly small and contains research on how race correlates with coopera tion/defection in Prisoners Dilemma games. The second and larger category, Gender, contains resear ch addressing how gender correlates with cooperation/defection in the Pris oners Dilemma. Below is a brief overview of the Race and 2 The tit-for-tat strategy in the Prisoners Dilemma game co nsists of a player cooperating on their first move and then copying the previous rounds cooperation/defection decision of their opponent for the duration of the game.

PAGE 18

18 Gender categories each followed by a discussion of the theoretical foundations for predicting variance in cooperation rates between racia l/gender groups. Race and Cooperation The research literature on how race affect s cooperation in the Prisoners Dilemma is miniscule. Early research on race and the Pr isoners Dilemma shows that participants demonstrate higher rates of cooperation when they play against a member of the same racial group (Sibley, Senn & Epanchin 1968; Wilson & Ka yatani 1968; Baxter 1973). Furthermore, Cederblom and Diers (1970) showed a tendency among White students to behave in a less cooperative manner toward cooperative Black game partner confederates3 than toward cooperative White game partner confederates. Additionally, Cederblom and Dies showed that Whites were more cooperative toward Black confed erates who at first displayed competitiveness and later cooperated, versus Black confederat es who mostly cooperated throughout the game play. These findings were take n by the researchers to mean th at Black assertiveness is both understandable and necessary in order to affect institutional change. In a later study of 42 same-sex pairs of eighth-grade students of varied (Black and White) racial composition, Downing and Bothwell ( 1979) discovered that females of either race tended to cooperate in same-race pairs and to compet e in mixed-race pairs. They also found that White males tended to cooperate and Black males to compete independent of their partners race. Owens (1998) designed a study to explore several facets of cross-racial interactions of young male African Americans and Caucasian Americ ans, finding evidence that for both African Americans and Caucasian Americans cooperation was in certain cases correlated with the participants respective stage of racial identity de velopment as measured by Racial Identity Scale 3 I use the term confederate to refer to a participants ga me partner who was (unbeknownst to the participant) hired by the researcher and typically employ ed a predetermined ratio of coopera tion and defection responses.

PAGE 19

19 (RIAS) or White Racial Atti tude Identity Scale (WRIAS ) (1988:2488). Although Owens study provides interesting insight into psychological aspects of cross-ra cial cooperation, it does not measure or compare same-race group cooperation rates, which is one of the aims of this present study. More recently, Heider and Skowrons ki (2007) found that in two 50-shot4 Prisoners Dilemma games (designed based on procedures adapted from Baxter [1973]) Caucasian participants unexpectedly cooperate d at a higher rate when presented with a computer image and name of an African American partner (M = 51.8 %) than when presented with a computer image and name of a Caucasian partner (M = 48.4%). They also found that measures of racial attitude such as the Implicit Associati on Test (IAT) and the Pro-Black subscale of the Pro-Black/AntiBlack Attitudes Questionna ire (PAAQ) predicted social behavi ors of Caucasian participants toward African American targets. In the following section, I offer a critiq ue of past studies and highlight some of their limita tions and weaknesses. Critique of Race and Prisoners Dilemma Literature A major problem with many of the past stud ies using the Prisoners Dilemma to measure racial attitudes and cooperation rates is the use of exclusively White part icipants. In doing so, studies such as Baxters (1973), Cederblom a nd Diers (1970) and Heid er and Skowronskis (2007) privilege the views and decisions of White s over those of Blacks. Moreover by recruiting only White participants, these studies provide no insight into Black same-race and cross-race cooperation. Another problem with past studies is the exclusive use of multi-shot games. Minkler and Miceli (2004) have noted that by employing a on e-shot game version of Prisoners Dilemma, rather than a multi-shot game version, you can assu re that any resulting cooperation or defection

PAGE 20

20 does not come from reputation or punishment effects associated with repeated games. Additionally, the one-shot version employed by this study protects against players cooperating or defecting as the result of a refined strategy deve loped over multiple plays. In terms of validity and generalizability, a one-shot game may also prove more similar to social dilemmas encountered by people on a daily basis. The decisi on of an individual to li tter, cut in front of a line, or use an express lane of a supermarket with more than the maximum allowed items is a decision that is typically made without consideration to past or future decisions in the same social dilemma-style situation. Also, the law of diminishing marginal utility suggests that a participant will place more sign ificance on a one-shot Prisoners Dilemma decision, than on, for example, the 47th-shot decision of a Prisoners Dilemma ga me. This suggestion is supported by the work of Clark and Sefton (2001) who found th at the most important variable influencing cooperation in sequential multi-shot Prisoners D ilemma games is the first-movers choice. Their study yields support for th e argument that cooperative be havior in multi-shot social dilemmas reflects reciprocation rather than un conditional altruism. Moreover, they found their cooperation in Prisoners Dilemm a games decreases with repetition. Hence, single-shot Prisoners Dilemma games are also better barometers of altruistic behavior than are multi-shot games. As will become evident in the Methods Chapte r (3), I seek to address these deficiencies of past research in this area. In the next section, I will explore so me of the theoretical foundations for possible differences in cooperatio n rates between different racial groups in a social dilemma, as well as the theoretical foundati ons for no variation by raci al groups in a social dilemma. 4 The word shot refers to how many iterations of the game were played.

PAGE 21

21 Theorizing Racial Cooperation Though much has been written on the rationale behind actors decisions to c ooperate or defect in social dilemmas, the possible influence of race on these decisions is a topic that has not yet been thoroughly explored. While race is a fluid and ever-changing soci al construct that should no longer be used as a system of classification in e ither the biological or so cial sciences, it is a concept that is nevertheless accep ted as a real and meaningful distinction by most Americans (Vera & Gordon 2003a). W.I. Thomas (1931) noted th at, situations that ar e defined as real, are real in their consequences, hence, since most Americans define race and racial groups as real, being categorized into a partic ular racial group has multitude of real consequences for an individual that will affect th eir life outcomes and perspectiv es (145). This study seeks to examine what influence, if any, racial group me mbership has on a partic ipants decision to cooperate or defect with another participant of either the same or a different racial group. For this purpose, the study employs Feagin and Feagin s (2002) definition of a racial group as a social group that persons inside or outside the group have decided is important to single out as inferior or superior, typically on the basis of real or alleged p hysical characteristics subjectively selected (7). Despite the paucity of research in this area, there is a sufficient theoretical basis to suggest that racial group affiliati on may influence a participants de cision to cooperate or defect in a social dilemma. Theoretical Basis for Greater Black Cooperation I begin with the hypothesis5 that Black pairs will have higher same-race group cooperation rates than White pa irs (referred to later as hypothe sis 1). This hypothesis is grounded in several historical and sociopolitical considerations, th e most important of which is 5 All hypotheses are formally presented in Chapter 4 (page 45).

PAGE 22

22 the Black subjugation and response to White racism White racism as defined by Feagin, Vera, and Batur (2001) can be viewed as the socially organized set of practices attitudes, and ideas that deny blacks and other people of color the privileges, dignity, opportu nities, freedoms, and rewards that this nation offers to White Americans (17). As Simmel (1955) noted in his classic work Conflict and the Web of Group Affiliations threats to a group (such as White racism) have the function of increasing group solidarity and causing group members to transcend individual in terests in favor of the moral imperative to defend the group. As the level of violence against a group escalates, so does the groups internal solidarity. Assensoh (2000) sums up this view, stating nothing is known to bind and promote both collaboration and cooperation more than y ears and shared experi ence of fighting common and easily identifiable enemies (115). Examples of this increased solidarity and intra-group cooperation among Blacks reacting to the social dilemma of White racism can be found in the work of McAdams (1995) and Chong (1991). Both scholars note that various forms of social protest employed by Blacks during the civil rights movement (such as demonstrations, sit-ins and boycotts) were all extremely cos tly to the individuals who participated in them, though it was known to all that any gains made by the proteste rs would benefit all Blacks in the community whether they participated or not. To explain this phenomenon, Chong contends th at social incentives such as increased social standing and avoidance of ridicule and ostracism for not participating worked as the impetus for participation in social protests among small social groups like Black Churches. Though social incentives also influence racial groups other than Black Americans, Americas racist past has provided Black Americans (compared to White Americans) much more motivation to sacrifice individual ga in for the good of the community. I contend that the Black

PAGE 23

23 history of subjugation to slaver y, Jim Crowism, and the continui ng racial discrimination faced by Blacks today, has created a culture of cooperatio n among Blacks that is rooted in what Dawson (1994) argues is a collective hist orical memory of racial oppressi on and sense of shared racial fate. He writes, [T]he collective memory of the African American community continued to transmit from generation to generation a sense that race was the defining interest in the individuals life and that the we ll-being of blacks individually a nd as a group could be secured only by social and political agitation [that memo ry] has been reinforced historically (Dawson 1994:58). Since individual participation in social and political agitation can itself be viewed as a social dilemma situation, it is reasonable to suggest that this coll ective memory may also increase intra-group cooperation among Blacks. This outcome is also supported by what McCormick and Franklin (2000) have called the inclusionary dilemma, a dialectical process wherein emergent racial conscious ness on the part of African Americans can be seen as response to the racial impediments that African Americans encounter as they attempt to gain access to the economic, social, and political opportunity stru cture of the larger society (330). In contrast to Blacks, Whites have no e quivocal long-standing hi storical memory of oppression and racism in the United States (Rodgers 2000). Moreover, in a nation where Whiteness is taken to be the norm and therefore beco mes invisible, it is no surprise that Blacks tend to be more conscious and aware of their r acial group status than are Whites (Doane 2003). What past research has failed to examine is wh ether or not that awareness will translate into increased racial group coopera tion in social dilemmas. Theoretical Basis for Greater Cross-Racial Group Defection In addition to the obvious and very pre ssing problems of racism, prejudice, and discrimination that plague American society and that may manifest themselves as increased

PAGE 24

24 defection in a cross-racial pa iring during a social dilemma ga me, there is also significant historical and sociological s upport for greater cross -racial group defection between Black and White actors (referred to later as hypothesis 5). American labor history has an array of examples where when faced with a social dilemma, White workers chose defection (and its benefits) over cooperation with Blacks and/or other minority groups (Feagin 2000). Though cooperation by White workers would have significantly increased the likelihood of higher wages for all workers White and Black, defection maintained what Du Bo is (1935) and later Feag in (2001) describe as the psychological wage of white ness (Du Bois:7 00; Feagin:27). The psychological wage of whiteness as presented by Feagin (2001) is a complicit support for a White dominated racially hierarch ical society in which working class Whites receive superior privileges, opportun ities and cultural resources than do minorities, at the cost of lower wages from the White elites who employ bo th. In other words, White workers choose the benefits of being White, over the greater econo mic wages they might have had if they had joined in strong organizations wi th black workers (Feagin 2001:30). Historical examples of this social dilemma can be found at numerous times in American history: in Iron Cages Takaki, (2000) recounts how the 1870s strike by The Secret Order of th e Knights of Crispin an allWhite labor union founded in 1867, failed because of the importation of Chinese laborers to replace the White union members. By excl uding minorities from their labor union and workplace, white laborers did not improve their bargaining position, but detrimentally undermined it. By the time the Crispins tried to recruit Chinese workers into their union it was too late; the strike would soon end with the un ion having lost (Takaki 2000:239-40). In this example, racial stratification was the primary stratification because it decentered class oppression in the thinking and orientation of most White wo rkers (Feagin 2001:31). This

PAGE 25

25 emphasis on race was of course the paramount goa l of the White industrial elite who wanted White workers to buy into a racist ideology and its psychological wage s, so as to divert attention from the fact that these elit es were exploiting both White an d Black workers (Feagin 2001). Labor movements, however, are not alone in their choices to exclude minority participation in favor of psychological wages and privilege. The U.S. womens rights movement of the late 19th century dismissed the concerns of bl ack women and developed segregated interests and organizations (Feagin 2000:31). In stead of working for the betterment of women of all racial groups, this movements negati on of minority womens issues undermined its objective of equality, fair pay and womens right s. Though since the civil rights movement, the psychological wage of Whiteness and White privilege have become less de jure and overt, they nevertheless still exist, albeit in more de f acto and hidden forms (Feagin 2000; Bonilla-Silva 2003). The existence of both the psychological wa ge of Whiteness and of White privilege may make Whites more likely to cooperate with each other and less likel y to cooperate with Blacks in a social dilemma. Also relevant to the hypothesis that there wi ll be greater cross-raci al group defection in a social dilemma is the statistical theory of di scrimination, which posits that people use race as a proxy to evaluate others by assuming in the abse nce of more specific information that the person has the average qualities of her race (McAdams 1995:1021). Considering that negative stereotypes of Blacks are broadcast to American television and movie viewer s on a daily basis, it would not be surprising to find th at in our highly segr egated society, many White people reify these stereotypes and take them to be true, ther efore making them less likely to cooperate with a Black person about whom little is known (Hacker 1995).

PAGE 26

26 Conversely, among White dyads facing a social dilemma, what Vera and Gordon (2003a) have identified as sincere fictions of the white self may lead to incr eased cooperation (15). Sincere fictions of the white self are deliberat ely constructed images of what it means to be white (Vera & Gordon 2003a:15). Many television programs and film s are sincere fictions in so far as they are rooted in the self-concept that we seldom examine, that we take for granted (Vera & Gordon 2003a:16). In other words, these fictions are embraced w ithout considering real or potential alternatives to them. The white self is the concept of white Americans proposed by white American moviemakers, who present Wh ites as morally and intellectually superior persons who are powerful, brave, cordial, kind, firm, good-looking, generous [and] natural born leaders (Vera & Gordon 2003b:114). Since people de fine themselves in relation to others and considering that many Whites internalize these sincer e fictions as true, it would not be surprising to see Whites more likely to cooperate with one a nother in a social dilemma, than with a Black person. As social science resear ch continues to show, groups ma tter people have a loyalty to groups that goes beyond what serves their na rrow pecuniary self-i nterest (McAdams 1995:1084). Alternative Theoretical Arguments: A gainst Differing Cooperation Levels Despite the theoretical and historical s upport for differing levels of social dilemma cooperation by race, there is also theoretical sup port to the contrary. The argument for different levels of cooperation hinges on a group interest ve rsus individual interest argument. Simply put, for reasons described above, group interest is more likely to affect th e social decision making process in favor of cooperation for Blacks than it is for Whites. That being said, it is difficult to create experimental designs in which pairs of participants take racial-group identity and

PAGE 27

27 allegiance into consideration without overtly or co vertly revealing to them that their decisions will be compared to that of participants from di fferent racial groups. As McAdams (1995) reports, in same-race pair s in which participants know they are being compared to pairs of another racial gr oup, the proxy effect encourages cooperation since participants may speculate that they will be eval uated in pairs and thus go for the best pair outcome cooperation. Furthermore, Blacks may experience a stronger desire than Whites to achieve the best pair outcome in order to comb at racial stereotypes a nd racist theories of inferiority. Not revealing that they are being compared to othe r racial groups protects against this bias, however, it also poses th e risk that racial-group interest wi ll not be seen as relevant to the decision making process. It is for this reason that the soci al location of the study and its participants is of critical im portance. Factors such as ag e, education, income, geographic location, and experiences of racism can all influence the ex tent to which group interest is seen as important or relevant to the decision to cooperate /defect. The role of so cial desirability should also not be discounted. Particip ants desire to be pe rceived by both their game partner and the researcher as good, nice, and non-r acist people may also increase their likelihood of cooperation. Another theoretical considera tion that could negate any r acial-group differences is the idea that White racism is also a form of coopera tion. So while I have proposed that Blacks are more likely to have higher cooperation rates as a result of responding to White racism, one could contend that Whites have forged an identity and cooperative spirit based on centuries of practicing white racism in both its de jure and de facto forms. Also pertinent is what critical theorist Derrick Bell (1980) calls the Interest-Convergence Hypothesis. According to Bell, whites support the cause of equality and justice for blacks only when it is in their interest to do so (Bell 1980: 252). In a so cial dilemma such as the Prisoners

