Analytical Sociology E-Book
66,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Serie: Wiley Series in Computational and Quantitative Social Science
- Sprache: Englisch
Demonstrates the power of the theoretical framework of analytical sociology in
explaining a large array of social phenomena
Analytical Sociology: Actions and Networks presents the most advanced theoretical discussion of analytical sociology, along with a unique set of examples on mechanism-
based sociology. Leading scholars apply the theoretical principles of analytical sociology
to understand how puzzling social and historical phenomena including crime, lynching,
witch-hunts, tax behaviours, Web-based social movement and communication,
restaurant reputation, job search and careers, social network homophily and instability, cooperation and trust are brought about by complex, multi-layered social mechanisms. The analyses presented in this book rely on a wide range of methods which include qualitative observations, advanced statistical techniques, complex network tools, refined simulation methods and creative experimental protocols.
This book ultimately demonstrates that sociology, like any other science, is at its best
when it dissects the mechanisms at work by means of rigorous model building and testing.
Analytical Sociology:
• Provides the most complete and up-to-date theoretical treatment of analytical sociology.
• Looks at a wide range of complex social phenomena within a single and unitary theoretical framework.
• Explores a variety of advanced methods to build and test theoretical models.
• Examines how both computational modelling and experiments can be used
to study the complex relation between norms, networks and social actions.
• Brings together research from leading global experts in the field in order to
present a unique set of examples on mechanism-based sociology.
Advanced graduate students and researchers working in sociology, methodology of social sciences, statistics, social networks analysis and computer simulation will benefit from this book.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 994
Veröffentlichungsjahr: 2014
Ähnliche
CONTENTS
Cover
Wiley Series Incomputational and Quantitative Social Science
Title Page
Copyright
Preface and Acknowledgments
About the Editor
List of Contributors
Introduction
Editor's Introduction to Chapter 1
Chapter 1: Data, Generative Models, and Mechanisms:
More
on the Principles of Analytical Sociology
1.1 Introduction
1.2 The Principles of Analytical Sociology
1.3 Clarity (P1)
1.4 Description (P2)
1.5 Generative Models (P3)
1.6 Structural Methodological Individualism (P4a)
1.7 Logics of Action (P4b)
1.8 Structural Interdependency (P4c)
1.9 Agent-Based Modeling (P5)
1.10 Back to Data (P6 and P7)
1.11 Concluding Remarks
1.12 How to Read This Book
References
Further Reading
Part I: Actions
Foundational Issues
Editor's Introduction to Chapter 2
Chapter 2: Analytical Sociology and Rational-Choice Theory
2.1 Rational-Choice Theory
2.2 Sociological Rational-Choice Theory
2.3 Analytical Sociology as a Meta-Theory
2.4 The Key Ideas of Analytical Sociology
2.5 The Puzzle
2.6 The Assumed Special Role of RCT
2.7 Conclusion
References
Further Reading
Crime and Voting
Editor's Introduction to Chapter 3
Chapter 3: Why Crime Happens: A Situational Action Theory
3.1 Situational Action Theory
3.2 Explaining Crime
3.3 The Situational Model
3.4 The Situational Process
3.5 The Social Model
3.6 Integrating the Social and Situational Models
3.7 Testing SAT
3.8 Explaining Crime Concentrations (Hot Spots)
3.9 Coda
References
Further Reading
Editor's Introduction to Chapter 4
Chapter 4: Frames, Scripts, and Variable Rationality: An Integrative Theory of Action
4.1 Introduction
4.2 The Model of Frame Selection (MFS)
4.3 Hypotheses and Previous Applications
4.4 An Exemplary Application Using Survey Data: Explaining Voter Participation
4.5 Applying the MFS to Study Social Dynamics
4.6 Conclusion
References
Further Reading
Historical Violence
Editor's Introduction to Chapter 5
Chapter 5: Analytical Sociology and Quantitative Narrative Analysis: Explaining Lynchings in Georgia (1875–1930)
5.1 Strange Fruits on Southern Trees
5.2 Analytical Sociology
5.3 Quantitative Narrative Analysis (QNA)
5.4 Of Sequences
5.5 Of Time and Space
5.6 Conclusions
Acknowledgments
References
Further Reading
Editor's Introduction to Chapter 6
Chapter 6: Identity and Opportunity in Early Modern Politics: How job Vacancies Induced Witch Persecutions in Scotland, 1563–1736
6.1 Introduction
6.2 Theories About Witches and Research on State Making
6.3 Towards a Theory of Persecution
6.4 Witch-Hunting in Scotland
6.5 Findings
6.6 Discussion
Acknowledgements
References
Further Reading
Trust and Cooperation
Editor's Introduction to Chapter 7
Chapter 7: Mechanisms of Cooperation
7.1 Introduction
7.2 Cooperation Problems in Dyadic Settings
7.3 Cooperation Problems Involving More Than Two Actors
7.4 Discussion and Concluding Remarks
References
Further Reading
Editor's Introduction to Chapter 8
Baldassarri's Preface to Chapter 8
Chapter 8: The Impact of Elections on Cooperation: Evidence from a Lab-in-the-Field Experiment in Uganda
*
8.1 Theoretical Framework and Hypotheses
8.2 Research Site, Sampling, and Experimental Design
8.3 Research Site
8.4 Sampling and Data Collection
8.5 Experimental Design
8.6 Experimental Findings
8.7 Monitors' Sanctioning Behavior
8.8 Discussion of the Experimental Part
8.9 Observational Data
8.10 Comparing Behavior in the Experiment and Real Life
8.11 Conclusion
Supporting Information
Appendix 8.A
Acknowledgments
References
Further Reading
Part II: Networks
Collective Action
Editor's Introduction to Chapter 9
Chapter 9: Social Networks and Agent-Based Modelling
9.1 Social Network Properties
9.2 Network Construction Techniques
9.3 Networks as Pipes: A Basic Demonstration
9.4 Discussion
References
Further Reading
Editor's Introduction to Chapter 10
Chapter 10: Online Networks and the Diffusion of Protest
