Essays on actor models in exchange networks and social dilemmas

217 Pages·2002·1.21 MB·English

by  Marcus Adrianus Leonardus Maria van Assen.

ESSAYS ON ACTOR MODELS IN
EXCHANGE NETWORKS AND
SOCIAL DILEMMAS
RIJKSUNIVERSITEIT GRONINGEN
ESSAYS ON ACTOR MODELS IN
EXCHANGE NETWORKS AND SOCIAL DILEMMAS
Proefschrift
ter verkrijging van het doctoraat in de
Psychologische, Pedagogische en Sociologische Wetenschappen
aan de Rijksuniversiteit Groningen
op gezag van de
Rector Magnificus, dr. D.F.J. Bosscher,
in het openbaar te verdedigen op
donderdag 28 juni 2001
om 14.15 uur
door
Marcus Adrianus Leonardus Maria van Assen
geboren op 1 april 1972
te Waalwijk
Promotores:
Prof. dr. F.N. Stokman
Prof. dr. W. Raub
Prof. dr. T.A.B. Snijders
Leescommissie:
Prof. dr. W.B.G. Liebrand
Prof. dr. S.L. Lindenberg
Prof. D. Willer
ISBN 90-367-1446-x
Preface
It proved difficult to write a preface that was, at least to a certain degree, satisfying to me.
Initially I thought that a short and to the point preface, such as “Folks, thanks for everything”,
would suffice. This would not be satisfactory though, because it does not answer the question
about which folks I thank and for what. The other extreme would have been a preface in
which I provide a long list of folks who have made my life enjoyable over the first 348
months of my life general, and in particular over the last 15.8% of it. However, this preface
would then have consisted of an enumeration of at least five pages. Moreover, a similar
enumeration would be more appropriate in an autobiography than in a preface of a
dissertation. Therefore, I have chosen to thank in the preface only those people who have
helped me in constructing this dissertation. A consequence of this choice is that I do not
mention some folks here who are very much more important to me than the present
dissertation.
I owe many thanks to Frans Stokman and Tom Snijders, two of my promoters. They
formulated the project “Parameter estimation in models of collective decision making” for
which I applied and was accepted. Although in my dissertation I estimated zero parameters in
models of collective decision making and although I mentioned collective decision making
only twice, Frans and Tom have had a big impact on this dissertation. I especially thank Tom
for his suggestions regarding the statistical tests of some of my hypotheses. I thank Frans in
particular for involving me in his discussions with Dave Willer, for being my guide at
conferences, and for introducing me to the research field of network exchange (chapter 6 and
7 of this dissertation) 42 months ago.
I also owe many thanks to my promoter Werner Raub and to Chris Snijders. Firstly,
Werner stimulated me to work out the relation between individuals’ risk aversion and
cooperation in social dilemmas. Later Werner, Chris, and I jointly prepared an experiment
and the articles that are the basis of chapters 3 to 5 of this dissertation. The joint discussions
with Werner and Chris were not only motivating, but secretly I even enjoyed them. I want to
thank Werner in particular for his involvement with the last phase of my writing and his
suggestions with respect to the introductory chapter. I appreciated and still appreciate the
cooperation with Chris very much. If everybody were like Chris then I would never again
work on an article alone.
Chapter 2 was the result of a traineeship during my study of mathematical psychology
at the University of Nijmegen. However, without Hein Fennema and Peter Wakker this
chapter would never have been accomplished.
For inspiring and instructive discussions with respect to the subjects in this
dissertation I thank Phil Bonacich, Dudley Girard, Károly Takács, Jeroen Weesie, and Dave
Willer. I owe many thanks to my wife Marieke van Onna for editing the dissertation, for
which she sacrificed days of her time. I thank Jan Kratzer and Ruud Brekelmans for their help
with constructing figures, Tamás Bartus for his help with running analyses in Stata, Graeme
Blake for correcting my English, and Rita Smaniotto for checking and improving parts of the
text.