PAGE 28

28 Dilemma, White participants could easily see it as in their best interest to cooperate with Black participants for a mutually bene ficial outcome. If this is the case, then different levels of cooperation will be unlikely. Finally, theoretical support against lower cr oss-racial group cooperati on levels and larger cross-racial group defection rate s can be found in Wilsons (1978) declining significance of race argument. In his classic work The Declining Significance of Race Wilson contends that in the modern industrial period, Black subordination and social advancement are more associated with economic class, than with race. Following this argument, cooperation/de fection in a social dilemma could be more closely associated with cl ass status than with ra ce, and any measures of cooperation should closely control for th e class status of the participants. One final concept that is germane to the i ssue of social dilemma cooperation/defection is that of White racism as societal waste. F eagin (2001) notes that th e tremendous amount of talent, energy, and resources that are spent in perpetuating White privilege are an extravagant waste that ultimately benefit no one and which ha ve the potential to destroy society. Similarly, individual and mutual defection in a social dilemma ultimately results in a waste of resources that in the long term benefits no one. In the plethora of social issues that can be viewed through the lens of a social dilemma, when defection becomes the rule, everyone loses. It is for this reason th at understanding if and why different racial groups are more likel y to defect or cooperate is important. Gender and Cooperation In contrast to the literature that examines racial cooperation in the Prisoners Dilemma, the literature that examines gender cooperation in th e Prisoners Dilemma is vast; it is also often times contradictory. Prisoners Dilemma gender cooperation results vary according to the time

PAGE 29

29 the study took place, the geographic location of th e study, as well as the studys design. Hence, certain studies have found evidence that females have higher cooperation ra tes (see Sibley, Senn & Epanchin 1968; Fisher & Smith 1969; Bonaci ch 1972; Smith, Vernon & Tarte 1975; Dawes, McTavish & Shaklee 1977; James, Soroka & Be njafield 2001; Hu & Liu 2003), while others have found support for the converse (see Kahn, Hottes & William 1971; Kahn, Hottes & Davies 1971; Sell & Wilson 1991; Brown-Kruse & Hummels 1993). Most studies, however, have found no significant differences for the cooperation ra tes of women versus th ose of men (Caldwell 1976; Goering & Kahn 1976; Orbell, Dawes & Schwar tz-Shea 1994; Sell 1997). The following section will revi ew some of the theoretical support for the hypothesis that women will have higher cooperati on rates than will men in a m onetary Prisoners dilemma. Theorizing Gender Cooperation Gender as a Social Construction and Social Structure In The Social Construction of Gender Lorber (2001a) notes that gender is so pervasive in our society that we assume it is bred into our genes (as cited in Shaw 2001:121). Gender, however, is not a genetic attribute; it refers to the personal tr aits and social positions that members of a society attach to being female or male (Macionis 2005: 325). Furthermore, conceptions of gender norms are fluid, varyi ng, and changing across cultures and over time; gender is constantly created and reproduced through human inte raction a process commonly referred to as doing gender (West and Zimme rman 1987). At the individual level, gender construction begins at birth (and possibly even before, thanks to medical technologies such as sonograms) when newborns are assigned a sex ba sed on the appearance of their genitalia. The newborns sexual classification quickly become s a gender status through naming, dressing, and the use of other gender indicators (Lorber 2001a). People respond to these indicators according

PAGE 30

30 to gendered norms and the child soon learns a nd internalizes their assigned gender identity and begins routinely doing (performi ng) gender for themselves (Butler 1990; Messner 1990). At an early age gender identity-construction is accomplis hed by the individual who then, according to West and Zimmerman (1987), becomes hostage to its production. In addition viewing gender as a social cons truction, it can also be seen as a social structure. Risman (2004) makes the case for conceptualizing gender as a social structure6 (others such as Patricia Martin (2004), prefer the term institution) and using this conceptualization as a tool for analyzing the ways in which gender is grounded in the individua l, interactional, and institutional dimensions of societ y. Focusing on gender as a social structure also suggests that gender can be a constraint that opposes i ndividual motivation (B lau 1977). Gender is structurally buttressed through gendered norms and expectations that are enforced through informal sanctions of gender-inappropriate beha vior by peers and by formal punishment by those in authority should behavior de viate too far from socially imposed standards for men and women (Lorber 2001a:123). The function of gender categories is (accord ing to some feminist scholars) to keep women subordinated through a system of stratification so that they can be exploited in both the marketplace and the household (Lorber 2001b). Despite the glari ng inequality gender classifications produce, they persist because they are deeply embedded as a basis for stratification in our individual personalities, cultural norms, and societal institutions (Risman 2004). Thus, the status inequalities that gende r divisions create can potentially manifest themselves in an individu als social decision making. 6 Social structure is defined as any relatively stable pattern of social behavior (Macionis 2005:655).

PAGE 31

31 Gender Cooperation and Social Decision Making Many scholars have employed social learning theories, such as Carol Gilligans (1982) work on differences in the moral development of women and men, as the foundation from which to project differences in cooperation in soci al decision making games such as Prisoners Dilemma (Brown & Taylor 2000; Stockard, va n de Kragt & Dodge 1988). According to Gilligan, boys are taught to hold a justice view of morality that emphasizes formal rules and the rights of individuals under thos e rules, while girls (because of socialization) tend to favor a care view moral perspective that emphasizes so lidarity, community, and close relationships. Additional theoretical support fo r increased cooperation rates among women can be found in the work of Janet Lever (1978), w ho after examining single-sex pl ay groups found boys favor games that nearly always have winners and losers; these games reinfor ce masculine traits of aggression and competition. Conversely, girls are more likely to play games (such as jump rope, hopscotch, dancing, and singing) where vict ory is not the ultimate goal a nd interpersonal skills of communication and cooperation are fostered. This finding is importa nt considering the construction of the Prisoners Dilemma game and its flexibility to end in either a draw or with a clear winner and loser. Applying Levers conc lusion to Prisoners Dilemma suggests that women are more likely to cooperate and play for the draw that most benefits both players, and that men are more likely to attempt to out-compete one another by defecting. As theoretical foundations for hypothesizing ge nder differences in social decision making scenarios the above explanations fit nicely with the social co nstruction and social structure theories of gender described earlier. Though gende r is a social construc tion, it is nonetheless considered as a real and meaningf ul distinction by most people. Since, as discussed earlier, situations that are defined as real, are real in th eir consequences, the societal belief that women

PAGE 32

32 place more value on interpersonal relationships, friendships, solidarity, and cooperation than do men who in turn are believed to favor aggression, domination, and competition becomes a selffulfilling prophecy (Thomas 1931:145). In othe r words, since as women and men we are expected to act in certain gende red ways, we internalize that e xpectation and act accordingly. Similarly, gender as a social structure can both act as a constraint that tempers individual motivations to cooperate/defect ac cording to socialized gender roles or as a prescribed response to the cooperate/defect decision. Brent Simpson (2003), however, argues that previous research theoretically grounded on work by Gilligan (1982) and other feminist schol ars often fails to find gender differences in cooperation levels because researchers have consis tently used the Prisoners Dilemma game to investigate if a difference exists. Simpson cont ends that the Prisoners Dilemma is problematic because several theories of gender/sex related beha vior support the notion that females are more likely to defect out of fear (the prospect that ones cooperation may be e xploited) while males are more likely to defect out of greed (the tempta tion to free-ride on othe rs cooperation) (Simpson 2003:36). Since the Prisoners Dilemma contains both fear and greed, Simpson argues that one should not expect to observe sex differences in this classic greed and fear version (his study supports this hypothesis). Instead, Simpson predicted that in a dilemma that emphasized the fear component, fewer females than males would c ooperate. In Simpsons study, this was not supported by the evidence since there where equa l levels of cooperation among men and women. Lastly, in a dilemma that emphasized greed, Simp son predicted males would be less likely than females to cooperate. In this greed dilemma st udy, he found that 57% of females versus 33% of males cooperated, thus supporting his hypothesis. In a repeat of this study using similar methods, Simpson found similar results.

PAGE 33

33 Simpson brings to light some interesting c onsiderations; his work shows that for all but the original (greed and fear) version of the Prisoners Dile mma, women in his study cooperated more than men despite the fact that their coop eration levels often did not reach the level of statistical significance. This be gs the question of whether a la rger and broader sample would have yielded differing results. Also questionable is whethe r Simpsons tweaking of the Prisoners Dilemma to create a fear dilemma and a greed dilemma is valid and is actually measuring fear and greed influences on cooperati on levels. Since he does not ask participants the reason for their defection, we can only theorize as to their real motives. In contrast, the openended questionnaire component of th e present study allows for an examination of the reasons and rationale participants give fo r cooperating and defecting. A final important factor that is presen t in the work of Simpson and many others employing the Prisoners Dilemma in their research, is that study participan ts are not allowed to see or meet one another throughout the experiment. As a result, participants have no ability to take racial and gender group membership into acc ount when playing against a person they have never seen. By allowing participants to see the person they will play the monetary Prisoners Dilemma game with before the game instructio ns are explained to each separately, this study gives insight into whether race and/or gender ar e somehow being used as a proxy to gauge trust and altruism. It should also be noted that additional theoretical grounding for the hypothesis (referred to later as hypothesis 2) that women pairs wi ll have higher same-gender group cooperation rates than male pairs in a Prisoners Dilemma ga me, can be based on several historical and sociopolitical considerations, the most important of which is the existence of patriarchy and womens resistance to this soci al-structural system of domi nation. Hence, Simmels (1955)

PAGE 34

34 work on Conflict and the Web of Group Affiliations can also provide theoretical grounding for the view that patriarchy and sexi sm, if viewed as a group threat to women, may have the function of increasing female group solidarity and fost ering group members to transcend individual interest in favor of the moral imperative to suppo rt the group. From this perspective, womens subordinate social position will result in higher female cooperation levels rather than higher female defection levels that are based on fear and self-protection, as is Simpsons contention. Thus, in the case of the monetary Prisoner s Dilemma game I employ, increased female cooperation theoretically hinges on whether women see them selves as a subjugated and/or oppressed group. In summation, the primary th eoretical rationale past re searchers have employed to predict gender differences in the Prisoners Dilemma has typically been that womens socialization would result in higher cooperati on rates for women relative to men (Stockard, van de Kragt & Dodge 1988; Brown & Taylor 2000). I propose that in addition to socialization, female experiences of subjugation under a patria rchal system may also motivate greater female cooperation in the Prisoners Dilemma. Inconsiste nt results and contradi ctory findings, however, suggest that the question of wh ether gender affects cooperation in Prisoners Dilemma and other social decision making games remains unanswe red (Ledyard 1995). By employing traditional experimental and quantitative methods along with more qualitative methods (that allowed for the analysis of open-ended responses to the questi on of why participants chose to cooperate or defect) this study was able to further explore Prisoners Dilemma gender differences and better elucidate how gender influences the deci sion-making processes of participants.

PAGE 35

35 The Intersection of Race and Gender Various scholars (particularly Women of co lor) have pointed out that when considering the system of gender stratification, it is valuable to consider vari ous other dimensions of social stratification such as race and class since all of these systems of oppression operate simultaneously, intersect, and can have cumulative effects (Anderson and Collins 2001; Carby 2000; Collins 2000; Williams 2000; Baca Zinn 1996; Delgado 1995; Baca Zinn and Thorton 1994; Frankenberg 1993; Rothenberg 1992; hooks 1991; King 1988). In light of this, some have noted that the theoretical divisions among scholar s who specialize in race or class or gender or queer theory have undermined a mo re sophisticated analysis of inequality (Reskin, 2002; Tilly, 1999). Others such as Collins (1998), Calhoon ( 2000) and Risman (2004) advocate a both/and strategy that accents the need to understand gender structure, race structur e, and other structures of inequality as they currently operate while also systematical ly paying attention to how these axes of domination intersect (R isman 2004:429). Though gender and race are social constructs they are nevertheless universally us ed to rationalize inequality. The implications of intersectionality rese arch for Prisoners Dilemma studies are that racial and gender considerations are intertwined, and that participants race and gender status cannot be divorced from one another. In othe r words, Black women cannot respond separately as Black and as women, but rather these two statuses work in tandem in shaping social perceptions and responses. The ra ce and gender status of a partic ipants game partner may also influence social perceptions and expectations in ways that vary across race and gender categories. As a result, there is a need to shif t from research that generally examines race or gender but not both together, to research that ex amines both race and gender and the intersection

PAGE 36

36 of these statuses. By doing so we can learn if one status dominates or if the intersection of race and gender creates unique perceptions and res ponses for different race and gender groupings7. In the following chapter, I present the met hodology employed in this study and describe a research design that was created to measure both race and gender group differences in Prisoners Dilemma cooperation/defection rates, while al so addressing the need for intersectional exploration. 7 The intersection of race and gender co mpose the theoretical basis for hypoth eses 3 and 4 presented in Chapter 4 (page 45).

PAGE 37

37 CHAPTER 3 RESEARCH METHODOLOGY Research Design In order to examine differences in levels of cooperation betwee n Blacks and Whites and women and men in a social dilemm a, an experimental design base d on a monetary version of the Prisoners Dilemma game was created. This specif ic research design makes it possible to test the following key questions: 1. Is there a correlation between race and likeli hood of cooperation in a social dilemma? 2. Is there a correlation between gender and like lihood of cooperation in a social dilemma? 3. What are the reasons and rationale given by par ticipants for choosing to cooperate or defect? 4. Do levels of trust and expect ations of a game partners decision differ between race and gender groups? To investigate these questions a sample with the following race and gender composition was recruited: 36 White males 36 White females 36 Black males 36 Black females This created a total sample size of 144 particip ants, split into the following six experiment groups: 1. White male group ( n = 24) 2. White female group ( n = 24) 3. Black male group ( n = 24) 4. Black female group ( n = 24) 5. Black and White male cross-race group ( n = 24) 6. Black and White female cross-race group ( n = 24)

PAGE 38

38 The purpose of creating these experiment groups and dividing participants accordingly was to allow the researcher to compare and contrast the cooperation/defec tion decisions and rationale of participants in a given group with those of participants in differe nt groups. Participants in the study were scheduled in dyads a nd played a one-shot game of th e Prisoners Dilemma with a member of their same-gender group. For race, both same-group and cross-race group differences were measured and analyzed. Cross-gender gr oup differences were not examined due to financial constraints and the multitude of previous studies that have explored this question. At the beginning of the experiment (after participants read an d signed an informed consent form8), participants were briefly allowed to see one another and to ld that they would both soon be participating in a so cial decision making game. They were then separated and told that the person they would be pl aying the game with was the same person they just saw and that they would leave the experiment in isolation (and not encounter th e person they played) once the game was complete. Next, the participants were escorted to separate ro oms where they received an extensive briefing on a monetary Prisoners D ilemma game until the re searcher was confident that they understood the various possible choi ces and the consequences and payoffs of each combination of choices. During this time they were told that if both players cooperated they would each receive $6. If they both defected, they would each r eceive $3, and if one cooperated and the other defected, the person who coopera ted would receive $0 while the person who defected would receive $9. Like the classic vers ion of the game, which potentially punishes cooperation9 with a nine-year prison term, the prosp ect of not receiving compensation for your time in this monetary Prisoners Dilemma game se rved to temper players desire to cooperate. 8 The informed consent form is located in Appendix B 9 As mentioned in Chapter 1, cooperation in the classic Pr isoners Dilemma scenario refers to not confessing to the crime and thus cooperating with the co-criminal.