10.1 Diffusion Dynamics
10.2 Thresholds and Critical Mass
10.3 Networks and Social Influence
10.4 Conclusion: Digital Data and Analytical Sociology
References
Further Reading
Homophily and Status Hierarchies
Editor's Introduction to Chapter 11
Chapter 11: Liability to Rupture: Multiple Mechanisms and Subgroup Formation. An Exploratory Theoretical Study
11.1 Introduction
11.2 A Formal Framework
11.3 Balance Theory
11.4 Homophily (H-theory)
11.5 Baseline Structures
11.6 Developing a Dynamic Mechanism for Balance Theory
11.7 Developing a Dynamic Mechanism for H-theory
11.8 The Dynamic Interaction of Balance and H-theories
11.9 Conclusions
Appendix 11.A: Micro–Macro Inferences and Scale
References
Further Reading
Editor's Introduction to Chapter 12
Chapter 12: Network Size and Network Homophily: Same-Sex Friendships in 595 Scandinavian Schools
12.1 Introduction
12.2 Theoretical Considerations
12.3 Empirical Application: Same-Sex Ties in School Classes
12.4 Results
12.5 Conclusion
References
Further Reading
Editor's Introduction to Chapter 13
Chapter 13: Status and Participation in Online Task Groups: An Agent-Based Model
13.1 Introduction
13.2 Previous Models
13.3 E-State Structuralism: A Very Brief Review with an Add-On
13.4 Case Study: Strategies and Discussions in Massively Multi-Player Online Games
13.5 Analysis of the Model
13.6 Empirical Test/Validation of the Model
13.7 Conclusions
References
Further Reading
Labour Market Inequality
Editor's Introduction to Chapter 14
Chapter 14: Turbulent Careers: Social Networks, Employer Hiring Preferences, and job Instability
14.1 Introduction
14.2 Background
14.3 Networks
14.4 Methods
14.5 Results
14.6 Summary and Conclusions
Technical Appendix 14.A: Detailed Description of jobMatch Simulation Model
Acknowledgments
References
Further Reading
Editor’s Introduction to Chapter 15
Chapter 15: Employer Networks, Priming, and Discrimination in Hiring: An Experiment
15.1 Introduction
15.2 Method
15.3 Results
15.4 Discussion
Acknowledgments
References
Further Reading
Organization Similarity
Editor’s Introduction to Chapter 16
Chapter 16: The Duality of Organizations and Audiences
16.1 Introduction
16.2 Similarity and the Duality of Organizations and Their Audiences
16.3 Organizational Similarity, Audiences, and Arguments for Extending Structural Equivalence
16.4 A Representation for Dual Similarity of Organizations and Their Audiences
16.5 Empirical Illustration: The Duality of Restaurants and Their Reviewers
16.6 Similarity as a Basis for Prediction: Validating the Model
16.7 Discussion, Implications, and Limitations
16.8 Connections to Analytical Sociology
References
Further Reading
Coda
Chapter 17: Problem Shift in Sociology: Mechanisms, Generic Instruments, and Fractals
Index
End User License Agreement
List of Tables
Table 3.1
Table 3.2
Table 4.1
Table 4.2
Table 4.3
Table 5.1
Table 5.2
Table 6.1
Table 8.1
Table 8.2
Table 8.A.1
Table 8.A.2
Table 8.A.3
Table 9.1
Table 9.2
Table 9.3
Table 9.4
Table 9.5
Table 12.1
Table 12.2
Table 12.3
Table 13.1
Table 13.2
Table 14.1
Table 14.2
Table 14.3
Table 14.4
Table 14.A.1
Table 14.A.2
Table 15.1
Table 15.2
Table 15.3
Table 15.4
Table 15.5
Table 16.1
Table 16.2
List of Illustrations
Figure 1.1
Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 4.1
Figure 5.1
Figure 5.2
Figure 5.3
Figure 5.4
Figure 5.5
Figure 6.1
Figure 6.2
Figure 7.1
Figure 7.2
Figure 8.1
Figure 8.2
Figure 8.3
Figure 9.1
Figure 9.2
Figure 9.3
Figure 9.4
Figure 9.5
Figure 9.6
Figure 10.1
Figure 10.2
Figure 10.3
Figure 10.4
Figure 10.5
Figure 10.6
Figure 10.7
Figure 12.1
Figure 12.2
Figure 12.3
Figure 12.4
Figure 13.1
Figure 13.2
Figure 13.3
Figure 13.4
Figure 13.5
Figure 13.6
Figure 14.1
Figure 14.2
Figure 14.3
Figure 14.4
Figure 14.5
Figure 14.6
Figure 15.1
Figure 15.2
Figure 15.3
Figure 15.4
Figure 15.5
Figure 16.1
Figure 16.2
Guide
Cover
Table of Contents
Preface and Acknowledgments
Introduction
Data, Generative Models, and Mechanisms: More on the Principles of Analytical Sociology
Actions
Pages
ii
iii
iv
xiii
xiv
xv
xvi
xvii
xviii
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
Wiley Series in Computational and Quantitative Social Science
Embracing a spectrum from theoretical foundations to real world applications, the Wiley Series in Computational and Quantitative Social Science (CQSS) publishes titles ranging from high level student texts, explanation and dissemination of technology and good practice, through to interesting and important research that is immediately relevant to social/scientific development or practice.
Other Titles in the Series
Guanglei Hong – Causality in a Social World: Moderation, Mediation and Spill-over
Patrick Doreian, Vladimir Batagelj, Anuska Ferligoj, Natasa Kejzar – Understanding Large
Temporal Networks and Spatial Networks: Exploration, Pattern Searching, Visualization and Network Evolution
Gianluca Manzo (ed.) – Analytical Sociology: Actions and Networks
Rense Corten – Computational Approaches to Studying the Co-evolution of Networks and Behavior in Social Dilemmas
Danny Dorling – The Visualisation of Spatial Social Structure
Analytical Sociology
Actions and Networks
Edited by
Gianluca Manzo
GEMASS, Centre National de la Recherche Scientifique (CNRS) and University of Paris–Sorbonne, France
This edition first published 2014
© 2014 John Wiley & Sons, Ltd
Registered office
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom
For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com.
The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Library of Congress Cataloging-in-Publication Data
Analytical sociology: actions and networks / editor, Gianluca Manzo.
pages cm. – (Wiley series in computational and quantitative social science)
ncludes bibliographical references and index.
ISBN 978-1-119-94038-8 (hardback)
1. Sociology. 2. Sociology–Research. 3. Sociology–Philosophy. I. Manzo, Gianluca, editor of compilation.
HM585.A52593 2014
301.01–dc23
2013042818
A catalogue record for this book is available from the British Library.