Contents
1 Introduction 1
1.1 The importance of the micro-level in predicting and understanding macro-level
outcomes in the social sciences 3
1.2 The importance of studying the micro-level in order to understand macro-level
outcomes when macro-level predictions are accurate 8
1.3 The importance of studying the micro-level in order to predict macro-level
outcomes: the case of social dilemmas 15
2 Measuring the Utility of Losses by Means of the Tradeoff Method 19
2.1 Introduction 21
2.2 History 22
2.3 The tradeoff method 24
2.4 The experiment 27
2.5 Results 29
2.6 Discussion 34
3 Effects of Risk Preferences in Repeated Social Dilemmas: A Game-theoretical
Analysis and Evidence from Two Experiments 37
3.1 Introduction 39
3.2 Effects of risk preferences on cooperation in repeated social dilemmas:
theoretical background 41
3.3 Experiment 1 (Raub and Snijders, 1997) 45
3.4 Utility measurement 47
3.5 Experiment 2 51
3.6 Conclusion and discussion 61
Appendix 3.1: Simulation of other actor's strategy in repeated PDs 64
4 Effects of Individual Decision Theory Assumptions on Predictions of
Cooperation in Social Dilemmas 67
4.1 Introduction 69
4.2 The effects of loss aversion on cooperation in social dilemmas 69
4.3 Probability weighing 72
4.4 Discussion 73
Appendix 4.1: Proof of Theorem 75
5 The Effect of Nonlinear Utility on Behavior in Repeated Prisoner’s Dilemmas 77
5.1 Introduction 79
5.2 Theoretical background and previous empirical results 79
5.3 Hypotheses 84
5.4 Experiment 85
5.5 Results 89
5.6 Conclusion and discussion 94
Appendix 5.1: Estimation of utility function 98
Appendix 5.2: Simulation of other agent's strategy in repeated PDs 99
Contents
6 Bargaining in Exchange Networks 101
6.1 Introduction 103
6.2 Background 105
6.3 Models and hypotheses 114
6.4 Hypotheses 116
6.5 Data 119
6.6 Results 120
6.7 Conclusion and discussion 131
7 Two Representations of Negotiated Exchange: A Review and Comparison 137
7.1 Introduction 139
7.2 Exchange representations in economics and sociology 143
7.3 Relations between representations of negotiated exchanges 152
7.4 Consequences of representation inequivalence 159
7.5 Advantages of the non-reduced exchange representation 175
Appendix 7.1: Equivalence of equiresistance formulations 179
Appendix 7.2: Nash solution in bilateral monopoly situation (II) 180
8 Epilogue 183
8.1 Introduction 185
8.2 Measuring utility by means of the tradeoff method 186
8.3 The relation between actors’ utility and their behavior in PDs 187
8.4 Bargaining in exchange networks 189
8.5 Two representations of negotiated exchange: a review and comparison 191
8.6 Conclusions 192
References 195
Samenvatting (Summary in Dutch) 211
1
Introduction
2 Chapter 1
Introduction 3
1.1 The importance of the micro-level in predicting and understanding macro-level
outcomes in the social sciences
1.1.1 The importance of the micro-level
A central problem in sociology and economics is that of accounting for collective
phenomena. Collective phenomena that are addressed in sociology are numerous and diverse,
examples of which include suicide rates (Durkheim, 1951 [1897]), group solidarity (Hechter,
1987), revolutions (Coleman, 1990, Ch. 18), and the existence of norms (Coleman, 1990, Ch.
10-11). In economics, a central problem is to predict market prices. The present dissertation is
an attempt to contribute to our understanding of two other collective phenomena that receive
a lot of attention in social science research in general, and sociological research in particular:
outcomes of social dilemmas (chapters 3 to 5) and outcomes in exchange networks (chapters
6 and 7).
How do economists and sociologists attempt to explain collective phenomena? One
possible mode of explanation of collective phenomena entails the examination of processes
inherent to the system, involving its component parts or units at a level below that of the
system, known as the micro-level. Coleman (1990, p. 2) refers to this mode of explanation as
the internal analysis of system behavior. The internal analysis of the system can be
represented in a diagram containing three kinds of relations. The diagram is shown in Figure
1.1 (also in Coleman, 1994, pp. 167-168; Raub, 1984, 1.2; Lindenberg, 1992). Following
Coleman, in Figure 1.1 and in the remainder of the text the terminology macro-level is used
instead of the collective level.1 The three relations are: (1) Effects of macro- or systemic level
phenomena on orientations, preferences, information, and the set of feasible actions of actors
(which can be either individuals or organizations) on the micro-level. (2) Actor behavior on
the micro-level. (3) The combination or aggregation of these actions, in some institutional
structure, to bring about outcomes at the systemic or macro-level. Internal analysis of the
system is practised by both economists, in particular microeconomists (Kreps, 1990, p. 3),
and some sociologists, in particular those adhering to rational choice theory (Coleman, 1987,
p. 167; Wippler and Lindenberg, 1987; Lindenberg, 1992; Raub and Voss, 1981, 1.4).