PAGE 39

39 Following the briefing, participants indica ted their cooperation or defection by circling either cooperate or defect on a game sheet which was located on the first page of a questionnaire that was given to them during the briefing period (see appendix A). After indicating their cooper ation/defection game decision, participants then filled out the questionnaire by responding (in writing) to a se ries of open-ended quest ions about their game decision and providing demographic informa tion. Upon learning the cooperation/defection decision of both players the researcher compensated each player accordingly. The researcher then debriefed both players individually to prevent any possible animosity resulting from defection during game play. Measures and Instruments The data for this study came from two sour ces: (1) participants recorded decision to cooperate or defect and (2) a questionnaire that asked particip ants about their demographic characteristics and the following open ended questions10: Why did you cooperate or defect ? Please explain in detail. Do you feel you can trust the ot her player? Why or why not? Do you believe the other player will cooperate? Why or why not? The participants recorded decisions to coopera te or defect were ente red into a statistical analysis computer program, Sta tistical Package for the Social Sciences (SPSS), which was used to calculate whether differences between groups we re statistically significan t (greater than zero) and whether the studys hypotheses (presented in Chapter 4) were supported by the data. The researcher coded the open-ende d questionnaire responses accordi ng to any common themes that emerged. Of particular interest were the reasons and rationale put forward by participants for cooperation/defection, and how th ese differed between and within the four categorical gender and racial groups and the two cross-race groups.

PAGE 40

40 The strengths of this study are that: (1) it ex amines if there is a correlation between race and cooperation. This is major gap in the soci al dilemma literature and this study is a step toward filling it; (2) Unlike many studies inco rporating the Prisoners Dilemma, this study uses a one-shot version of the game that guarded agai nst revenge, reputation and changes in strategy affecting cooperation rates; (3 ) Incorporating an open-ended su rvey component allows for the examination of why participants cooperate or de fect in addition to the data on their actual cooperation and defection rates; and (4) Unlike most previous re search that has tended to exclusively measure the likelihood of cooperation, irrespec tive of the gender and/or racial group membership of each individual and pair of partic ipants, this study specifically examines intragroup and cross-group cooperation rates. Experimental Design According to Bernard (2000), the experiment al method is not one single technique but an approach to the development of knowledge (139). Experiments are often used by social science researchers to control th e effects of confounds to understa nd the effects of a particular independent variable. However, since this expe riment (like most) entailed investigating the effects of several independent vari ables at once, a factorial design that allows for the systematic analysis of each independent variable was most appropriate. Moreover, factorial design made it possible to measure interaction effects on depend ant variables (such as cooperation/defection) that occur as a result of intera ction between two or more independe nt variables (such as race and gender). The strength of using this method is that it enabled the researcher to see if there are any statistically different levels of cooperation/defection both among and between racial and gender group pairings. Additionally, a general benefit of f actorial experimental designs is that they can 10 The questionnaire can be viewed in Appendix A.

PAGE 41

41 be replicated, thus augmenting the confidence that can be placed in the fi ndings internal validity and generalizability (Babbie 2001). Perhaps the greatest strength of this method, however, is that a controlled experiment can measure social dilemma cooperation in an isolated environment and under such conditions that would be difficult, if not impossible, to observe otherwise. A general limitation of the experimental met hod used is its artificiality (Babbie 2001). In other words, what happens in a controlled setting may not be reflective of what happens in more natural social encounters. Any study concerni ng social dilemma cooperation/defection should take this into account. In the case of social dilemma cooperation/defection the possibility of spurious correlations were also of concern. Fa ctors other than race or gender, such as income, class, and/or religiosity may be at the root of greater group cooperation/de fection and these were statistically controlled for when appropriate. Recruitment of Participants Participants for this study were Black and White undergraduate students attending a major public southeastern university. They were recruited through the use of flyers that were posted on campus and distributed in general elective introducto ry level sociology classes. Recruitment flyers stated that participants co uld earn from $0-$9 dollars for playing a social decision making game that would take between 15 and 20 minutes to complete. The recruitment flyers also stated that th e study was seeking African-Amer ican and White non-Hispanic participants. Demographic Characteristics of the Sample Participation in the study was restricted to self-described Black and White students to examine if race (or perhaps racism) influenced a participants decision to cooperate or defect. Due to the systemic form of White-on-Black racism that has persisted in the U.S. from the 17th

PAGE 42

42 century until the present, any racial group diffe rences in social dilemma cooperation resulting from subordination are most likely to be evident in same-race and cross-race pairings of Blacks and Whites. Latinos were not recruited11 for the sample because th ey trace their origins to a multitude of countries including Mexico, Puerto Rico, Cuba, Dominican Republic, as well as all the Spanish speaking countries of Central and South America, each of which have their own differing systems of racial hier archy and domination. Moreover, racially Latinos are white, black, indigenous, and every possible combinati on thereof (Suarez-Orozco & Paez 2002:3). A general weakness of this study is that th e sample was limited to undergraduate college students. Participants for the study ranged from 18-28 years of age. The mean age was 20 with a standard deviation of 1.63. Also, the majority of participants report ed coming from a middleclass, upper-middle class, or upper-class b ackground with 80.5 percen t reporting a family income of $41,000 or higher. Fifty-five college majors were represented with sociology being the most common major (n = 25). In these character istics, the sample is reflective of the typical ages, social class positions, and va riety of educational majors of traditional college students and therefore is susceptible to any biases present in this population. For example, studies on aging employing the Prisoners Dilemma provide eviden ce that cooperation rates may be higher for children than for adults (Sjoberg, Bokander, De nick, & Lindbom 1969). So cial class may also influence cooperation and researchers have show n that college major can be correlated to increased or decreased cooperation rates (James Soroka, & Benjafield 2001; Hu & Liu 2003). By only studying college undergraduates, this st udy did not address the question of whether having a broader and more diverse sample popula tion would have yielde d different results. 11 A few participants who considered themselves White non -Hispanic, but who indicated they had a Latino ancestor on the questionnaire were permitted to participate. Si milarly, participants who considered themselves African American, but indicated that they had ancestors from th e Caribbean were also permitted to participate.

PAGE 43

43 More specific weaknesses of this studys sample include the large contingent of sociology majors ( n = 25) who participated, the use of sociology facilities to conduct experiments, and the participation of approximate ly 40 students who were enrolled in a course taught by this researcher. To combat possible influences on responses that a participants connection to the sociology department and/or to this researcher may have, it was explained to participants (and explicitly stated in the inform ed consent form) that their responses would be anonymous to all but the researcher and that th e responses would have no bearing on any course grade. Analyzing the Data Quantitative procedures for analyzing the da ta in this study included cross-tabulations, bivariate and multivariate analysis, analysis of variance and logistic regression. Qualitative procedures included a content an alysis of the open-ended res ponses to the question Why did you cooperate or defect? Please explain in de tail. Questionnaire responses were coded according to various thematic categories and comp ared within and between the six experiment race and gender group categories. In addition to being qualitatively coded, responses to the questions Do you feel you can trust the other play er? and Do you believe the other player will cooperate? were quantitatively coded and assigned into one of the three following categories: (1) mostly yes; (2) mostly no; and (3) uncertai n. Statistical procedures to analyze race and gender group differences to the responses to these questions were th en conducted using SPSS. Role of the Researcher The researcher/social dilemma game facilita tor attempted to remain as neutral to the process as possible. He employed the same verbal script to explain the ru les of the game to each participant and provided cl arification when needed. Furthermore, he made sure that participants

PAGE 44

44 were familiar with all of the possible outcome scenarios and the corresponding pay-out for each one. At the conclusion of the game he paid pa rticipants accordingly and debriefed each one individually. He also asked participants not to reveal the dynamics of th e game to any of their peers who might participate in the study. At the games conclusion he made sure participants left their separate rooms at staggered times so as not to encounter one another on the way out. Finally, he made every effort to maintain the same professional demeanor throughout each game and debriefing period. For the purpose of reflexivity, the researcher would like to disclose that he is a Latino, male, graduate student, 28-years of age and that his presence in itself may have influenced the outcome of the study in unforeseen and unknown ways.

PAGE 45

45 CHAPTER 4 QUANTITATIVE RESULTS Hypotheses In consideration of the first two research que stions, (1) Is there a correlation between race and likelihood of cooperation in a social dilemma? and (2) Is there a correlation between gender and likelihood of cooperation in a social dilemma? outlined in Chapter 1 (page 12), I posit the following five hypotheses which are theoretica lly grounded in the works of Gilligan (1982), Chong (1991), McAdams (1995), James, Soroka a nd Benjafield (2001), and Hu and Liu (2003) among others. Hypothesis 1: Black participants will have high er intra-group cooperation rates than White participants. Hypothesis 2: Women will have higher intra-group cooperation rates than men. Hypothesis 3: Black women will have the highest c ooperation rates of any of the four same-race and same-gender experiment groups. Hypothesis 4: White men will have the lowest cooperation rates of any of the four same race and same-gender experiment groups. Hypothesis 5: same-gender cross-racial group pairi ngs will have lower cooperation rates than the intra-group pairings of both Blacks and Whites. The following sections discuss the quantitative results of the study, which yield support for hypotheses numbers one and three. The findings from a quantitative analysis of trust and expectations are also discussed, as is the role of relig iosity in social dilemma decision making. Quantitative Findings Cooperation Tables 4-1 and 4-2 display the cooperation ra tes and distribution of outcomes for each of the six experimental groups. Hypothesis 1 pr edicts Blacks will have higher intra-group cooperation rates than Whites in the monetary Prisoners Dilemma. Concordant with this

PAGE 46

46 hypothesis, 69% of Blacks and 50% of Whites in same-race groups cooperated. This difference in same-racial group cooperation ap proached statistical significance ( p = .061) (Table 4-3). Additionally, if one also takes into account the participants in cross-racial group experiments, the overall cooperation rate is 72% for Blacks and 54% for Whites. This difference in Black and White cooperation is significan t (p = .025) (Table 4-4). Hypothesis 2 predicts that women will have higher intra-group cooperation rates than men will have in the monetary Prisoners Dilemma. Contrary to this hypothesis, 61% of women and 65% of men cooperated (Table 4-5). This difference was not significant ( p = .604) and indicates virtually no difference between female a nd male cooperation rates. Furthermore, there were also no statistically significant differen ces between females and males on the variables trust in partner and expectation of partners deci sion to cooperate or defect These findings are concurrent with the majority of previous stud ies that have found no signi ficant difference in the cooperation rates of women compared to thos e of men (see Caldwell 1976; Goering & Kahn 1976; Orbell, Dawes & Schwar tz-Shea 1994; Sell 1997). Hypothesis 3 predicts that Black women will have the highest cooperation rates of any of the four same-race and same-gender categoric al groups. Supporting this hypothesis, the cooperation rates for same-race and same-gender group pairings are as follow: Black women 75% cooperation; Black men 62% cooperation; White men 54% cooperation; White women 46% cooperation (Table 4-1). Though th e differences between the coopera tion levels of all the samerace and same-gender groups is not statistically significant ( p = .202), the cooperation level difference between the Black female group and White female group is significant ( p = .039). Hypothesis 4 suggests White men will have th e lowest cooperation rates of any of the four same-race and same-gender categorical gr oups. This hypothesis was not supported.

PAGE 47

47 Although the White male group had the lowest coope ration rate of any of the categorical groups besides White women, the differences between the White male groups cooperation rate and most of the other groups was not statistically significant. The final hypothesis predicts that same-g ender cross-racial group pairings will have lower cooperation rates than the same-gender same-race group pairings of both Blacks and Whites. Surprisingly, the exact c onverse proved true in that the male cross-racial group pairings had the highest cooperation rate (79%) of any group and the female cross-racial group pairings cooperation rate (63%) tied for th ird (out of the six groups) with the Black male group (Table 41). This finding contradicts the findings of ear lier research on how race affects cooperation in the Prisoners Dilemma (see Sibley, Senn & Epanchin 1968; Wilson & Kayatani 1968; Cederblom & Diers 1970; Baxter 1973). On th e other hand, this study corroborates the more recent cross-racial group coope ration findings of Hieder and Skowronski (2007). Additionally, this study demonstrated that male cross-racial group pairings had cooperation rates that approached significantly greater levels than White male group pairings ( p =.066) and that were statistically greater than the White female group pairings ( p = .017). These findings run counter to thos e of Dowing and Bothwell (1979) a nd suggest that either racial discrimination in social dilemmas is subsiding or th at the desire to appear non-racist is leading to increased cross-racial cooperation. Finally, it should be noted that overall Black and White differences in cooperation (that indicate that Blacks have highe r cooperation levels than White s irrespective of race of game partner) remain significant even in logistic regr ession models that control for the influence of gender and family income ( p = .044).

PAGE 48

48 Quantitative Analysis of Trust and Expectations The quantitative component of this study al so revealed strong re lationships between whether a participant trusted their game partner, their expectation that th eir game partner would cooperate, and their decision to cooperate or defect (Table 46). Of the participants who indicated in the questionnaire that they mostly trusted thei r game partner, 86% cooperated whereas 37% of those who mostly did not trust their partner coopera ted. Of the participants who indicated they where uncertain as to whether or not they trusted their partner, 64% cooperated. In this, cooperation levels for part icipants who identified themselves as uncertain more closely resembled that of participants who trusted their pa rtner than those who did not trust their partner. The relationship between trust of game part ner and cooperation is significant at the p < .001 level. Expectation of game partners decision wa s also a significant pr edictor of cooperation (Table 4-7). Participants who mostly belie ved their game partner would cooperate had a cooperation rate of 82%. Conversely, particip ants who mostly believed their game partner would defect had a cooperation rate of 25%. Of the participants w ho were mostly uncertain as to whether their partner would cooperate or def ect, 68% choose to coope rate. The relationship between expectation of part ners decision and cooperat ion is significant at the p < .001 level. Additionally, there is a significant ( p < .001) relationship between wh ether a participant trusts his/her game partner and his/ her expectation that the ga me partner will cooperate. Also of note, when excluding participants who indicated they were uncertain as to whether their game partner would cooperate or defect, statistically significant ( p = .051) differences between Black and White participan ts became visible (Table 4-8). Seventy-two percent of Blacks believed their ga me partner would cooperate comp ared to only 56% of Whites.