ISBN: 978-1-119-94038-8
Preface and Acknowledgments
Analytical Sociology: Actions and Networks contains 16 essays that discuss the principles of analytical sociology and apply them to the analysis of a wide range of macro-level dynamics and microscopic behaviors, such as crime, voting, lynching, witchcraft, trust and cooperation, collective action, homophily, status hierarchies, labor market inequality, and organization patterns. The book is conceived as a suite of variations on a common theme. In my opening chapter, I overtly address the question of the identity of analytical sociology in the sense of its singularity and uniqueness within contemporary sociology. I venture to propose a specific understanding of analytical sociology. The remaining chapters are conceived as a suite of variations on this understanding. The purpose behind this architecture is neither to “speak for analytical sociology” nor to polish and police its present boundaries. More modestly, the book is an endeavor to deepen our understanding of what analytical sociology may become in the future and to enhance the exchange between analytical sociology and other theoretical and methodological approaches. In short, this collection of essays undertakes the difficult operation of arguing in favor of analytical sociology by challenging its own theoretical and methodological principles.
Analytical Sociology: Actions and Networks originates from a conference that I organized at the University of Paris–Sorbonne on June 9 and 10, 2011. This was the Fourth Conference of the European Network for Analytical Sociologists (ENAS) – since then renamed the International Network of Analytical Sociologists (INAS). But the book's final content only partially overlaps with the conference. This is for two reasons. First, among the 30 Paris presenters, only those engaged in the analysis of specific world-related explananda were asked to submit a chapter. The rationale behind this choice was to help correct the imbalance between meta-theoretical discussions and empirically oriented analyses often observed in debates on analytical sociology. Second, some submitted chapters were not finally accepted, or did not arrive in time; others developed the Paris presentation along different lines; and two papers were simply unrelated to the 2011 ENAS meeting.
My warmest gratitude goes to the book's contributors. They have made admirable efforts to relate their own research agendas to their understanding of analytical sociology. Their commitment to the book project should be praised and appreciated. They have agreed to revise their chapters several times and patiently replied to all my requests for clarification and precision. I have greatly enjoyed the long back-and-forth with each of them, and during which I have learned a great deal. I hope that the final product will be equally enjoyable for each contributor.
I wish also to thank Olivier Galland, the Director of the Groupe d'Etude des Méthodes de l'Analyse Sociologique de la Sorbonne (GEMASS), for providing me with the funds necessary to organize the 2011 ENAS conference in Paris. Special thanks go to Heather Kay and Debbie Jupe at Wiley, who believed in my editorial project from the beginning; to Richard Davies, my project editor, who professionally and patiently followed the book's production process in its entirety; to Prachi Sinha Sahay, Sharib Asrar, Ajay Gupta and Neville Hankins, who wonderfully assisted me during the book's typesetting and copyediting stages and to Adrian Belton for revising my English. Finally, it should be acknowledged that preparatory work for this book was also supported by the “ERC Advanced Grant on Analytical Sociology” and the RJ program “Segregation: Micro mechanisms and macro-level dynamics” both currently run at the Institute for Futures Studies (Stockholm) under the direction of Peter Hedström.
Last but not least, my wife's love was essential for completion of the book. I sometimes feel that there is a negative correlation between the time and concentration needed for writing and the energy and attention that one is able to give to one's family. My wife intelligently accepts this unfairness and continuously works to counteract its possible undesirable effects. To her and to our splendid Eléonore and Mathilde, this book is dedicated.
Gianluca Manzo 2013, Paris
About the Editor
Gianluca Manzo earned a PhD in Social Sciences from the University of Paris-Sorbonne and a PhD in Epistemology and Methodology of Social Sciences (2006) from the University of Trento (Italy). He is a permanent research fellow in sociology at the Centre National de la Recherche Scientifique (CNRS) and holds a teaching appointment at the University of Paris-Sorbonne, where he teaches statistics and simulation methods. Gianluca Manzo is an international research affiliate at the Institute for Futures Studies (Stockholm) and has served as visiting scholar and professor at several universities, including Columbia University and the University of Oxford. He investigates interaction-based reinforcing mechanisms in connection with educational inequalities, inequality subjective perceptions, status hierarchies, ethnic boundaries, and diffusion of innovations. He is also concerned with mechanism-based explanations, the theory of action, and comparative advantages of statistical and computational modeling.
List of Contributors
Peter Abell, Department of Management, London School of Economics, UK.
Delia Baldassarri, Department of Sociology, New York University, USA.
Davide Barrera, Department of Culture, Politics, and Society, and Collegio Carlo Alberto, University of Turin, Italy and ICS/Department of Sociology, Utrecht University, Netherlands.
Javier Borge-Holthoefer, Institute for Biocomputation and Physics of Complex Systems, University of Zaragoza, Spain.
Giangiacomo Bravo, Department of Social Studies, Linnaeus University, Sweden, and Collegio Carlo Alberto, Italy.
Marco Castellani,Department of Economics and Management, University of Brescia, Italy.
Christine Fountain, Department of Sociology and Anthropology, Fordham University, USA.
Roberto Franzosi, Department of Sociology, Emory University, USA.
Simone Gabbriellini, Department of Informatics: Science & Engineering, University of Bologna, Italy.
Sandra González-Bailón, Annenberg School for Communication, University of Pennsylvania, USA.
Guy Grossman, Political Science Department, University of Pennsylvania, USA.
Thomas Grund, Institute for Futures Studies, Sweden.
Peter Hedström, Institute for Future Studies, Sweden.
Balázs Kovács, Institute of Management, University of Lugano, Switzerland.
Clemens Kroneberg, Institute of Sociology and Social Psychology, University of Cologne, Germany.
Gianluca Manzo, GEMASS, Centre National de la Recherche Scientifique (CNRS) and University of Paris–Sorbonne, France.
Anna Mitschele, Department of Sociology, Columbia University, USA.
Yamir Moreno, Institute for Biocomputation and Physics of Complex Systems, University of Zaragoza, Spain.
Meredith Rolfe, Department of Management, London School of Economics and Political Science, UK.
Flaminio Squazzoni, Department of Economics and Management, University of Brescia, Italy.
Katherine Stovel, Department of Sociology, University of Washington, USA.
Károly Takács, MTA TK “Lendület” Research Center for Educational and Network Studies (RECENS), Hungarian Academy of Sciences and Institute of Sociology and Social Policy, Corvinus University of Budapest, Hungary.
Per-Olof H. Wikström, Institute of Criminology, University of Cambridge, UK.