Why do microeconomists and rational choice sociologists ascribe importance to the
micro-level in order to predict and understand collective or macro-level phenomena? In the
first chapter of his standard work Foundations of social theory Coleman (1990, pp. 3-5)
offers five theoretical arguments in favor of the internal analysis of system behavior. Most
relevant for the present discussion is his argument that “… an explanation based on internal
analysis of system behavior in terms of actions and orientations of lower-level units is likely
to be more stable and general than an explanation which remains at the system level”
(Coleman, 1990, p. 3)2. Coleman’s argument can also be reformulated more forcefully. A
relation between variables on the macro-level is mediated by actors’ behavior on the micro-
level. Therefore, if we do not understand processes on the micro-level (relation (2) in Figure
1.1) and the links between micro- and macro-level (relations (2) and (3)), then we have an
1 Although collective phenomena mainly occur at the macro-level, they can also occur at the level of
individuals. Therefore, Figure 1.1 and the terminology macro-level are somewhat misleading. However, I have
chosen to remain consistent with Coleman’s work and terminology because of its familiarity.
2 Coleman’s arguments in favor of the internal analysis of system behavior are discussed in more detail in
section 6.1 of this dissertation.
4 Chapter 1
System or
macro-level
1 3
2
Actor or micro-level
Figure 1.1: Internal analysis of system behavior represented by three kinds of relations.
incomplete understanding of why and how the relation between the macro variables
originated.
1.1.2. How much emphasis on the micro-level?
After Max Weber, Homans (1958, 1974) was among the first sociologists to propose the
basing of theoretical analyses of macro-level phenomena on a micro-level model of
individual behavior. According to Homans there are no general laws that are specifically
social, and only psychological general propositions are needed to explain human activity
(Willer, 1999, p. 6). Homans adopted these general propositions from behaviorism in learning
psychology, according to which the actions of an individual must be understood in terms of
the learning history of that individual. Homans’ approach can therefore be regarded as an
attempt to reduce sociology to psychology (Wippler and Lindenberg, 1987, p. 139).
Homans’ approach has been criticized for two reasons. Firstly, his approach fails in
explaining system behavior and its relation to actor behavior in the micro-level. Homans
asserts that social structures are a consequence of characteristics of individuals, and that they
are an incidental by-product of the everyday activities of individuals (Willer, 1999, p. 6).
However, much research, for example the network exchange research as discussed in
Chapter 6 and Chapter 7 of this dissertation, demonstrates that structures, independent of
characteristics of individuals, can have a large effect on both actor behavior at the micro-level
and outcomes at the macro-level. Secondly, the principal task of sociology lies in the
explanation of macro-level phenomena, not of the behavior of single actors. Focusing on the
behavior of actors at the micro-level can divert attention from the functioning of the system at
the macro-level (Coleman, 1990, p. 2). Therefore, Coleman and other rational choice
sociologists propose to keep the model of actor behavior at the micro-level as simple as
possible.3 They argue that psychological theories of action, such as behaviorism, are too
complex for the analysis of collective phenomena (Coleman, 1990, pp. 13-21). Rational
choice theory in sociology distinguishes itself from Homans’ approach, which places both
analytical and explanatory primacy at the micro-level, by placing analytical primacy at the
3 The phenomenon to be explained at the macro-level is not independent of the specifications and
simplifications of the micro-level model (Lindenberg, 1998). A consequence of simplifying the micro-level
model is that less aspects of the macro-level phenomenon can be explained. The simplest model assumptions
should therefore always be realistic enough to allow a description of the phenomenon to be explained.
Lindenberg (2001) calls this principle sufficient complexity. In order to avoid complicating the discussion in the
present chapter, I do not attend to this dependency and principle.