PAGE 49

49 A breakdown of participants exp ectation of game partners d ecision by race and gender is shown on Table 4-9. The Role of Religiosity Of the variables measured, the only one othe r than racial variables to correlate with cooperation was religiosity. The re ligiosity variable was originally coded with three categorical options for the response to the question How relig ious are you? These options were (1) below average: (2) average; and (3) above average. La ter this data was collaps ed into the categories less than average religiosity, and average or higher than average religi osity in order to better explore how the presence and absence of relig iosity is correlated with social dilemma cooperation and defection. In th is, participants who identified as having less than average religiosity had a cooperation rate of 50%, whereas those who iden tified as having average or higher than average religiosity had a cooperation rate of 69% ( p = .029) (Table 4-10). Religiosity was also associated with trus t in partner. Alt hough participants who identified as having less than aver age religiosity were just as likel y (39%) to trust their partner as participants who identified as ha ving average or above average religiosity, the former group was more likely (45% versus 30%) to indicate that th ey did not trust their partner. Furthermore, participants with below average religiosity were half as likely to indicate uncertainty as to whether they trusted their partner (16%) than wh ere participants who identified as average or above average religiosity (32%). These measur es approached statisti cal significance at the p = .069 level. While religiosity and race are clearly associ ated to cooperation levels in the monetary Prisoners Dilemma, the quantitative results of this study are complicated by an intersection of race and religiosity (see tables 4-11 and 4-12). Eighty-nine pe rcent of Black participants

PAGE 50

50 indicated having average or above average religiosity compared to 50% of White participants. This difference in religiosity was significant at the p < .001 level. The higher level of Black participant religiosity was also evident in the six experiment r acial/gender groups with the Black female group and Black male group tied for the great est level (88%) of group participants in the average and above average religiosity category. Differences in the six racial/gender groups' religiosity levels are significant at the p < .002 level. In logistic regressions that incorporate religiosity and race, race as a predictor of cooperation is no longer a significant measure ( p = .136). Therefore, it is possibl e that cooperation level differe nces by race are a reflection differences in religiosity and vice-versa; however further exploration of this intersection is needed in order to draw a more definitive conclusi on. It should also be noted that reported levels of religiosity correlated with cooperation rates in a similar manner for both Black and White participants. In terms of the religion of participants, 76% of Blacks identified as either Protestant, Christian or both compared to 38% of Whites. Differences in religious affiliation between Blacks and Whites was significant at the p < .001 level. While religious affiliation was not significantly associated with c ooperation rates or expectation of partners decision rates, participants who indicated they were either Protestant or Christian were more likely to trust their partner (47%) than were non-Prot estant and non-Christians (29%). This difference approached statistical significance at the p = .081 level. Also of note, in logistic regressions that controlled for gender and family income of participants, reli giosity as a variable re mained significant at the p = .042 level.

PAGE 51

51 Discussion of Quantitative Findings The quantitative findings of this study provide support for the hypothesis that Black participants will have higher same-race group coope ration rates than White participants in a social dilemma. They also show that in this particular social dilemma, Black participants are generally more likely to cooperate with a game pa rtner (irrespectiv e of that game partners race) than are White participants. In Chapter 2, I prop osed that hypothesized racial differences in cooperation rates would be based on Black subjugation and responses to White racism, as well as greater Black racial identity awar eness. However, this studys quantitative component measures only if differences in cooperation exist, not what the reasons for these differences are. With that said, the quantitative findings are in line with the presented theoretical arguments for greater Black cooperation, although further analysis is n eeded (and will be presented in Chapter 5) to verify the validity of these claims. It should also be noted, that racial differences in participants levels of religiosity may be the underlying cause of greater Black coope ration, thus presenting the possibility that associations between race group and cooperation levels are the result of a spurious correlation. On the question of whether women have higher same-gender group cooperation rates than men (presented as hypothesis 2), the quant itative findings indicate no significant evidence of gender difference and thus are concordant with the majority of previous Prisoners Dilemma studies (see Caldwell 1976; Go ering & Kahn 1976; Orbell, Dawe s & Schwartz-Shea 1994; Sell 1997). This finding contradicts the view that gender socialization and the status inequalities that gender divisions create (in the Un ited States) will by themselves lead to greater female social dilemma cooperation levels.

PAGE 52

52 However, the finding (supporting hypothesis 3) that the Black female group had the highest cooperation levels of any same-race experiment group while the White female group had the lowest, suggests that the in tersection of race and gender cr eates unique perceptions and responses for different race and gender groupings. These differen ces in cooperation levels of Black and White females may be a reflection of the different social positions and power these women hold. As Lengermann and Wallace (1985) point out, White women are less likely to see themselves as an oppressed minority group than are Black women because the former are well represented on every level of U.S. social class structure, whereas the latter are found primarily on the lower and middle levels. With regard to so cial dilemma cooperation, this suggests that if White women do not see themselves as a minority group it is not likely they will have high cooperation rates that are due to their shared experiences of minor ity group subjugation. This is not to imply that gender group membership does play an important role in the cooperation/defection decisions of some White women, but rather that the influence of gender status on social dilemma coopera tion may be different for White women than it is for Black women. The negation of hypotheses 4 and 5 further speak s to the complicated nature of the results that the intersection of race and gender can crea te. In particular, the findings of higher than expected levels of both male and female cro ss-racial group cooperation can be viewed in many ways. These quantitative findings can lend support for Bells (1980) Interest Convergence Hypothesis, Wilsons (1978) declining significance of race argument, and a host of other theoretical perspectives. In Ch apter 5, I report and discuss the ope n-ended data findings of this study, which help to assess and in terpret the reasons behind the re latively high levels of crossracial group cooperation.

PAGE 53

53 In light of the above findings it is important to note that one of the biggest challenges faced by gaming studies, and more particularly th e Prisoners Dilemma, is questions about the external validity of laboratory experiments and their relevance to real-life social dilemmas (Nemeth 1972; Apfelbaum 1974; Pruitt & Kimmel 1977). As Pr uitt and Kimmel (1997) note, most researchers who use games simply report th eir results with no atte mpt to speculate about the real-life implica tions (367). This is in part due to th e view of certain researchers such as Rapoport (1964) who feel labo ratory results are chiefly valuable for the perspectives they bring forth and questions they give rise to. Thus, they feel no need to generalize beyond the laboratory. Others, such as Pruitt and Kimmel (1997), feel it is preferable for researchers to try to generalize their findings to settings with si milar background conditions, however, side against those, such as Rubin and Brown (1975), who atte mpt to generalize their game results to all settings involving interpersonal dependence. The current study utilizes race and gender theo ry as the basis for testable hypotheses and employs content analysis in an effort to show the relevance of the findings to social dilemma research without over generalizing. Hence, what can be said a bout the quantitative findings presented above is that minority racial group membership and reli giosity are posi tively correlated with cooperation in the social dilemma game Pr isoners Dilemma. Additionally, it is reasonable to suggest that these variables may also be co rrelated in settings involving interpersonal dependence with similar background conditions. Ho wever, as some of the open-ended responses that will be discussed in the next section show, the ideas of what qualifies as a similar condition can vary greatly from person to person. In the following chapter, a content analysis of questionnaire responses will help address some of the validity concerns about employing games to study social dilemma cooperation.

PAGE 54

54 Table 4-1. Proportion of cooperation and de fection by race/gende r experiment group Participants race/gender experiment group % Cooperated % Defected White male group(n = 24) 54.2 45.8 White female group (n = 24) 45.8 54.2 Black male group (n = 24) 62.5 37.5 Black female group (n = 24) 75.0 25.0 Cross-race male group (n = 24) 79.2 20.8 Cross-race female group (n = 24) 62.5 37.5 Total 63.2 36.8 Table 4-2. Distribution of game pair outcomes by race and gender experiment group Experiment Group Cooperate/Cooperate Cooperate/Defect Defect/Defect White Males 3 pairs 7 pairs 2 pairs White Females 1 pair 9 pairs 2 pairs Black Males 5 pairs 5 pairs 2 pairs Black Females 7 pairs 4 pairs 1 pair Cross-race Males 8 pairs 3 pairs* 1 pair Cross-race Females 4 pairs 7 pairs* 1 pair *In cooperate/defect category of the cross-race male group a ll defectors were White partic ipants. In cooperate/defect category of the cross-race female grou p 4 defectors were White participan ts and 3 were Black participants. Table 4-3. Proportion of cooperation and defection in exclusively White and exclusively Black experiment groups Experiment group % Cooperated % Defected Exclusively White groups 50.0 50.0 Exclusively Black groups 68.8 31.3 Chi-square = 3.498 p = .061 n = 96 Table 4-4. Proportion of cooperation and defection by race Race of participants % Cooperated % Defected Whites 54.2 45.8 Blacks 72.2 27.8 Chi-square = 5.046 p = .025 n = 144

PAGE 55

55 Table 4-5. Proportion of coopera tion and defection by gender Gender of participants % Cooperated % Defected Women 61.1 38.9 Men 65.3 34.7 Chi-square = .269 p = .604 n = 144 Table 4-6. Proportion of cooperation and defection by trust in game partner Participant trusts game partner % Cooperated % Defected Mostly yes 85.7 14.3 Mostly no 36.7 63.3 Uncertain 64.1 35.9 Chi-square = 26.974 p < 0.001 n = 144 Table 4-7. Proportion of cooperation and defecti on by expectation of game partners decision Expects game partner to cooperate % Cooperated % Defected Mostly yes 82.3 17.7 Mostly no 25.0 75.0 Uncertain 68.4 31.6 Chi-square = 39.944 p <0.001 n = 142 Table 4-8. Expectation of ga me partners decision by race Expects game partner to cooperate % White % Black Mostly yes 55.7 72.6 Mostly no 44.3 27.4 Chi-square = 3.796 p = 0.051 n = 142 Table 4-9. Expectation of game pa rtners decision by race and gender Expects game partner to cooperate % White % Black Mostly Yes Women 56.7 71.0 Men 54.8 74.2 Mostly No Women 43.3 29.0 Men 45.2 25.8 Chi-square = 3.889 p = 0.274 n = 142 Table 4-10. Proportion of cooperatio n and defection by religiosity Level of religiosity % Cooperated % Defected Average and above 69.0 31.0 Below average 50.0 50.0 Chi-square = 4.742 p = 0.029 n = 144

PAGE 56

56 Table 4-11. Level of religiosity by race Level of religiosity % White % Black Average and above 50.0 88.9 Below average 50.0 11.1 Chi-square = 25.658 p < 0.001 n = 144 Table 4-12. Level of religio sity by race and gender Level of religiosity % White % Black Average and above Women 52.8 91.7 Men 47.2 86.1 Below average Women 47.2 8.3 Men 52.8 13.9 Chi-square = 26.182 p < 0.001 n = 144

PAGE 57

57 CHAPTER 5 O PEN-ENDED DATA RESULT S Open-ended Data Findings As stated earlier, the aim of the open-ende d questionnaire component of this study is to explore the reasons and rationale pa rticipants give for their decisi ons to cooperate or defect and to see if there are general differences in the wa y different gender and/or racial groups explain their choices. It is also to see if the common explanations give n by participants differ when their game partner is of a different racial group. To examine the wr itten responses to the question, Why did you cooperate or defect? Please explai n in detail, a coding system was developed with 27 categories that captured the most typical responses (see ta ble 5-1). These responses were then compared within and betw een the six racial and gender e xperimental groups. Responses that were atypical were also noted and compared in the same fashion. Responses to questions on the variables of trust and expectations though mostly utilized for quantitative analysis were also examined and are discussed here. It should be noted that women generally wrote longer answers than did men and perhaps concomitantly gave mo re reasons for their de cisions. Unexpectedly, open-ended responses also revealed information about the validity of the game and how certain participants equated this labor atory social dilemma to other social dilemmas they might encounter in natural settings. In this chapter, I will first discu ss the findings on validity and then present a summary of the general reasons each of the six experimental racial and gender groups gave for both cooperation and defection. Next, I will discuss experiment group differences to questions three and four of the questionnaire. I will conclude the chapte r with a discussion of what these findings mean for social dilemma research.

PAGE 58

58 External Validity Although none of the questions on the questionnaire were specifi cally designed to test the validity of the study, in their written responses to the open-ended questio ns some participants made direct comparisons between the social dilemma game they were playing and other social dilemmas that they have encountered in the past or may encounter in the future. One participant equated trusting someone to coopera te in the game to asking a stranger or associate to save my place in line please (Participant 77). Others co mpared game cooperation to trusting someone to return a pencil that they had borrowed or not tattling on someone (Participant 37 and Participant 5). An interesting analogy for coope ration in the game came from a participant who wrote: In life sometimes you will be confronted with life making decisions; therefore, you have to be able to cooperate with your peers even if you dont know their name. For example, you might be stuck in an elevator with a pregnant women. What do you do? There is another person inside, do you and him cooperate or do you argue? So, you see it is very important to cooperate (Participant 114). These types of responses were equally presen t in all of the racial/gender groups and serve to demonstrate that at least some participants ( n = 6) equate the social dilemma present in the game Prisoners Dilemma with social dilemmas th at are encountered in natural settings. Additionally, five participants referenced metaphors popular in other settings to explain their decisions in the laboratory. One participant compared his de cision to cooperate as tak[ing] one for the team (Participant 61). Another said although society is all about su rvival of the fittest, I do believe that we should work together to better survive [rather] than try to kill each other (Participant 135). Tw o participants alluded to the win-win philosophy popular in the corporate sector, one writing that In the long run, it is better for us to cooperate and not be

PAGE 59

59 greedy. Same idea as organizations or businesses. If they all helped eac h other out instead of working against one another, everyone would win (Participant 10). Support for the external validity of the study is also evidence d in responses such as I equate this game to life and I felt to cooperate is the risk I take everyday, so why not transfer that same value to a game (Participant 131). Sim ilar responses like In my opinion in a life perspective cooperating is the only way to succeed demonstr ate the parallels that some participants draw between their decision in the game and decisions they make in their daily life (Participant 33). On the other ha nd, a handful of participants make reference to the artificiality of the experiment with comments such as, I kno w Ill feel bad for her if she gets no money, but I kind of just convinced myself that its a game and the whole point of playing a game is to have a winner. This now as Im reading it sounds really terrible though because we should have fun together not compete (Participant 44). Overal l, however, most references to the external validity of the study suppor t the assertion that part icipants draw on their r eal world experience in their game decision making process. Furthermore, some participants equate their game decision to decisions they would make in other social d ilemmas. What these other social dilemmas are, however, can vary greatly betw een participants. An issue that negatively affects the extern al validity of the st udy is the amount of monetary pay-out given to participants depe nding on their game decision and the decision of their game partner. As noted by Simpson (2003), changes in the monetary reward and punishment pay-out matrix produce different types of social dile mmas with possibly different results. With regard to this, one to two coopera tive participants in ever y experiment group except the Black female group, expressed th at had the pay-out been larger they would have defected.

PAGE 60

60 Therefore any generalization of th ese findings to other contexts should consider the relatively minor stakes the social dilemma in this study presented participants with. In the following sections, I summarize, comp are, and contrast the general reasons each of the six experiment race and gender groups ga ve for both cooperation and defection (see Table 5-2). Reasons for Cooperation and Defection White male experiment group Among participants in this group there were fi ve frequently cited reasons for cooperation. As with all of the six expe rimental racial/gender groups, the most popular reason given by participants in the White male group for cooperati ng is that cooperation offers the most mutually beneficial outcome for themselves and their ga me partner. As one participant wrote I cooperated because thats the best outcome for the both of us. Its not the most important to care about my well-being but I would like to see ot hers benefit as well (Participant 20). The next most common explanation fo r cooperation amongst the White male group centered on the theme that they c ooperated because of their personal faith in humanity. Of the 13 participants in this group who cooperated, five expressed sentiments such as I chose to cooperate because I believe that humans are i nherently good and unfortunately we have been given all this incorrect, negative propaganda that it is us against the world. But in reality it is not; we are all in this together and the more we cooperate (Pun intended!) the better off we are all going to be (Partic ipant 2). And, I chose to cooperate because I tend to give people the benefit of the doubt and [believe ] that people are good-hearted (P articipant 13). It should be noted that faith in humanity as a reason for coopera tion was either not cited at all or not cited by more than one participant in any of the other five experimental groups.