Petri Ylikoski, Department of Social Research, University of Helsinki, Finland.
Introduction
Editor's Introduction to Chapter 1
The computer is even more revolutionary as an idea, than it is as a practical device that alters society – and we all know how much it has changed our lives. Why do I say this? Well, the computer changes epistemology, it changes the meaning of “to understand”. To me, you understand something only if you can program it. (You, not someone else!) Otherwise you don't really understand it, you only think you understand it.
(Gregory Chaitin 2006 [2005] Meta Math! The Quest for Omega, Vintage Books, p. xiii)
As the book's subtitle suggests, Hedström's Dissecting the Social had the fundamental goal of clarifying the theoretical and methodological principles underlying the research program of analytical sociology. Despite Hedström's admirable effort, many observers' reactions to analytical sociology over the last 10 years suggest that the intellectual project behind analytical sociology is still often misunderstood. Although the interest in analytical sociology is growing rapidly, criticisms are also recurrent. In this chapter I start with the idea that the current internal heterogeneity of analytical sociology and the complexity of its theoretical and methodological proposals help explain both its success and the criticisms that it receives. I then take these criticisms seriously and let them guide me in an attempt to remedy the most recurrent misunderstandings of analytical sociology's principles. To this end, I shall defend a specific understanding of analytical sociology as a set of principles defining a research program in the sense of Lakatos. The essay discusses each of these principles in detail, as well as their interdependence. It argues that the interdependence among the principles is the fundamental feature of analytical sociology's distinctiveness within contemporary sociology. The argument is that this interdependence arises from a specific understanding of the concept of mechanism. According to this understanding, a mechanism-based explanation amounts to a reverse engineering operation: an observation is explained only when it can be recreated. Once this is posited, it follows that some methods are more appropriate than others for designing models of mechanisms (i.e., “generative models”) and studying their high-level consequences. The chapter explains the basic generic elements composing generative models and why the technical foundations of agent-based computational modeling put this method at the core of analytical sociology. The chapter clarifies that accomplishment of this research program depends on a clear division of labor among quantitative and qualitative scholars, formal modelers, and experimentalists. When all these elements are brought together, analytical sociology clearly stands apart as an empirically oriented, experimentally and computationally based, macro-sociology with clearly explicated and empirically grounded dynamic micro- and network-level foundations. The chapter concludes by suggesting that Analytical Sociology: Actions and Networks should be read as a suite of variations on a common theme, this theme being the kind of analytical sociology discussed in this chapter. The book's aim is to accumulate elements that may foster the further development of this kind of analytical sociology. Analytical Sociology: Actions and Networks is not intended to be about the past or the present of analytical sociology: it points to (one of) its possible future(s).
1Data, Generative Models, and Mechanisms: More on the Principles of Analytical Sociology
Gianluca Manzo
GEMASS, Centre National de la Recherche Scientifique (CNRS) and University of Paris–Sorbonne, France
1.1 Introduction
The contemporary meaning of the term “analytical sociology” started to circulate informally through European academic space in the mid-1990s (see Manzo, 2010: 138). Still absent from the seminal collection of essays by Hedström and Swedberg (1998a) on social mechanisms, the expression “analytical sociology” officially entered the sociological vocabulary with Hedström's Dissecting the Social (Hedström, 2005) to denote the sociological perspective that seeks systematically to formulate and empirically test micro-founded, mechanism-based explanations of complex macro-level patterns and dynamics.
Despite the considerable efforts at theoretical clarification made by Hedström (2005), and despite the conceptual richness of the essays subsequently collected by Hedström and Bearman (2009a) and by Demeulenaere (2011a), doubts have been raised concerning the need for analytical sociology and its originality. Qualitative-oriented symbolic interactionists (see Sawyer, 2007; 2011), pragmatists (see Abbott, 2007a; Gross, 2009), cultural sociologists (Lizardo, 2012; Santoro, 2012), rational-choice theorists (Opp, 2007; 2013a), as well as philosophers of social sciences like Bunge (2007) or Little (2012a), have all criticized analytical sociology's understanding of mechanism-based thinking as based on narrow and unoriginal theoretical foundations.
This is an interesting puzzle for (historically oriented) sociologists of knowledge. Indeed, when one considers the arguments brought against analytical sociology (see, in particular, Lizardo, 2012), it seems as if some authoritative scholars have artfully constructed an unoriginal sociological approach with an uncanny ability to mobilize a large stock of institutional and cognitive resources and to attract a considerable amount of attention, including that of scholars who feel it necessary to attack this new intellectual construct and denounce its emptiness, thereby opening the eyes of its blind followers.
At first glance, this puzzle can be resolved by positing that both the construction of analytical sociology and the critical reactions against it simply result from a struggle for academic identity in which false problems and transitory novelties arise because actors intentionally emphasize minor points while ignoring the fundamental ones. I prefer to take seriously, and believe in the intellectual honesty of, both the advocates and critics of analytical sociology. It well may be that the diversity and complexity of the cognitive content of analytical sociology explain both the attention received by the approach and the objections brought against it.
First, there are diverse understandings as to the purpose of analytical sociology. Some maintain that the task of analytical sociology is to clarify what a good sociological explanation is in general, thus endorsing a strong normative stance which ultimately decrees what is scientific and what is not (see Demeulenaere, 2011b: 1). This position (with reason) irritates some observers (see Little, 2012a, and, partly, Gross, 2013). Others reject this imperialistic attitude and claim that analytical sociology “only provides a ‘syntax’ for explanation: that is to say, a set of rules on how hypotheses about mechanisms underlying the regularities of social life can be theoretically designed and empirically tested” (see Manzo, 2010: 162; see also Hedström and Ylikoski, this volume) without implying that those who do not conform with this “syntax” are ipso facto mistaken. Even more liberally, others claim that analytical sociology is only one of the possible ways to conduct “good” sociology, thus implying that the quest for mechanism-based explanations is not necessarily to be considered the priority (see Bearman, 2012).
Analytical sociology is also diverse with respect to some fundamental theoretical and methodological choices. Not all advocates of analytical sociology make the same assessment of the role that rational choice theory should play in model building (see Hedström and Ylikoski, this volume; Manzo, 2013b). From a methodological point of view, some of them distrust quantification and formalization (see Boudon, 2012; Elster, 2007; 2009a), whereas others consider the formal modeling of a mechanism to be a crucial research step (see Hedström and Bearman, 2009b; Hedström, 2005: Ch. 6; Manzo, 2012a).
This diversity has an advantage. Different scholars with different theoretical and methodological orientations can become interested and involved in analytical sociology. This is the success part of the story. The advantage comes with a cost, however. The heterogeneity of analytical sociology dilutes and obscures the perception of its originality. This facilitates the task of skeptical observers.
The complexity of the cognitive content of analytical sociology is likely to generate a similar twofold effect on its reception. From its very beginning, in fact, this intellectual movement has relied on a multi-dimensional combination of conceptual, epistemological, ontological, and methodological elements (see Manzo, 2010). As the topics covered by Hedström's Dissecting the Social show, analytical sociology requires us to reflect at the same time on the principles of scientific explanation, the meaning of methodological individualism, the content of the theory of action, the role of social networks, the problem of the micro–macro transition, and the advantages and shortcomings of statistical methods and formal modeling for the empirical testing of sociological theories.
These are difficult questions that bear upon some of the most fundamental aspects of social inquiry. They have long occupied philosophers of social sciences and social scientists. It is therefore not surprising that a large number of scholars have become interested in analytical sociology. This approach is seen by many as a new intellectual space in which old questions can be again addressed and hopefully developed further. At the same time, given the fundamental importance of these questions, the answers proposed by analytical sociology are likely to provoke controversies. This explains the (strong) critical reactions against the approach: in particular against some of its crucial assertions on methodological individualism and rational-choice theory (see Little, 2012a; Opp, 2013a).
The complexity of the analytical sociology research program also helps explain the criticism that it lacks originality. For assessment of analytical sociology's novelty requires the effort to consider the entire set of questions addressed and the coherence of the entire set of replies provided. What matters is the overall picture. Many of the theoretical and methodological proposals of analytical sociology have deep roots in sociology, and several areas of contemporary sociology also focus on some of them. However, the originality of analytical sociology stems from its integration of these elements into a unitary meta-theoretical framework (see Manzo, 2011a). By contrast, as the writings of the critics show, the discontents with analytical sociology systematically focus on one or only some of the components of analytical sociology's research program. They consequently neglect the source of analytical sociology's novelty: the interdependence among the elements. This is not the critics' fault. Analytical sociology is made up of a complex web of conceptual, epistemological, ontological, and methodological choices, some of which do not go undisputed even by those who are supposed to help develop the approach. Hence, it should not come as a surprise that the overall picture is still missed by many observers.
Some critics have considered this line of reasoning to be a purely rhetorical strategy whereby analytical sociology's advocates –“chameleon like,” to use the expression by Lizardo (2012) – artificially mobilize new elements whenever a criticism is made. In my opinion, this interpretation is wrong. Like any research perspective that has reached a minimum level of maturity, analytical sociology is a complex intellectual construct. To cite only a few examples, Gross (2009), Goldthorpe (1998), or Back and co-authors (2012) all depict pragmatism-oriented sociology, rational-choice theory, and cultural sociology as highly heterogeneous, multi-faceted research orientations with several variants. Critical discussion of these approaches requires knowledge of their internal complexity. The same holds for analytical sociology.
Analytical Sociology: Actions and Networks has two goals. On the one hand, it aims to advance the discussion on the two theoretical pillars of analytical sociology, that is, “actions” and “networks.” My concern here is to remedy the recurrent misunderstanding which views analytical sociology as a reductionist form of methodological individualism and another instance of rational-choice-based sociology. From different points of view, and with different emphases, the 15 following essays all contribute to demonstrating that analytical sociology is all about the complex interplay between “actions” and “networks” (and social structures, more generally). On the other hand, Analytical Sociology: Actions and Networks aims to develop the cognitive content of analytical sociology further. It does so by focusing on one specific understanding of analytical sociology's research program. The present chapter conducts detailed discussion on this variant of analytical sociology, while the remaining essays provide specific theoretical and methodological insights that help to develop and/or challenge the conception of analytical sociology proposed here. Because one of the objections brought against analytical sociology, sometimes by its own proponents (see Bearman, 2012; Manzo, 2011a), is that programmatic statements still tend to outweigh their empirical application, virtually all the chapters contribute to the discussion on analytical sociology by studying specific empirical phenomena. This is also the case of the present essay, whose main arguments represent the meta-theoretical counterpart to empirical analyses presented elsewhere (see in particular Manzo, 2013a).
The chapter is organized as follows. In the next section, I present the set of principles constituting a particular variant of analytical sociology (Section 1.2). Each principle will then be detailed in one of the following eight sections (Sections 1.3–1.10). I summarize the main arguments in the concluding remarks section, while the final section on how to read this book gives more details about the book's orientation and content.
1.2 The Principles of Analytical Sociology
Commentators on analytical sociology focus extensively on the epistemological features of mechanism-based explanations (compared to other types of explanation), on the concept of mechanism, on the meaning of methodological individualism, or on the theory of action defended by analytical sociologists (among the most detailed analyses, see Abbott, 2007a; Gross, 2009; Little, 2012a; Opp, 2013a). By contrast, analytical sociology's methodological proposals have been subject to only limited discussion, which has essentially focused on analytical sociology's supposedly mistaken dismissal of regression-based methods (see Opp, 2007) and on the over-importance given to simulation methods (see Lucchini, 2007: 236–240; Lucchini, 2008: 9–12; Winship and Morgan, 2007: 233, note 10).
This imbalance is problematic because analytical sociology is in fact a set of research guidelines for both theoretical model building and empirical model testing in sociology. Hence, the meaning and the scope of analytical sociology can only be appreciated if the approach is understood as the intersection between one set of principles concerning the construction of explanatory theoretical models and another set of principles referring to the empirical validation of those models.