PAGE 61

61 After faith in humanity, the next major r eason for cooperation stated amongst the White male group was that his game partner seemed nice/friendly. Though I never knew that student [before today], I get the impression that hes a nice, genuine person from the first impression (Participant 10). Another writes, I cooperated because the other participant appeared that he would be more willing to cooperate. By that I m ean when I saw him he di d not look like the type of person that would make a decision to go ag ainst a complete stranger. He looked like a genuinely good-mannered person. Another reason given by multiple par ticipants to explain their cooperation was simply that cooperation is the better choice (P articipant 13). To participants employing this reasoning it seems that cooperation is the most rational alternative. The primary explanation for defection gi ven by the White male experimental group participants was that defection guarantees money. Ten out of the eleven participants in this group who defected attributed their choice, at least in part, to the fact th at irrespective of their game partners decision, they would receive either $3 or $9. Statements such as I choose to defect because there is a guara ntee I will receive money were frequently observed in this experimental group (Participant 1). A second, less major reason for defection cited by members of this experimental group is that defection made the maximum individual pay-out possible. I defected because it offers me the chance to win the most [money] possible (Participant 4). Finally, the sentiment that a persons game partner will likely think and/or ac t like me was equally given as a rationale for both defection and cooperation decisions. White female experiment group Participants in the White female experime nt group primarily cited cooperation is the most mutually beneficial outcome as the reason fo r their cooperation. Slightly more than half

PAGE 62

62 of the women who cooperated expres sed this logic in their written response to the question of why they cooperated of defected. Another popular response for this group was, I did not want to sucker my game partner. One participant w ho cooperated noted that by choosing to defect I feel as if I am being unjust to th e other participant (Participant 31). A participant with a similar view wrote, if I chose to defect and she chose to cooperate, I would feel guilty about the money. Choosing to cooperate means th at I am still electing to win money, but not hurting the other person with my motives (Participant 27). Un like the White male group, faith in humanity was not a popular response in the White fe male group and was mentioned by only one person. Interestingly, three respondents in this group wrote that they cooperated because they did not need the money. Even if she chose de fect, losing the money woul d not be an issue for me (Participant 47). This sentiment was also expressed multiple times in the Black female experiment group and in the female cross-racial group; however, only one male (White) in the entire study mentioned not caring about the money (Participant 105). One final cooperation theme that became evident in the White female experiment group is that the female gender of the game partner pl ayed a role in some participants decision to cooperate. As one participant put it, There wa s the possibility that I would cooperate and she would defect to receive the full money, but I felt that isnt something a girl would do, more of a sabotage act a male would commit (sexist, I agree) (Participant 25). This view was shared by another participant who wrote, M y first instinct is to cooperat e to be nice but I would figure most people would defect because they will make m oney and they cant lose, so if she chooses as I think she will we both will win. Its just an in itial judgment to go by. If it were a male I would have no question but to defect, because shes female she might cooperate (Participant 38). A third expressed I think th e other participant, as a girl, is mo re likely to cooperate than defect

PAGE 63

63 (Participant 41). Notably, White women were the most likely to list gender as the reason or partial reason for their decision. None of the men and only two Black women in the study mentioned gender in describing their decision-ma king rationale. Despite the view by some White women in this group, as well as in the cr oss-race group, that females are more likely to cooperate as the quantitative pa rt of this study demonstrate d, White women had the overall lowest cooperation rates of any of the six racial/gender expe rimental groups. Like the White male group, the large major ity of defectors in the White female group cited the guarantee of money as the principal reason for thei r defection decision. One White women stated I chose to defect because I figured I cannot leave w ith $0 (Participa nt 30). Also, similar to the White male group, the White fe male groups second most commonly given reason for defecting is that defecting makes the maximum pay-out possible. Due in part to their higher defection rates, however, White women listed th is reason twice as ofte n as White men did. Three women also mentioned that they defected (among other reasons) because they did not know their game partner. I did not know the stranger, wrote one participant of her decision to defect (Participant 28). This reason was given either one or two times in each of the other five racial/gender experiment groups. Finally, the be lief that a persons game partner will likely think and/or act like me, was listed slightly more often as a rationale for cooperation than for defection. Black male experiment group Participants in the Black male group (with a few notable exceptions) generally gave the briefest written answers to th e questionnaire of all the six ra cial/gender experiment groups. Concomitantly, they provided fewe r types of reasons for their coope ration or defection. Like the previous groups discussed, the most common reas on for cooperation given by participants in the

PAGE 64

64 Black male group centered on the theme that c ooperation is the most mutually beneficial outcome. Ten out of the 15 players in this gro up who cooperated gave this as their entire or partial response. The only other rationale for cooperation given multiple times by participants in this group is variations on the sentiment, I cooperated because I am a good/nice/cooperative person. Statements such as that is just the ty pe of person I am, if we can both get paid why not cooperate, and, seriously, I am a nice person and I do not see any reason to defect were used to communicate that participants viewed coopera tion as a positive character trait (Participants 53 and 61). One participant in this group explicitly wrote, I am a cooperati ve person (Participant 63). Curiously, this response theme, although fairly common among Blacks in the study, was solely mentioned by one White male and one White female. On the other hand, the popular White male group cooperation response, faith in humanity, was only mentioned by a single Black (male) participant in the entire study. Unlike the majority of defectors in the Wh ite male and White female groups who listed guaranteed money as the most common reason fo r their defection, the most common response participants in the Black male group gave for defection is that defection made the maximum game pay-out possible. Six of the nine defect ors in this group mentioned the possibility of achieving the maximum pay-out as a motivating factor. One participan t wrote, I chose to defect because I am a very competitive person. This is a game and defecting gives me the chance to win the most (Participant 54). Another stated, [ I] defected in hopes that my partner will pick cooperate and I will get nine dollars. Sounds se lfish but oh well, thus again highlighting how this social dilemma decision can be viewed by some as a reflection of individual character (Participant, 59). The second mo st typical rationale for defec tion among participants in the

PAGE 65

65 Black male group was that defection guaranteed mone y. Four respondents gave this as part or all of their answer. Members of this experi ment group gave few other reasons for defection. Black female experiment group Participants in the Black female group ma de frequent reference to the mutually beneficial outcome of cooperati on and like all other racial/gende r groups listed it more often than they did any other reason for cooperation. Other, moderately common responses to the question Why did you cooperate or defect? given by participan ts in the Black female group are: (1) I cooperated because I am good/nice/cooperative person; a nd (2) I think my game partner will likely think and/or act like me. In an illustration of the latter reason a respondent wrote [I cooperated] because I like to think positiv ely and I believe that she will do the same (Participant 95). The other major theme to emerge from this groups responses (more clearly than from any others) was that of a moral/a ltruistic imperative to cooperate Examples of altruistic and morally laced statements made by participants in this group in clude responses such as [I cooperated] because I was thinking about helpin g the other person out and we could both leave with money (Participant 74) a nd I cooperated because I like to share and am not a very selfish person so, it was OK to cooperate; the more the be tter. Why would I be greedy and not share the wealth (Participant 75). Another respondent stated, I coope rated because, why would I want to work against someone for seemingly no reason, it seems a bit wrong and selfish Id feel a tiny bit bad if I somehow robbed her of an easy, free $6 (Participant 77). Related to the moral/altruistic theme was that of I cooperated because I do not need the money. A handful of Black female respondents e xpressed statements such as if she gains all the money and I gain nothing, its OK because I w ould feel that she needs the money. Whereas I

PAGE 66

66 really am not in need (Participant 87). Additionally, assertions such as the other person seems caring enough to want to help someone with a surv ey, so she seems trustwor thy; also, I figure if she decides to defect Im not at a great loss are also examples of this line of reasoning (Participant 92). An additional type of response that loosely corresponds to the moral/altruistic theme is that of choosing cooperation to avoi d feelings of guilt which was cited by three participants in the group. Though this group had relatively few defecti ons, the two most common reasons given for defection by participan ts in this group were that def ection guarantees money, and that, defection made the maximum pay-out possible. Participants in this group gave no other highfrequency responses for why they chose defection. Male cross-race experiment group The male cross-race group had the highest c ooperation rate of any of the six experiment groups. The most common reason given among both Black and White respondents was that cooperation provided the most mutually benefici al outcome. The next most common reasons for cooperation given by Black respondents cen tered on the I coopera ted because I am a good/nice/cooperative person theme and the moral/ altruistic theme. Both of these response types typically expressed that th e participant felt a strong connec tion between their character and values and their decision to cooperate. On e Black participant in this group wrote: I chose to cooperate because the money being issued collectively is maximized if both parties decide to cooperate ($12 in total), along with the idea that I dont want to seem like a selfish jerk if I happened to choose defect and he chooses cooperate. Sure, Id receive $9, but the f act that he received no money because of my selfish action I believe significantly outweighs the monetary value of $9. Of course, he could choose to defect, leaving me with no money; that would simply convey the type of character he has. In this case, it would bother me much more that he received no money because of my selfishness, than vice-versa. Money isnt something that could deter my fee lings and values (Participant 117).

PAGE 67

67 Another Black participant in this group wrote, maybe I chose not to defect because of my parents and upbringing (Participant 112). White males in this group who cooperated mostly wrote about cooperation as mutually beneficial, but were le ss likely to invoke moral or character arguments in their response. One White participant wrote, I chose to cooperate because the net benefit to both people cooperating ($12) is greater than the net be nefit of either both defecting ($6) or one defecting and one cooperating ($9) (Participant 101). Another wrote, I c ooperated because I felt making $6 would be a decent gain for me and the other, an d another $3 to defect isnt worth losing more [total] dollars. However if the game was fo r millions, I would probably defect (Participant 120). Interestingly, no other cooperation themes am ong White male participants in the crossrace group emerged. The theme faith in human ity that was the second most popular response among the exclusively White male group was not me ntioned by a single White male in the crossrace group. Similarly, the response, I c ooperated because my game partner looks nice/friendly, which was the third most popular response in the White male group, was not mentioned by a single White male in the crossrace group (although one Black respondent listed it). Notably the cross-race group had only one Bl ack participant who defected (compared to four White participants who chos e defection). Also notable, is the rationale given by the Black participant for his defection. He wrote, I really wanted to cooperate but I dont know the other person Im playing with my theory: with me being who I am, if he chooses to cooperate, then I would have much respect for him and I would give him half of the money I received when I see him again. However, I couldnt put cooperate be cause Im pretty sure he wouldnt do the same

PAGE 68

68 for me (Participant 116). This participant was one of only two players in the entire study to suggest he would split the maximum defecti on money (though this task would have been difficult because participants departure time wa s intentionally staggered and they did not know one and other). The most frequent response White particip ants in this group who chose defection gave was based on the view that defection was the most economically rational choice. The other most common responses for defection mentioned by Whites in this group were that defection guaranteed money, and that defection was the safest option. One participant employing all of these views wrote, I think choosing coopera te is a risky and unwise choice since you are not guaranteed money and even if you do win, you cannot win the highest amount possible (Participant 110). Notably, every White who defected in the male cross-race group cited economic logic as the primary reason for their defection. Female cross-race experiment group The female cross-race group had the most varied responses for why they cooperated than any of the six racial/gender e xperiment groups. Still, the most common reason for cooperation given by both Black and White women in this group was that cooperation offered the must mutually beneficial outcome. The second most common response among Black women was that their game partner will likely think and/or act like me. It should be noted that Black women in this group were twice as likely as White women in this group to employ this response. The second most common types of responses fo r White women was I cooperated because I have nothing to lose, and, defecting is greedy, which were each mentioned three times. Employing the former type of reasoning, one Wh ite women wrote, If, by chance the other

PAGE 69

69 player decides not to cooperate, I am no better or worse off than when I entered this room, having received $0 (Participant 127). The most common response for defection gi ven by both Black and White women in this experiment group was that defection guarant eed money. A couple of White women also utilized the defection is the safer option them e in describing why they defected. One White women wrote of her defection, I didnt want to end up with $0 and I also prevented myself from being left in a vulnerable position (Parti cipant 137). This emphasis on safety was a theme that was generally more common among women th an among men and which I further discuss in the section titled, Gender and Race Influences. In the next section, I explore and discuss the responses to the questions deali ng with trust and expectations. Open-ended Data Analysis of Trust and Expectations Though responses to the questi ons Do you feel you can trust the other player? Why or why not? and, Do you believe the other play er will cooperate? Why or why not? were primarily incorporated into the questionnaire fo r quantitative coding and analysis (see Chapter 4), a content analysis of respons es to both these questi ons uncovered a difference in experiment group responses for the White male and Black ma le experiment groups. Before discussion of this difference, however, it should be stated that responses to these questions were generally very short (one to three sentences) and mostly simila r across all groups. Typi cal responses read, no, because she probably wants to receive the larger amount also, or, I have faith that he would pick cooperate (Participants 31 and 63). Mo reover, quantitative anal yses found no significant race or gender differences on the trust variab le and no gender differences on the variable measuring expectation of game partners de cision. However, some overall race group differences were found on the latte r variable (see Chapter 4).

PAGE 70

70 Despite the limited racial and gender group qua ntitative variation in the overall coding of responses, content analysis of experiment group responses found a major difference in the language used by different experiment groups to explain the decision to cooperate. This difference centered on the use of the physical ap pearance of a game partner as a factor in determining that persons trustw orthiness and predicting their ga me decision to cooperate or defect. Labeled as the halo effect, this phe nomenon causes others to assume that those who are physically attractive are also more generous, trustworthy, and sociable than those who are not attractive (Katz, 2003). Additiona lly, several studies have also established a positive correlation between a persons level of physi cal attractiveness and their chan ces of being the recipient of helping behavior (see Benson, Karabenick & Le rner 1976; Mims, Hartnett & Nay 1975; West & Brown 1975). With regard to the Prisoners Dile mma, Morse et al. (1974) showed that physical attractiveness plays an important role in the cooperation/def ection decision. In their study, females playing the Prisoners Dilemma made more cooperative responses when playing with an attractive male partner than with an unattractive male partner. They also found that males work harder for an attractive female experimenter th an for an unattractive one and that males allow attractive females to exert more social influence over them than they do an unattractive female. Even among players of the same gender, pe rceived physical attract iveness could possibly influence cooperation and defection rates. In this study, over one-third of participan ts in the White male group cited the physical appearance of their game partner as a factor in their c ooperation. One respondent wrote, I feel like I can trust him because of his disposition an d clothing style. I believe judging a book by its cover is completely necessary, often useful, and yes sometimes results in false assumptions, however with his clean-cut looks I dont think hes got too ma ny radical tricks up his sleeve

PAGE 71

71 (Participant 21). ) Other partic ipants wrote, he l ooked like a trustworthy person. He appeared well maintained and intellectual. He looked li ke someone that would trust me as well, and, simply put he looks like a nice dude (Participants 13 and 19). Interestingl y, all participants in the White male group who referenced physical app earance as a factor in their decision to cooperate used the term looks to describe th e assumed attributes of their game partner, whereas most women in the study employed the term seemed to de scribe assumed attributes. The Black male group, in contrast, had onl y one participant who employed the physical appearance of his game partner to explain his cooperation. This pa rticipant wrote, he looks like he needs money just like I know I need money, so I feel we both will make the same decision and cooperate so we can have a few extra dollars in our pockets (Participant 65). Black and White women, irrespective of their experiment group, were about equally likely to cite the physical appearance of their game partner in their responses to questions on trust and expectations, although White women showed a clear preference for th e word seems in describing their game partner. Statements such as she seems nice were very common in the White female group. An interesting finding is that no White males in the cross-race group (in contrast to onethird of the males in the exclus ively White male group) referred to the physical appearance of their game partner in their respons es to the questions regarding tr ust and expectations. Instead, a common response for White males in this group was th at they could not say if they trusted their game partner or if he would cooperate because they did not know him. One White male respondent wrote, I dont know him well enough to make that judgment (Participant 105). Although this was a common style of response fo r many participants irrespective of their experiment group, unlike White males in the cro ss-race group, one third of the Black males in

PAGE 72

72 the cross-race group cited the appearance of their White game partner as a factor in why they trusted him or expected him to cooperate. One participant wrote, I think he will cooperate because of the way he looks. His demeanor seem s to be laid back (Participant 107). Another stated, He seemed like a cool person (Partici pant 97). The absence of references to the physical appearance of Black game partners by White males and greater mention of perceived visible attributes of White game partners by Black males in the Cross-race group poses some interesting questions about who is allowed to talk about physical appearance and in what context. Though the use of physical appearance as a proxy to measure assumed social and personality characteristics of strangers is well documented (see Zebrowitz, Waltham & Montepare 2005; Katz 2003; Mulford, Orbell, Shatto & Stockard 1998; Kahn, Hottes & Davis 1971), whether physical appearance matters more for certain racial and gender groups is a question that needs further exploration. This st udys open-ended data findings show that many participants rely on stereotypes in their judgm ents of the trustworth iness and cooperative potential of others. Yet while many participants freely admit to the use of certain stereotypes (such as stereotypes about socioeconomic status, educational level, or social group membership), the mentioning of race and/or raci al stereotypes is clearly taboo. For example, differences in response styles according to the race of a partic ipant and the race of their game partner suggest that judgments based on physical appearance are permissible so l ong as you are a racial-minority group member or if your game partner is of your same racial group. Thus, a White male can list the look of a White game partner as his reason for trusting him or expecting him to cooperate, but is socially prohibited (lest he be considered racist) from listing the look of a Black game partner as a reason for his coope ration/defection decision. Additio nally, it is more common and

PAGE 73

73 appears to be more acceptable to list a game partners physical appearance as a reason for trusting or cooperating with them ra ther than as a reason for not tr usting them or defecting. The fact that the male cross-race experime nt group had the highest cooperation rates of any in the study may mean that participants in this group are taking racial considerations into account in their decision making a nd are cooperating at a higher rate because of it. In the next section, I analyze the references made to gende r and race by participants and explore the ways these demographic factors may influen ce cooperation/defection decisions. Gender and Race Differences Only one male participant in the entire study cited gender in his writ ten responses to the open-ended questionnaire. He was in the Black male experiment group and wrote, I generally would trust someone that never gave me a reas on not to trust them, but there can also be underlying trust because the other player was the same gender and race as me (Participant 51). Significantly, no other males in the study cited ge nder in their written responses. As stated earlier in the chapter, White women (in both the White women and cross-race experiment groups) were most likely to cite gender as fact or in their cooperation/defection decision. One White female participant expressed, I believe sh e is more likely to cooperate than to defect because she is a woman. Women are more prone to cooperation, and would be more likely to try to guarantee the maximum total payout between tw o players. I think she will comply with this stereotype (Participant 126). Anot her stated, I feel that I can to some degree, trust her. I suppose its because she is female as well as soft-spoken. She does not appear as a threat (Participant 31). A theme of being more willing to trust anot her woman rather than a man also emerged among both White and Black female participants. Explaining why she trusts her female game

PAGE 74

74 partner, a White women wrote, [her] being a fema le instead of a male influences my safety decisions (Participants 28). Simila rly, another participant wrote, t he other player is female like me and I more willingly trust a female I dont kn ow than I would a male (Participant 124). In one case, a woman who was perceived as portr aying masculine qualities was discriminated against because of them. Her ga me partner (after defecting) wr ote, I dont think I trust her because she seemed more tom-boyish. But maybe she is more nice than her exterior predicts. I dont really need to trust her one way or another (Participant 38). Inte restingly, this greater trust of females by certain participants did not ma nifest itself in high cooperation rates for the White female group that was most likely to em ploy this rationale in explaining their game decision. Instead, a stated emphasis on issues of threat, risk, and sa fety resulted in the more selfserving and competitive choice, defection. There are several explanations for White wo men having the highest defection rates. The most common is that White women did not want to be taken advantage of (s uckered) or be seen as weak. A second, less obvious reason is that th ey are using a gender ster eotype to predict and take advantage of their female partner. In othe r words, certain female participants may perceive cooperation as a gendered w eakness (but one they need not adhere too) that they can exploit for their own gain. This is evident in defection st atements such as I choose to defect because I think [she] would cooperate or normally I coope rate in my day-to-day life, choosing defect breaks me out of my normal routine and with my personality, I am usually one to take an opposite and independent stand on is sues (Participants 46, 40, and 28). Rather than cooperating and thus conforming to gendered expectati ons, many White women may be doing the opposite for either personal gain or as a result of the high personal risk that c ooperation entails.

PAGE 75

75 Mens responses also reflect gendered pattern s in that they are le ss concerned about the personal risk of cooperation and more concerne d about being perceived as bad, greedy, selfinterested, and/or biased. Masc ulinity places them in a position of strength where from they can take perspectives similar to that of the male part icipant who wrote, I coul d defect and be almost positive Ill get $9 but Ill feel lik e an ass and $3 is not worth the social punishment... I dont think too many would risk the norm al cordial social lives they have with strangers for a few bucks (Participant 19). Thus, for men (and pe rhaps Black women) it may be that cooperation yields more social rewards than potential economic risks, and that the social punishments defection may entail take precedence over the economic advantages defection provides. In addition to, and sometimes coinciding w ith, gender, a game partners race was also listed by a handful of participants as a factor in their decision making pr ocess. However, it should be noted that no participants in the White male, White female, or male cross-race experiment groups mentioned race in any of their written responses. Instead, discussion of race as a decision making factor appe ared exclusively in the Black male, Black female, and female cross-race groups. One White woman in the female cross-race e xperiment group wrote, I feel that because she is female, the other player stands a greater chance of cooperating than defecting. However, I also feel that an African-American woman may be more likely to stand up for what she wants by defecting, than to take the chance of cooperating. Based on these fact s I feel that I have a greater chance of winning more money, and I have guara nteed that I will not leave empty-handed (Participant 126). Another example of racial infl uences on a participants decisi on to defect were evident in this response by a Bl ack participant:

PAGE 76

76 I feel that in this world, the one person we must fi rst look out for is you. This by no means is suggesting that people should be selfish with one another. I just feel as though in order to help the people ar ound us, we must be at our most optimal level (financially, psychol ogically, socially, emotionally, etc.). I was always taught not to trust anyone, but still treat others the way you want to be treated. I can say that I believe most people have a more selfish (for the lack of a better word) mindset than an altruist ic one. I briefly met with th e girl in the other room. She seemed very outgoing and willing to talk. I would assume she is of Caucasian descent, and although I would never automatically assume that she has any negative feelings toward me, my expe riences growing up have taught me that whites secretly do not like blacks on a stan dard equal to them. Ive even been told they have a secret agenda toward blacks (Participant 129). Asked whether he believed his game partner wo uld cooperate, a Black male wrote, I would hope so because we are both Black and Greek12. I think that hell coope rate (Participant 53). Though the quantitative component of this study showed that certain racial differences in cooperation levels exist, the so cietal taboos against openly disc ussing race make the precise nature of and reasons for these differences difficu lt to uncover. In the following section I will summarize and discuss the open-ende d data findings of this study. Discussion of Open-ended Data Findings The open-ended data findings on the external validity of the study support the view that the social dilemma present in a monetary Prisoners Dilemma game is to a certain extent relevant to other social dilemma situations. Defining exactly which social dilemmas Prisoners Dilemma findings can be used to genera lize about, however, is a question whose answer seems to vary largely from person to person based on subjective fr ames of reference and personal histories. It is also a question that need s more in depth examination. The analysis of open-ended questions also demonstrates that th e most common reasons given for both cooperation and defection are very similar across various racial and gender group 12 By Greek the participant presumably means affiliated with a fraternity. Since participants in the study were not allowed to participate with someone they knew previously, measures were adopted to ensure members of the same

PAGE 77

77 pairings. Secondary response themes for coope ration, on the other hand, ha ve greater variance by racial and gender groups. Blacks were more likel y to view the decision to cooperate or defect as a moral choice and reflection of an individuals character than were Whites. This helps to explain the greater cooperation le vels of Blacks and the statistica lly significant correlation of the religiosity, race, and cooperation variables. Differences in responses by gender lend limited support for the view advocated by Simpson (2003) that women are more likely than men (particularly Black men in this study) to defect out of fear rather than out of greed. Nonetheless, a significant difference between Black female and White female cooperation rates point to the important role that cultural factors such as racial and religious group member ship can play in social dilemm a decisions. In light of this, the findings of past studies that rely exclusiv ely or overwhelmingly on White participants should be revisited and duplicated using broader, more diverse samples. Furthermore, it should be noted that among the various race and gender experiment groups, participants belonging to minority groups are much more likely to take racial and/or gender group membership into account in their game decisions. This is support ed by the finding that not a single White male participant mentioned either the race or gende r of a game partner as a factor in their cooperation/defection decision. Similarly, only one Black male mentioned gender as a factor in his cooperation/defection decision. White women in the study were most likely to mention the gender of their game partner as a factor in their decision making, yet only one White women cited race. In line with intersection theory, Black women who may have faced subjugation and discrimination due to both their race and gender, were the only racial/gender group participants in the study to list both race and gender as f actors in their cooperation/defection decision fraternity or sorority group were not part nered with one another. In this case the participants knew of each others Greek affiliation because they were both wearing shirts identifying their respective fraternities.

PAGE 78

78 multiple times. In this group the intersection of race and gender coupled with the influence of religiosity resulted in significantly higher coope ration rates than were present in the White female experiment group. The complete absence of race talk in the cr oss-race male experiment group, despite clear differences in the language participants in this group versus same race groups used to describe their game partners, suggests th at race, as a basis for cooperati on/defection and as a topic of discussion, remains taboo. Additionally evidence of this taboo is apparent in the lack of references to the physical appearance of Black game partners by White males in the cross-race group. Though Bonilla-Silva (2003) has documented how many times the alleged colorblindness of most Whites serves as a mechanism for maintaining White supremacy, the greater cooperation rates of White participants in cross-race groups is likely not the result of covert racism, but instead may be evidence of Bells (1980) Interest Convergence Hypothesis In other words, White participants may be viewing the experiment as a chance to show they are non-racist by cooperating with Black participants for a mutually beneficial out come. There is also modest evidence suggesting that Black male participants at this predominantly White university may have bought into the sincere fic tions of Whiteness presented by our society and discussed earlier (Chapter 2), and that they are more likely to c ooperate with White males as a result. Due to the limitations of the open-ended data gathered, more in-depth research is needed to substantiate these theoretical conclusions. In the final sect ion of this paper, I will conclude and present suggestions for future research.

PAGE 79

79 Table 5-1. Thematic coding categories with frequenc ies for responses to the question, Why did you cooperate or defect? Please explain in detail. Cooperation Themes Defection Themes 1) Mutually benefici al (56) 1) Defection guaranteed money (38) 2) Person will think or act like me (26) 2) Made maximum pay-out possible (22) 3) I am a good/cooperative person (14) 3) I do not know the other person (11) 4) Morally superior choice (11) 4) Defection is the safer option (9) 5) Wanted to avoid feelings of guilt (10) 5) Defection is the logical choice (4) 6) Person seemed nice/friendly (9) 6) I think other pers on will cooperate (3) 6) Did not want to sucker the player (9) 7) Defected because I am greedy (2) 8) I do not need the money (8) 7) Defected because the stakes are low (2) 8) Faith in humanity (8) 9) Defected due to persons appearance (1) 8) Defecting is greedy (8) 9) Cooperation could result in zero gain (1) 11) Cooperating is fair (7) 12) Cooperation is th e better choice (6) 13) May have defected if more at stake (5) 14) Gender influenced my decision (4) 14) I have nothing to lose (4) 16) Makes me feel better about myself (3) 17) Makes me the better person/man (1) *Frequency distribution of response s is in parenthesis after each reason for cooperation and defection. *Multiple responses were counted for participants who provided them. Table 5-2. Reasons for cooperation and def ection by experiment group with frequency of response Experiment Group Reasons for cooperation Reasons for defection White Males n = 13 n = 11 n = 24 1. Mutually beneficial (9 ) 1. Guaranteed pay-out (10) 2. Faith in humanity (5) 2. Maximum pay-out possible (3) 3. Nice/friendly partner (4) 3. Partner will act like me (3) 4. Better choice (3) 5. Partner will act like me (3) White Females n = 11 n = 13 n = 24 1. Mutually beneficial (6 ) 1. Guaranteed pay-out (10) 2. Against suckering pa rtner (3) 2. Maximum pay-out possible (6) 3. Do not need money (3) 3. Partner is a stranger (3) 4. Gender of partner (3) 5. Partner will act like me (3) Black Males n = 15 n = 9 n = 24 1. Mutually beneficial (10) 1. Maximum pay-out possible (6) 2. I am a good person (4) 2. Guaranteed pay-out (4)

PAGE 80

80 Table 5-2 continued Experiment Group Reasons for cooperation Reasons for defection Black Females n = 18 n = 6 n = 24 1. Mutually beneficial (7) 1. Guaranteed pay-out (3) 2. I am a good person (4) 2. Maximum pay-out possible (3) 3. Partner will act like me (4) 4. Moral/altruistic choice (3) 5. Do not need money (3) 6. Guilt avoidance (3) Cross-race White Males n = 8 n = 4 n = 12 1. Mutually beneficial (5) 1. Economically rational (3) 2. Guaranteed pay-out (2) 3. Safest option (2) Cross-race Black Males n = 11 n = 1 n = 12 1. Mutually beneficial (8) 1. Partner is a stranger (1) 2. I am a good person (3) 3. Moral/altruistic choice (3) Cross-race Black Females n = 8 n = 4 n = 12 1. Mutually beneficial (7) 1. Guaranteed pay-out (3) 2. Partner will ace like me (4) Cross-race White Females n = 7 n = 5 n = 12 1. Mutually beneficial (4) 1. Guaranteed pay-out (5) 2. Nothing to lose (3) 2. Safest option (2) 3. Defecting is greedy (3) *Frequency distribution of response s is in parenthesis after each reason for cooperation and defection. *Multiple responses were counted for participants who provided them.