Without a doubt, this characterization is insufficient to set analytical sociology apart from other research traditions that also seek to devise conceptual models and to prove their empirical appropriateness. To a large extent, this is the purpose of all scientific research. The specificity of analytical sociology should thus be sought in the distinctive way in which its model-building and testing practices are conceived and concretely organized. I suggest that analytical sociology's uniqueness within contemporary sociology can only be fully appreciated if the following combination of principles (hereafter, P) is considered (for a graphic illustration, see Figure 1.1):1
P1: use concepts that are as clear and precise as possible to describe both the facts to be explained and the explanatory hypotheses/facts mobilized to explain them, while avoiding all linguistic obscurity and convolutedness;
P2: mobilize the best quantitative and qualitative empirical information available and use the technical tools best suited to describing the facts to be explained;
P3: in order to explain the social outcome(s) described, first formulate a “generative model,” that is, a model of a (set of) mechanism(s), where a mechanism is a set of entities and activities likely to trigger a sequence of events (i.e., a process) likely to bring about the outcome(s);
P4: in order to formulate the “generative model,” provide a realistic description of the relevant micro-level entities (P4a) and activities (P4b) assumed to be at work, as well as of the structural interdependencies (P4c) in which these entities are embedded and their activities unfold;
P5: in order rigorously to assess the internal consistency of the “generative model” and to determine its high-level consequences, translate the “generative model” into an agent-based computational model;
P6: in order to assess the generative sufficiency of the mechanisms postulated, compare the agent-based computational model's high-level consequences with the empirical description of the facts to be explained;
P7: in order to prove that the hypothesized micro- and network-level assumptions are not only generative sufficient but also empirically grounded, inject as much individual- and relational-level quantitative, qualitative, and/or experimental data as possible into the agent-based computational model and reanalyze its behavior and high-level consequences.
Figure 1.1 Stylized ideal–typical research cycle underlying analytical sociology.
If one considers that the facts of primary interest to analytical sociology are cross-sectional population-level patterns and their temporal trends (see Hedström, 2005: 67), then P1–P7 turn analytical sociology into an empirically oriented, experimentally and computationally based, macro-sociology with clearly explicated and empirically grounded dynamic micro- and network-level foundations.
Before I discuss each principle in detail, let me clarify how, in my opinion, these principles should be understood, and what we may gain from conceiving analytical sociology in this axiomatic form.
In regard to the meanings of P1–P7, it would be a mistake to interpret them as a set of universal normative imperatives. Figure 1.1 should not be understood as describing a rigid sequence of research steps that must necessarily be followed. Sometimes a researcher does not have the time, resources, and/or cognitive skills to meet the requirements contained in the seven principles. Sometimes, the researcher may refer to the results of previous studies of relevance to one (or some) of the research step(s), thus directly implementing only a subset of the operations proposed. Hence, P1–P7 should be regarded as a set of logically organized guesses as to the fruitfulness of a specific list of theoretical and methodological options. This is a set of guesses that should be borne in mind even if it is not possible or not necessary to perform all the requisite operations within a given piece of research. This is the sense of what Lakatos (1972: 132) considered a “research program,” that is, a set of “methodological rules: some tell us what paths of research to avoid (negative heuristics), and others what paths to pursue (positive heuristics).”2
Although specific, this conception of analytical sociology as a research program defined by P1–P7 has, in my opinion, three general advantages.
First, listing the principles sequentially, from the most general (P1) to the most specific (P7), helps to assess analytical sociology's uniqueness within contemporary analytical sociology. Indeed, as the number of principles considered increases, their combination makes it increasingly difficult to find one sociological perspective defined by the same combination of elements. The initial apparent overlap between analytical sociology and the rest of sociology thus tends progressively to disappear (on this point, see also the concluding remarks section).
Second, P1–P7 allow the better mapping of analytical sociology's internal heterogeneity. The main dividing line seems to be between those who accept the entire set of principles and those who restrict analytical sociology to P1–P4, plus the application of P7 without formal modeling, thus rejecting the idea that formal modeling is necessary to prove that there is a real connection between the explanans and the explanandum. In this regard, the description by Hedström (2005: 143–144) or Hedström and Bearman (2009b: 16) of analytical sociology's core research strategy differs markedly, for instance, from the non-formalized but deep explanatory analyses contained in Gambetta (2009).
Third, P1–P7 help to visualize why analytical sociology needs its own internal heterogeneity, other theoretical and methodological sociological perspectives, as well as specialties lying outside sociology. Going from P1 to P7 is extremely demanding in terms of time and cognitive resources. Even the best equipped scholar may be unable to fulfill all the requirements contained in the seven principles in a single piece of research. Thus, P1–P7 provide guidelines with which to locate potential collaborators within and outside sociology, and they suggest research areas that can help with developing some or other item on the analytical sociology research program.
This is the spirit that animated the selection of essays collected in Analytical Sociology: Action and Networks. While only a few chapters approximate the full research cycle depicted by Figure 1.1 (see the contributions by Gabbriellini, Grund, and Fountain and Stovel), all of them show how analytical sociology communicates with, and benefits from, other research traditions – like game theory, social network analysis, cognitive psychology, or behavioral economics – and how studies implementing only some of P1–P7 at a given point in time may help create the conditions for complete application of the research program in the long run. It is this conception of analytical sociology as a constantly evolving web of elements that we may want to pursue and develop further. Analytical Sociology: Action and Networks modestly seeks to contribute to this endeavor.
1.3 Clarity (P1)
Within analytical sociology, P1 – the quest for clarity and precision in the definition of concepts and in writing style – has evident philosophical roots (see Hedström, 2008: 331–302). In particular, it stems from one of the axioms of analytic philosophy that the ambiguity of natural language is responsible for many conceptual problems and misleading observations (for a thorough survey of analytic philosophy, see Glock, 2008). From the point of view of an empirically oriented discipline, this implies that both the concepts describing the facts to be explained and the facts mobilized to explain them must be formulated in clear and simple terms. Otherwise, analytical sociologists argue, the connections among events are difficult to see and the empirical testing of competing theoretical hypotheses is difficult to perform.
Building on Pareto's distinction between a theory's ideological utility and its empirical descriptive accuracy, Boudon (2002: 375, emphasis added) noted: “A false and useful theory is often perceived as true, as long as its falsity is not too visible. If in addition it is obscure, it may even be perceived as profound.” Analytical sociology's P1 aims to avoid this undesirable cognitive effect. P1 relies on the conviction that the complexity of social reality does not require linguistic complexity to be described. Analytical sociology thus rejects the equation between linguistic intricacy and intellectual profundity.