PAGE 81

81 CHAPTER 6 CONCLUSION Though experimental games such as the Pr isoners Dilemma have long been popular, questions concerning their relevan ce to real-life settings and their general lack of theoretically grounded hypotheses have contributed to the resu lts of gaming studies being largely ignored by social scientists outside the field of experime ntal gaming (Pruitt & Kimmel 1977). Despite this disregard, experimental games are valuable in studying social dilemmas in part because they permit precise measurements of elusive variables such as extent of cooperation, and because they allow hostile and competitive behavior to transp ire without injury to participants or their relationships (Pruitt & Kimmel 1977). The need to comprehend social dilemmas is of critical importance since as Van Lange and Messick note, the functioning of societie s, groups, and relationshi ps is perhaps most strongly challenged by social dilemmas because the well-being of these larger units is threatened when most or all individuals pursue their own well-being rather than the collective well-being (1996:93). In the mode rn era, technological advancements in nuclear and biological weaponry, communications systems, and industry ha ve created with them new social dilemmas in which defections mutually deficient outco mes (and the human suffering those outcomes may entail) take place on a global scale. Since the wellbeing of much of humanity lies in our ability to cooperate on major social dilemmas, we must understand and be capable of overcoming socially constructed differences in race, gender, ethnicity, nationality, id eology, and religion. As stated by civil rights leader Martin Luther King Jr., We must learn to live t ogether as brothers or perish together as fools (1964). Keeping in mind the need to employ theo retically grounded hypothe ses to uncover the racial and gender intricac ies of social dilemma cooperation, th e aims of this study consisted of:

PAGE 82

82 (1) determining if Black college students, as me mbers of a subordinated group, are more likely to cooperate in a social dilemma th an are White college students w ho are members of the dominant group; (2) determining if female college students are more likely to cooper ate with each other in a social dilemma than are male college students; (3) exploring the reason s and rationale given by respondents for cooperation or defection and (4) tes ting if levels of trust and expectations of a game partners decision differ between race and gender groups. In regard to the first aim, this study provi des support for the hypot hesis that members of subordinated racial groups are more likely to coop erate in a social dilemma than are members of a racially dominant group. In the construc ted social dilemma employed by this study, quantitative analyses demonstrat ed that Blacks had both higher intra-group cooperation rates than did Whites (69% intra-gr oup cooperation for Blacks compared to 50% for Whites) and significantly higher overall cooper ation rates than did Whites (72% overall cooperation for Blacks compared to 54% for Whites). In other wo rds, Black participants were more likely to cooperate with their game partner (irrespective of that game partners race) than were White participants. This finding lends limited support to the theory that Blacks have developed a culture of cooperation as a response to White racism, and additionally suggests that Blacks (especially Black males) are willin g to extend that culture of coopera tion to include Whites. With regard to the second aim, overall gender cooperation levels did not significantly differ between males (65% cooperation) a nd females (61% coope ration), although the cooperation rates in the Black female experime nt group (75% cooperation) and White female experiment group (46% cooperation) did. These differing levels of female cooperation by race suggest that in the United Stat es cultural factorssuch as race and religiosityplay a more important role than gender in a participants de cision to cooperate or de fect. These differences

PAGE 83

83 also potentially reflect an important intersec tional difference in the way women of different racial groups are influenced by gender subordi nation. In this, subordi nation may be leading Black women to cooperate at higher rates and Wh ite women to defect (mostly out of fear) at higher rates when presented with a monetary social dilemma. With regard to the third aim, it was found that although the reasons and rationale given by participants in different racial and gender experiment groups for cooperation and defection were mostly similar, minority racial and gende r group members were more likely to account for their minority group status in their cooperation/de fection decision. This greater awareness of race and/or gender group status by minority gro up members, coupled with the higher cooperation rates of Black participants, lends support to the claim that expe riences of subjugation (such as McCormick and Franklins [2000] inclusionary dilemma) and the increased group awareness that they generate can manifest themselves as increased minority group cooperation in social dilemmas. Furthermore, the finding that White women ha d the lowest cooperation rates of any race and gender experiment group can be interpreted to mean that most White women do not see themselves as subjugated. This interpretation of the data is consistent with the work of Lengermann and Wallace (1985) who argue that Wh ite women are less likely to see themselves as an oppressed minority group than are Black women. Additionally, it was found that Black participants were more likely to vi ew the decision to cooperate or defect as a moral issue and as a reflection of character, than were Whites. This finding is likely correlat ed with the significantly higher levels of religiosity repor ted by Black participants compared to White participants. With regard to the fourth aim, quantitative analyses revealed no signi ficant race or gender differences on the trust variable and no gender differences on the variable measuring expectation

PAGE 84

84 of game partners cooperation de cision. There was, however, a st atistically significant difference between Black and White participants expectatio n of game partners cooperation decision when participants who indicated they were uncertain as to whether th eir game partner would cooperate or defect were excluded from the cross-tabulation. With pa rticipants who reported being uncertain filtered out, 72% of the remaining Black participants believed their game partner would cooperate compared to only 56% of White participants. Like the findings on cooperation rates, this finding also suggest s that subjugation increases a gr oups cohesiveness and ability to work together in social dilemmas. Lastly, content analyses on the trust and e xpectation variables reve aled that among males in the study, Whites were more inclined to report making judgments based on physical appearance when playing a member of their same racial group, but refr ained from mentioning physical appearance when paired with a Black ga me partner. Possible explanations for White males omitting mention of the physi cal appearance of Black game partners is that race is considered by Whites to be a taboo topic, and th at references to the physical appearance of a Black game partner may be interp reted as racist. Unlike most previous studies employing the Prisoners Dilemma, the present study did not rely on a sample composed almost exclusivel y of White participants and was therefore able to generate insights into Black same-race a nd cross-race social dilemma cooperation and defection. Additionally, the methodol ogical decision to assign participants to one of six race and gender groups allowed for cross-race group compar isons and for analyses of how the intersection of race and gender influences social dilemm a cooperation and defection decisions. Other methodological advantages of the present study in cluded the use of a singl e-shot rather than multi-shot game, and the utilization of theoretically based hypotheses. Both of these

PAGE 85

85 methodological design elements served to improve the validity and generali zability of the studys findings. Compared to previous Prisoners Dilemma re search on cross-race pairs, this studys findings on racial cooperation cont radict many of the earlier studi es conducted in the 1960s and 1970s that showed lower levels of cross-race cooperation comp ared to same-race cooperation. Nonetheless, this studys findings coincide with the most recent study of Heider and Skowronski (2007), which produced results that were the exact converse of the earlier studies. This reversal, coupled with the finding that White males refrai ned from referring to the physical appearance of Black game partners, may be evidence of a societ al shift toward less race -based discrimination. However, an equally plausible and less optimistic interpretation of the cro ss-race findings is that they merely reflect a desire among participants to appear non-racist. As Heider and Skowronki have noted, this latter interpre tation suggests that White participants in the study were aware of their potential prejudice and we re overtly trying not to treat the African American partner poorly (2007:61). Additionally, these findings may al so be evidence of Bells (1980) Interest Convergence Hypothesis whereby Whites realize it is in their intere st to cooperate with Blacks for a mutually beneficial economic outcome. This studys findings on racial cooperation add to the racial relations and social dilemma literature by showi ng both that (1) racial minority group status positively correlates with social dilemma cooperati on and (2) that anti-racism may be influencing Whites cooperation/defection decisionsin fa vor of increased cooperationwhen they are paired with a Black game partner. On the question of gender cooperation in the Prisoners Dilemma, this study like most before it, suggests no overall difference in coope ration levels by gender. This, however, does not

PAGE 86

86 necessarily imply that gender does not play a role in social dilemma cooperation since many studies conducted at other times and places have s hown that it does. What the lack of overall gender difference implies, rather, is that cult ural difference, includi ng differences in gender socialization and levels of pa triarchy and subjugation, are more important to social dilemma cooperation than is the biological cl assification of female or male. Considering that the sample was composed entirely of undergraduate college students, this studys findings may also be indicative that U.S. anti-discrimination educational curriculums, programs, and policies are redu cing overt discrimination against Blacks. Furthermore, the findings imply that minorities are more likely to account for their minority group membership in social dilemma decisionmaking, and that among Whites, concerns about political correctness may suppress competitive beha vior toward Blacks and make the mentioning of race in formal settings taboo. The finding that Blacks have significantly higher same-race and ove rall cooperation rates than do Whites, may be evidence that Black students at predominantly White universities are collectively resisting ex periences of subjugation and/or alie nation through cooperative strategies that emphasize group interests over i ndividual interests. In this, gr eater Black coope ration in this social dilemma may in part be the result of copi ng strategies developed by Black students to overcome social-structural and institutional hurdles faced by students of color. The policy implications of the above findings are limited by the usual issues that arise when trying to generalize from an experimental se tting to a target setti ng and by the variation in behavior that different types of social dilemmas may elicit. With this in mind, any attempts at generalization should be made cautio usly and should be limited to social dilemmas that involve the allocation of finite resources such as those that often occur in the economic, social, political,

PAGE 87

87 and interpersonal arenas. In th ese arenas, the most practical imp lication of this study centers on the finding that racial domination, be it in the fo rm of racism, discrimination, subjugation, and/or exploitation, increases a groups aw areness of its minority status a nd concomitantly also increases its cohesiveness and likelihood of social dile mma cooperation. Researchers from the various fields that employ games theory, social dilemm a theory, social psychol ogical theory, rational choice theory, and symbolic interaction theory may find the results of this study relevant to their work. Additionally, policy makers and activists in education, government, the military, the labor movement, non-governmental organizations and in ot her sectors that require an understanding of the intricacies of soci al dilemma cooperation and defection may also find this study useful. Considering this is one of the few studies to ut ilize Black participants in the examination of social dilemma behavior and to provide cross -race comparisons of social dilemma cooperation and defection rates, these findi ngs should not be overlooked. Suggestions for future Prisoners Dilemma a nd social dilemma research include the need for researchers to account for and study the religiosity leve ls of participants in their sample. As discussed in Chapter 4, the inte rsection of race and re ligiosity complicate this studys findings. Based on the theoretical rationale that grounds this st udys hypotheses, the observed differences in racial cooperation are likel y the result of shared experien ces of minority group subjugation. However, the significant differences in levels of religiosity by race make it feasible that observed racial differences in cooperation ar e in actuality the product of differen ces in levels of religiosity. In addition to the need to account for the infl uence of religiosity, th e need to incorporate significantly more racial minority group participants into future Prisoners Dilemma studies is also great. Latinos and AsianAmericans in particular are tw o large and growing US minority groups that have been habitually ignored by Prisoners Dilemma researchers. Corresponding

PAGE 88

88 with a greater emphasis on racial minority groups participants, more re search needs to be conducted to corroborate the finding s of this present study. In the written responses to this studys questionnaire, much about th e racial factors that influen ce cooperation/defection was left unsaid. Qualitative methods such as in-depth in terviews and focus groups should be utilized to uncover richer data on the more covert factors that influence a participants cooperation/defection decisions. Quantitatively, games other than the Prisoners Dilemma, such as ones based on monetary ultimatums, should be employe d to measure racial social dilemma cooperation/defection. Doing so c ould not only reinforce, but also speak to the exte rnal validity of the findings presented here. Also, em ploying a sample composed exclusively of undergraduate university students is a limitation in this study that future research should address by drawing from a more diverse pool of par ticipants. The uniqueness of the university environment may have influenced this st udys outcome in ways that limit broader generalizability. The focus of the present research has been to use the Prisoners Dilemma to identify the links between race, gender, and levels of coopera tion, trust, and expectations of a partner in social dilemmas. However, due to a limited am ount of research employing games theory to study race relations, many questions still remain unanswered. By employing games similar to the one adapted for this study, future researchers could gauge racial disc rimination by comparing intraand inter-race cooper ation rates for a given school, organization, town, or city. Additionally, these results could be compared across time and place, providing hard data on the relative levels of societal racial discrimination. In this, the application of soci al decision making games to measure racial discrimination would be us eful in that they perm it hostile and aggressive

PAGE 89

89 behavior to transpire without harm to participan ts or as Pruitt and Kimme l put it, these games permit conflict without tears (1977:366). Furthe rmore, these games have the added advantage of providing researchers with a way to measure racial discrimination with out revealing what is being measured to the participants of the study (who may otherwise in advertently taint the results). Finally, the potential for games in gene ral and the Prisoners Dilemma in particular to be utilized as a barometer of institutional and soci etal levels of racial di scrimination is great and should be further explored.

PAGE 90

90 APPENDIX A MEASUREMENT INSTRUMENT The following is a game in which you will have the opportunity to earn money for your participation in this social decision making study. You will be playing the game alongside another player (who is the person you just met). The game is simple; each of you can decide to cooperate with each other or to defect and work against one another. Be aware that there are benefits and risks attached to both cooperation an d defection. The games payout is a follows: If you both cooperate you each receive $6. If you both defect you each receive $3. If one player cooperates and the other defects, the person who cooperated will receive $0, and the person who defected will receive $9. Game 1. Please circle one of the following choices : Cooperate or Defect Total $12 $6 $6 Total $12 $9 $0 Total $6 $3 $3 Total $9 $0 $9 Other Player Cooperates Defects Cooperate You Defect

PAGE 91

91 2. Why did you cooperate or defect ? Please explain in detail. ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________

PAGE 92

92 3. Do you feel like you can trust th e other player? Why or why not? ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________

PAGE 93

93 4. Do you believe the other player will cooperate? Why or why not? ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________

PAGE 94

94 Demographic Background 1. Gender: (circle one) Male Female 2. Age (in years) ______ 3. How would you describe your racial or ethnic background? (circle all that apply) Black (please specify)________________________________________ African American ___________________________________________ Asian (please specify) _______________________________________ White (please specify) _______________________________________ Caucasian (please specify) ____________________________________ Hispanic (please specify)______________________________________ Latino (please specify)________________________________________ Other (please specify)_________________________________________ 4. College classification (please circle one) a. freshman b. sophomore c. junior d. senior e. graduate student 5. College Major? _________________________ 6. How many siblings do you have? ___________

PAGE 95

95 7. How many years of education do you expect to complete? (circle one) a. B.A. b. M.A./M.S./M.B.A./M.S.W. c. J.D. (law) d. Med./Vet/Dent. e. Other Doctoral 8. What is your religious preference? (circle one) a. Protestant -what denomination?_____________________________________ b. Catholic c. Jewish d. Other please specify ______________________________________________ e. Christian what denomination? ______________________________________ e. No religion 9. How religious are you? A. very religious B. average C. below average 10. What is your familys year ly household inco me? (circle one) a. $25,000 and below b. $26,000 $40,000 c. $41,000 $80,000 d. $81,000 160,000 e. Above $160,000

PAGE 96

96 11. Are you affiliated with a fraternity or sorority? (circle one) Yes or No If you answered yes, what fraternity or sorority ___________________________ 12. Which of the following political orient ations best describes you? (circle one) A. Conservative B. Liberal C. Centrist D. Other (please specify) _____________________ 13. What political pa rty do you identify with? A. Republican Party B. Democratic Party C. Green Party D. I am independent E. Other (please specify) ____________________