The macro-consequences of linguistic convolutedness have been specified by Sperber (2010). What he labels the “Guru Effect” corresponds to a causal chain that can be summarized as follows. Interpreting linguistically complex and convoluted sets of sentences is demanding in terms of cognitive effort. When, despite the efforts made, the reader is still unsure about the meaning of the argument, s/he looks for external cues to adjudicate on it. External signs of academic authority and reputation often serve as such cues. The larger the author's stock of such signs, the more likely it becomes that the reader will conclude that his/her lack of understanding reflects the profundity of the author's thought rather than author's lack of clarity. This belief may be reinforced by the interdependence of actors' beliefs. In search of external cues, the reader may look at the opinions of other readers, who, under pressure of the same cognitive mechanism, will tend to endorse the same belief as the focal reader's. Actors' beliefs thus dynamically reinforce each other. The reader's confession to others of his/her lack of understanding puts the reader in a potentially embarrassing situation which generates a “spiral of the silence” that may fuel an explosion of intellectual credibility for authors and articles.
There are consequently good reasons to pursue conceptual and linguistic clarity. Indeed, it is likely that linguistic convolutedness increases the probability that undesirable intellectual dynamics – in which hermeneutic problems become more central than the analysis of specific empirical facts – will arise, with a consequent waste of cognitive resources: namely, the resources needed to eliminate false debates and unjustified academic authorities from the academic market of ideas.
That said, P1 is certainly not sufficient on its own to confer originality on analytical sociology. Elster (2007: 455), for instance, considers the “near-obsessive concern with clarity and explicitness” to be the distinctive feature of “the analytical turn” that he sees at work in the social sciences at large. Thus, one may agree with those discontents of analytical sociology who have claimed that “clarity and precision”
is certainly sufficient to tell AS apart from the mass of sociological research that is unclear and imprecise (or from bad translations of French theory), but it certainly does not work well to make it different from the equally large mass of sociological research that is in fact clear and precise.
(Lizardo, 2012: 7)
However, the problem with this objection is that it isolates the clarity requirement from a larger set of principles. As I have argued above, it is the combination of these principles that matters when one wants to assess the intellectual distinctiveness of analytical sociology.
1.4 Description (P2)
When the explanandum is formulated in clear and explicit conceptual terms, it is easier to find appropriate empirical indicators for it. This facilitates the application of the second key principle of analytical sociology: that the facts to be explained should be precisely identified by mobilizing the best empirical information available and by using the technical tools best suited to describing the data.
In this respect, analytical sociology has two general ambitions. First, it wants to foster the development of new data collection procedures in which an explicit connection is established between the social mechanisms that one wants to study and the data collection design (see Hedström and Swedberg, 1996: 136–137). Second, analytical sociology wants to stimulate the more creative use of descriptive data (see Brückner, 2009). As testified by the essays collected in this book, the first goal induces scholars not to restrict the kind of empirical data that can be mobilized to describe the social outcome(s) of interest. Individual-level survey data (see Chapter 4), aggregate, historical administrative, and census data (see Chapters 6 and 14), geo-referenced data (see Chapter 3), video-recorded data (see Chapter 3), textual data (see Chapter 5), network-based data (see Chapters 6 and 12), digital, Web-based data (see Chapters 10, 13, and 16), and experimental data (see Chapters 7, 8, and 15) are regarded as equally relevant sources of empirical information with which to describe the facts to be explained. On the other hand, when empirical data are wrung in order to tackle the empirical signature of the individual- and network-level mechanisms at work, analytical sociology combines different types of data within the same study (see Chapters 3, 6, and 8).
The importance that analytical sociology attributes to descriptive tasks warrants special treatment because it is not always properly understood. Some commentators, indeed, have criticized the analytical sociology research program for what they consider to be its excessive and unjustified emphasis on explanatory goals (see Reiss, 2007: 164; Opp, 2005; Pisati, 2007). More explicitly than others, Bernardi (2007: 3, my translation from Italian) notes that “acknowledging the importance of description makes us aware of the risk of lapsing into what one may call ‘mechanismism’, that is, the obsessive quest for mechanisms behind phenomena that are not well defined and the existence of which is not well established.”
Two factors help explain this misperception of analytical sociology. First, it is true that there is some variability within the analytical tradition concerning the virtues of description. While Boudon (2002) explicitly distinguished between “scientific” and “descriptive” sociology – thus giving the impression that description can only play a secondary role within scientific research – Hedström and Swedberg (1998b: 17) made it explicit that “we do not wish to suggest that quantitative empirical research is of minor importance for the sociological enterprise. Quite the contrary: Quantitative research is essential both for descriptive purposes and for testing sociological theories.” More recently, Bearman (2012: 2) has provocatively declared:
Good sociology often involves explanation but I think good sociology can also be in the business of description without any explanation at all.…Some of the richest descriptions of things are those things that cannot be seen or known by individuals. And when those are described, I think we get some pretty good sociology.
Thus, according to the authors that one decides to consider more persuasive, analytical sociology may or may not be accused of privileging explanation over description. In my opinion, the important point is that it is not possible to explain something that has not been previously empirically described (see Goldthorpe, 2004). As a consequence, the most convincing position seems to be the one that gives equal importance to description and explanation and considers these tasks to be different steps in a more general research process – which is the meaning of the ordering between P2 and P3.
That said, there may be a more fundamental reason why some commentators see analytical sociology as a potential threat to description. This reason has to do with analytical sociology's critical assessment of the scope of multivariate statistical methods (see Hedström and Swedberg, 1998b: 15–17; Hedström, 2005: Ch. 5). The thrust of the criticism is expressed by Hedström (2005: 113) as follows: “causal explanations are not achieved by simply estimating parameters of generic statistical models, but by developing evidence-based generative models that explicate the mechanisms at work.” The crucial point here is that no matter how carefully the variables entering a statistical model are chosen; no matter how resistant the structure of the model's estimates is to different model specifications; no matter how large the amount of outcome variability accounted for by the predictors' variability – the model's coefficients cannot provide a detailed representation of the entities, the activities, and the relations among those entities and activities that are likely to be responsible for the observed outcome(s).
However, as testified by the above quotation from Hedström and Swedberg, it would be a mistake to equate this critical stance with an extreme, final dismissal of variable-centered statistical analysis (see also Brante, 2008). Analytical sociologists are perfectly aware that statistics is a powerful tool with which to figure out robust relations among factors measured at, and referring to, different levels of analysis. As P2 suggests, performing or referring to this kind of analysis is the first step in any serious mechanism-oriented analysis. Moreover, as my discussion of P7 will suggest (see Section 1.10), robust relations among context-, network-, and individual-level variables can be employed to increase the realism of formal, explicit models of social mechanisms. The argument among analytical sociologists, therefore, only concerns the scope and the appropriate task that can be legitimately attributed to multivariate statistical methods. No matter how carefully specified and sophisticated a statistical model may be, it can only provide a parsimonious description of a set of relations that represents the individual- or social-level signature of a (set of) social mechanism(s). But it cannot provide an explicit, detailed, and dynamic representation of that mechanism and of its high-level consequences (for a detailed application of these criticisms to a specific statistical technique, namely log-linear topological models, see Manzo, 2006).
It should be acknowledged that analytical sociology's critical assessment of variable-based analysis has long-standing roots in sociology (see Boudon, 1979; Sørensen, 1976) and in philosophy of social sciences (see Harré, 1972: 118). Among contemporary authors, scholars as different as Abbott (1988; 1992; 1997), Abell (2004), or Goldthorpe (2001) have also raised similar objections against regression-based methods. Statisticians like Freedman (1991; 2005) or Cox (1992) have urged resisting the temptation to interpret statistical coefficients as revealing underlying causal mechanisms.
Once again, however, analytical sociology's principles should not be assessed in isolation. The role that P2 attributes to description and to variable-centered analysis should be read in combination with the proposals contained in P5–P7. As we shall see, these principles attempt to build a complex interface between statistics and substantively oriented formal modeling which constitutes the constructive side of analytical sociology's critical stance toward variable-centered sociology (see also Manzo, 2007a).
1.5 Generative Models (P3)
While the rigorous (variable-based, when appropriate) empirical description of the social regularities to be explained is a fundamental task for analytical sociologists, P3 clarifies that description is only the first, preliminary step along a more complex research path whose core consists of explanation (see Figure 1.1).
From the point of view of philosophy of science, this explanatory ambition seems entirely legitimate (see Hempel, 1965: 245). However, given that a variety of understandings of how an explanation can be provided exist in social sciences (see Little, 1991) and that different explanatory modes co-habit within the ordinary and the academic world (see Mantzavinos, 2013), the specific conception of explanation that analytical sociology defends is likely to arouse resistance. In particular, within analytical sociology, explanation is understood as a model-based, mechanism-seeking activity. Let me first briefly discuss the concept of mechanism, and then explain the meaning of the “model-based” label.3
While the concept of mechanism has received a variety of definitions (for a collection of them, see Mahoney, 2001: 579–580; Gerring, 2008; Gross, 2009: 360–362; Hedström, 2005: 25; Hedström and Bearman, 2009b: 5–6), two simple ideas may be used to understand it. In terms of epistemic function, a mechanism is meant to make sense of the connection observed between (at least) two happenings. In this sense, a mechanism aims to eliminate black-box input/output relationships (see Bunge, 1997; Boudon, 1998a; Hedström and Swedberg, 1998b). In terms of content, by adapting a definition from biology (see Machamer, Darner, and Craver, 2000), a mechanism can be conceived as consisting of a set of organized entities whose properties and activities are able to trigger changes that generate the observed connections with some regularity.
It is essential to appreciate that the concept of mechanism is substantively empty. The specific entities, properties, activities, and connections, as well as the particular nature of these activities (for instance, probabilistic versus deterministic), should be defined only in connection with the specific outcome under scrutiny and in relation to the specific level of analysis at which the outcome is observed. It is for this reason that it is so difficult to find a consensual dictionary definition of the concept (see Hedström and Ylikoski, this volume). This analytical property should be regarded as an opportunity. The substantive emptiness of the concept of mechanism allows it to travel across the natural and social sciences, as well as across their research subfields, thus potentially enhancing knowledge accumulation, communication, and understandability. To borrow a concept from the sociology of science and technologies, a mechanism can be conceived as a “generic instrument,” that is, a (conceptual, in this case) device based on principles that can be adapted to different application domains and thus be reshaped again and again (see Shinn, 2008).4
It is also important to understand why a mechanism should not be equated with an intermediate/mediating variable (see Pawson, 1989: 130–131). From an epistemic point of view, the introduction of intermediate/controlling variables has the purpose of checking for the possibility that the order-zero relation is accounted for by elements that were not considered initially. In this sense, this operation aids understanding of the origin of the order-zero relation, and it echoes the goal of eliminating black-box input/output relationships, which also is the epistemic feature of guessing a mechanism. However, the content of a mechanism shows why this similarity is only apparent. The set of intermediate/controlling variables introduced does not amount to a set of entities, properties, activities, and connections that may be responsible for the social production of the order-zero relation. At best, these variables are fragmentary indicators of that underlying, potentially generative structured system. To give an example, a path-analytical diagram is indubitably able to dissect the (average) order-zero relation between, say, the occupations of parents and the final occupations of their offspring, hence increasing our initial understanding of the relation between the two variables. But the set of additional variables progressively introduced into the model only provides a remote (average) statistical signature of the underlying mechanism, which is likely to be made up of interacting actors and organizations with their own goals and opportunities.
Hence, one should be wary of statements like these: “Anyway, it is important to note that mechanism-based explanations are complex relationships between variables, which ultimately (i.e., on the micro level) are properties of actors” (Opp, 2007: 121) or
The appeal for mechanisms is a useful rallying cry, but the originality of a mechanism-based sociology has been oversold.…Arguing that mechanisms are concatenations of nonlinear functions is not an argument against the use of variables, since the primitive elements of functions – defined as inputs and outputs – can be redefined as variables.
(Morgan, 2005: 31)
These considerations on mechanism-based explanations make an important point explicit: the need for data structures and operations on these data structures to operationalize a theoretical representation of a (set of) mechanism(s). This is especially apparent when mechanisms are studied by means of formal modeling (see Section 1.9). However, the role performed by (numerical and logical) variables and functions relating and operating on these variables within a (formal model of a) mechanism is radically different from that of variables within a statistical model. While a formal model of a mechanism uses variables and functions to mimic the details of entities' properties and activities, and of connections among entities, with the aim of making these entities trigger changes over time that in the end may bring about the connection under scrutiny, variables and functions are used within a statistical model to detect a pattern of average effects which may reflect the (aggregate) statistical signature of the (unspecified) underlying mechanism. For this reason, while it is correct to say that the detailed theoretical representation of a mechanism entails the use of variables and functions, the main implications that some draw from this fact – that structured sets of intervening/mediating variables can be considered “mechanism sketches” (see Morgan and Winship, 2007: 238–242) and that multivariate statistics can be used to test mechanism-based explanations directly (see Opp, 2007: 121) – fail fully to appreciate the different functions and contents of variables and mechanisms.