PAGE 97

97 APPENDIX B INFORMED CONSENT Informed Consent Protocol Title: Social Decision Making Study Please read the consent form carefully before you decide to participate in this study Purpose of the research study: The purpose of this study is to examine how people respond to social dilemmas. What you will be asked to do in the study: You will be asked to meet another student and afterward pl ay a quick social decision making game. You will also be asked to fill out a short survey. Time required: 10-15 minutes Risks and Benefits : You may or may not experience minor feelings of anger or guilt after playing the game. We do not anticipate any other risks to you by participating in the study. The benefit of the study is the opportunity to earn $0, $3, $6, or $9 for completing the game and survey. Confidentiality: Your identity will be kept confidential to the extent provided by law and your name will not be used in any report. Voluntary Participation: Your participation in this study is completely voluntary. There is no penalty for not participating. Right to withdraw from the study: You have the right to withdraw from the study at anytime without consequence. Whom to contact if you have questions about the study: Victor Romano M.A., Sociology Department, 3357 Turlington Hall, 786-338-1027 Supervisor: Hernan Vera Ph.D., Professor of Sociology, 3229 Turlington Hall, 392-0265 ext. 232 For questions about your rights as a research participant, please contact the UFIRB office at 352-392-0433 or PO Box 112250, Gainesville, Florida 32611. Agreement: I have read the procedure described abov e. I voluntarily agree to participate in the procedure and I have received a copy of this description. Participant: __________________________________ Date:_________________ Principal Investig ator:__________________________ Date:_________________

PAGE 98

98 LIST OF REFERENCES Anderson, Margaret L., and Patricia H. Collins. 2000. Race, Class, and Gender: An Anthology (4th Edition) Wadsworth. Apfelbaum, Erika. 1974 On Conflicts and Bargaining. Advanced Experimental Social Psychology 7:103-56. Assensoh, Akwasi B. 2000. Conflict or Cooperation? Pp. 113-39 in Black and Multiracial Politics in America edited by Y.M. Alex-Assensoh & L. J. Hanks. New York University Press. Axelrod, Robert. 1980. Effective Choice in the Prisoners Dilemma. The Journal of Conflict Resolution 24:3-25. Axelrod, Robert. (1984). The Evolution of Cooperation. Basic Books, Babbie, Earl. 2001. The Practice of Social Research (9th Edition). Wadsworth/Thomson Learning. Baca Zinn, Maxine. 1996. What is Multicultural Feminism? Feminist Studies 22:321-31. Baca Zinn, Maxine, and Bonnie T. Dill. 1994. Women of Color in USA Society Temple University Press. Barash, David.P. 2004. Caught Between Choi ces: Personal Gain vs. Public Good. Chronicle of Higher Education 50:B13-16. Baxter, George W. Jr. 1973. Prejudiced Libera ls? Race and Information Effects in a TwoPerson Game. The Journal of Conflict Resolution 17:131-161. Bell, Derek. 1980. Brown v. Board of Educatio n and the Interest-Convergence Dilemma. Harvard Law Review 518:93. Benson, Peter., Stuart A. Karabenick, and Rich ard M. Lerner. 1976. Pretty Pleases: The Effects of Physical Attractiveness, R ace, and Sex on Receiving Help. Journal of Experimental Social Psychology 12:409-15. Bernard, H.Russell. 2000. Social Research Methods: Qualita tive and Quantitative Approaches. Sage Publications. Blau, Peter. 1997. Inequality and Heterogeneity. Free Press. Bonilla-Silva, Eduardo. 2003. Racism Without Racists: ColorBlind Racism and the Persistence of Racial Inequality in the United States. Rowman and Littlefield.

PAGE 99

99 Brown, Kelly M. and Laura O. Taylor. 2000. Do As You Say, Say As You Do: Evidence on Gender Differences in Actual and Stat ed Contributions to Public Goods. Journal of Economic Behavior and Organization 22:255-67. Brown, Kendrick T. 1999. Skin Tone Bias and African-American Well-Being: A Dual Influence Model Approach. Dissertation Abstracts Intern ational: Section B: The Sciences and Engineering, 59:5616. Butler, Judith. 1990. Gender Trouble: Feminism and the Subversion of Identity. Routledge. Calhoun, Cheshire. 2000. Feminism, the Family, and the Politic s of the Closet: Lesbian and Gay Displacement. Oxford University Press. Carby, Hazel V. 2000. White Wo men Listen! Pp. 389-403 in Theories of Race and Racism: A Reader edited by Les Back and J ohn Solomos. Routledge. Cederblom, Douglas, and Carol J. Diers. 1970. E ffects of Race and Stra tegy in the Prisoners Dilemma. The Journal of Social Psychology 81:275-286. Chong, Dennis 1991. Collective Action and the Ci vil Rights Movement. Chicago University Press. Clark, Kenneth, and Martin Sefton. 2001. The Sequential Prisoners Dilemma: Evidence on Reciprocation. Economic Journal 111:51-68. Collins, Patricia. H. 1997. Comment on He kmans Truth and Method: Feminist Standpoint Theory Revis ited: Wheres the Power? Signs 22:375-79. Dawson, Michael. 1994. Behind the Mule : Race and Class in African American Politics Princeton University Press Delgado, Richard. 1995. Race, Sex, Class, and their Intersections Introduction to part VI. P. 249 in Critical Race Theory: The Cutting Edge, edited by Richard Delgado. Temple University Press. Doane, Ashley W., and Eduardo Bonilla-Silva. 2003. White Out: The Continuing Significance of Racism. Routledge. Downing, Leslie L., and Kenneth H. Bothwell. 1979. Open-space Schools: Anticipation of Peer Interaction and Development of Cooperative Interdependence. Journal of Educational Psychology 71:478-484. Du Bois, W.E.B. [1903] 2004. The Souls of Black Folk. Paradigm Publishers. Du Bois, WE.B. [1935] 1992. Black Reconstruction in America 1860-1880 Atheneum.

PAGE 100

100 Dubois, Frdrique, and Luc-A lian Giraldeau. 2003. The Fora ger's Dilemma: Food Sharing and Food Defense as Risk-Sensitive Foraging Options. The American Naturalist 162:768-79 Feagin, Joe R. 2000. Racist America: Roots, Current Realities and Reparations Routledge. Feagin, Joe R. 2006. Systemic Racism: A Theory of Oppression Routledge. Feagin, Joe R., and Clarice B. Feagin. 2002. Racial and Ethnic Relations (7th Edition) Prentice Hall. Feagin, Joe R., Hernan Vera, and Pinar Batur. 2001. White Racism: The Basics. Routledge. Frankenberg, Ruth. 1993. White Women, Race Matters: The Social Construction of Whiteness University of Minnesota Press. Gilligan, Carol. 1982. In a Different Voice: Psychol ogy and Womens Development. Harvard University Press. Hacker, Andrew. 1995. Two Nations: Black and White, Separate, Hostel, Unequal Ballentine Books. Heider, Jeremy D., and John J. Skowronski. 2007 Improving the Predictive Validity of the Implicit Association Test. North American Journal of Psychology 9:53-76. hooks, bell. 1991. Yearning: Race, Gender, and Cultural Politics. South End Press. Hu, Yung-An, and Liu Day-Yang. 2003. Altruism Versus Egoism in Human Behavior or Mixed Motives. American Journal of Economics and Sociology 62:677-704. James,Tammy., Lewis Soroka, and John G. Benjaf ield. 2001. Are Econom ists Rational, or Just Different? Social Behavior and Personality 29:359-64. Kaplan, Michael, & Ellen Kaplan. 2006. Chances Are: Adventures in Probability. Penguin Books. King, Debra W. 1988. Multiple Jeopardy, Multiple Consciousness: The Context of a Black Feminist Ideology. Journal of Women in Culture and Society 14:42-72. King, Martin L. 1964. Speech in St. Louis, Missouri March 22. Ledyard, John O. 1995. Public Goods: A Survey of Experimental Results. Pp. 111-94 in Handbook of Experimental Economics edited by J.H. Kagel and A.E. Roth. Princeton University Press.

PAGE 101

101 Liebrand, Wim B.G. 1986. The Ubiquity of Soci al Values in Social Dilemmas. Pp. 113-33 in Experimental Studies of Social Dilemmas edited by H. Wilke, C. Rutte, and D.M. Messick. Frankfurt, Ger.: PeterLang. Lengermann, Patricia M., and Ruth A Wallace. 1985. Gender in America: Social Control and Social Change. Prentice-Hall Lever, Janet. 1978. Sex Differences in the Complexity of Childrens Play and Games. American Sociological Review 43: 471-83. Lorber, Judith. 1994. Paradoxes of Gender. Yale University Press. Lorber, Judith. 2001a. The Social Cons truction of Gender.. Pp. 121-24 in Womens Voices, Feminist Visions edited by S. M. Shaw and J. Lee. Mayfield Publishing Company. Lorber, Judith. 2001b. Gender Inequality (2nd Edition). Roxbury Publishing Company. Macionis, John J. 2005. Sociology (10th Edition). Pearson Prentice Hall. Macy, Michael W., and John Skovoretz. 1998 The Evolution of Trust and Cooperation between Strangers: A Computational Model. American Sociological Review 63:638-60 Martin, Patricia Y. 2004. Gende r As Social Institution. Social Forces 82:1249-73. Marx, Karl. [1862] 2006. Economic Manuscripts of 1861-1863: Part 3. Relative Surplus Value. Retrieved November 4, 2006 from the World Wide Web http://www.marxists.org/archive/marx/works/1861/economic/ch31.htm McAdams, Richard H. 1995. Cooperation and Conflict: The Economics of Group Status Production and Race Discrimination. Harvard Law Review 108:1005-85. McCormick, Joseph P., and Sekou Franklin. 2000. E xpressions of Racial Consciousness in the African American Community. Pp. 315-36 in Black and Multiracial Politics in America, edited by Y.M. Alex-Assensoh and L. J. Hanks. New York University Press. Messner, Michael A. 1990. Boyhood, Organized Sports and the Construction of Masculinities. Journal of Contemporary Ethnography 27:197-211. Mims, Patricia R., John J. Hartnett, and Robert W. Nay. 1975. Interpers onal Attraction and help volunteering as a Function of Physical Attractiveness. Journal of Psychology 89:125-31. Minkler, Lanse.P. and Thomas J. Miceli. 2004. Lying, Integrity, and Cooperation. Review of Social Economy 62:28-50.

PAGE 102

102 Morse, Stanley .J., Harry T. Reis, Joan Gruzen, and Ellen Wolff. 1974. The Eye of the Beholder: Determinants of P hysical Attractiveness Judgments in the U.S. and South Africa. Journal of Personality 42:528-42. Nemeth, Charlan. 1972. A Critical Analysis of Research Utilizing the Prisoners Dilemma paradigm for the study of bargaining. Advanced Experimental Social Psychology 6:20334. Nowak, Martin. 2004. Prisoners of the Dilemma. Nature 427:491. Orbell, John M., Robyn M. Dawes, and Peregr ine Schwartz-Shea. 1994. Trust, Social Categories, and Individual s: The Case of Gender. Motivation and Emotion 18:109-28. Owens, Timothy K. 1998. Effects of Fear and Bias Perceptions of Interracial Cooperation in Young Males. Dissertation Abstracts International : Section B: The Sciences and Engineering 59:2488. Pruitt, Dean G., and Melvin J. Kimmel. 1977. Twe nty Years of Experimental Gaming: Critique, Synthesis, and Suggestion for the Future. Annual Review of Psychology 28:363-92. Reskin, Barbra. 2002. The Proximate Caus es of Employment Discrimination. Contemporary Sociology 29:319-28. Risman, Barbara. J. 2004. Gender as a Social Structure: Theory Wrestling with Action. Gender and Society, 18:429-50. Rodgers, Reuel. 2000. Afro-Caribbean Immigrants, African Americans, and the Politics of Group Identity. Pp. 15-59 in Black and Multiracial Politics in America, edited by Y. M. Alex-Assensoh and L. J. Hanks. New York University Press. Rothenberg, Paula S. 1992. Race, Class, and Gender in the United States: An Integrated Study St. Martin Press. Russell, Kathy, Midge Wilson and Ronald Hall. 1993. The Color Complex: The Politics of Skin Color among African Americans Anchor Books. Sibley, Sally A., Sara K. Senn, and Alexis Ep anchin. 1968. Race and Sex of Adolescents and Cooperation in a Mixed-motive Game. Psychonomic Science 13:123-124. Simmel, Georg. 1955. Conflict and the Web of Group Affiliations. The Free Press.

PAGE 103

103 Simpson, Brent. 2003. Sex, Fear, and Greed: A Social Dilemma Analysis of Gender and Cooperation. Social Forces 82:35-52. Sjoberg, Lennart., Ingvar Bokande r, Lars Deneck, and Kerstin Lindbom. 1969. A Comparison between Children and Adults in the Prisoners Dilemma Game. Scandinavian Journal of Psychology 10:158-66. Spinney, Laura. 2004. Why We Do What We Do. New Scientist 183:32-36. Stockard, Jean, Alphons J.C. van de Kragt, and Patricia Dodge. 1988. Gender Roles and Behavior in Social Dilemma s: Are There Sex Differences in Cooperation and in Its Justification? Social Psychology Quarterly 51:154-63 Suarez-Orozco, Marcelo M., and Mariela M. Paez. 2002. The Research Agenda. Pp. 1 in Latinos: Remaking America, edited by M.M. Suarez-Orozco and M.M. Paez. University of California Press. Takaki, Ronald. 2000 Iron Cages: Race and Culture in 19th-Century America. Oxford University Press. Thomas, William I. 1931. The Relation of Rese arch to the Social Process. Pp. 175-94 in Essays on Research in the Social Sc iences: Papers Presented in a General Seminar Conducted by the Committee on Training of Th e Brookings Institution, 1930-31. Brookings Institution. Tilly, Charles. 1999. Durable Inequality. University of California Press. Van Lange, Paul A., and David M. Messick. 19 96. Part 1: Social Di lemmas: Psychological Processes Underlying Cooperation in Social Dilemmas. Pp. 93-112 in Social Agency: Dilemmas in Education Praxiology, edited b y W. Gasparski, M. Mlicki & B. Banathy. Transaction Pub. Vera, Hernan, and Andrew M. Gordon. 2003a. Screen Saviors: Hollywood Fictions of Whiteness. Rowman and Littlefield Vera, Hernan, and Andrew M. Gordon. 2003b. The B eautiful American: Sincer e Fictions of the White Messiah in Hollywood Movies. Pp. 113-25 in White out edited by A. Doane & E. Bonilla-Silva. Routledge. Wallace, Ruth A., and Alison Wolf. 1999. Contemporary Sociological Theory: Expanding the Classical Tradition. Prentice Hal. West, Candace, and Don H. Zimmerman. 1987. Doing Gender. Gender and Society 1:125-51. West, Stephen G., and Jan T. Brown. 1975. Physic al Attractiveness, th e Severity of the Emergency and Helping: A Field Expe riment and Interpersonal Simulation. Journal of Experimental Social Psychology 11:531-38.

PAGE 104

104 Williams, Patricia J. 2000. R ace and Rights. Pp. 421-30 in Theories of Race and Racism: A Reader edited by Les Back and John Solomos. Routledge Wilson, Warner U., and Myra Kayatani. 1968. Int ergroup Attitudes and St rategies in Game between Opponents of the Same or a Different Race. Journal of Personality and Social Psychology 9:24-30. Wilson, William J. 1978. The Declining Significance of Ra ce: Blacks and Changing American Institutions (2nd Edition). University of Chicago Press Yamagishi, Toshio, and Karen S. Cook. 1993. Gener alized Exchange and Social Dilemmas. Social Psychology Quarterly 56:235-48

PAGE 105

105 BIOGRAPHICAL SKETCH Victor Eduardo Romano was born in 1979, in Queens, New York. The only son of Argentine immigrants, he grew up mostly in Miami, Florida, which is where he currently resides with his wife Kelli.