Details

Agent-based Models and Causal Inference


Agent-based Models and Causal Inference


Wiley Series in Computational and Quantitative Social Science, Band 1 1. Aufl.

von: Gianluca Manzo

66,99 €

Verlag: Wiley
Format: EPUB
Veröffentl.: 28.01.2022
ISBN/EAN: 9781119704461
Sprache: englisch
Anzahl Seiten: 176

DRM-geschütztes eBook, Sie benötigen z.B. Adobe Digital Editions und eine Adobe ID zum Lesen.

Beschreibungen

<p><b>Agent-based Models and Causal Inference</b></p> <p>Scholars of causal inference have given little credence to the possibility that ABMs could be an important tool in warranting causal claims. Manzo&rsquo;s book makes a convincing case that this is a mistake. The book starts by describing the impressive progress that ABMs have made as a credible methodology in the last several decades. It then goes on to compare the inferential threats to ABMs versus the traditional methods of RCTs, regression, and instrumental variables showing that they have a common vulnerability of being based on untestable assumptions. The book concludes by looking at four examples where an analysis based on ABMs complements and augments the evidence for specific causal claims provided by other methods. Manzo has done a most convincing job of showing that ABMs can be an important resource in any researcher&rsquo;s tool kit.<br /><br /><b><i>Christopher Winship,</i></b> Diker-Tishman Professor of Sociology, Harvard University, USA</p> <p><i>Agent-based Models and Causal Inference&nbsp;</i>is a first-rate contribution to the debate on, and practice of, causal claims. With exemplary rigor, systematic precision and pedagogic clarity, this book contrasts the assumptions about causality that undergird agent-based models, experimental methods, and statistically based observational methods, discusses the challenges these methods face as far as inferences go, and, in light of this discussion, elaborates the case for combining these methods&rsquo; respective strengths: a remarkable achievement.&nbsp;&nbsp;</p> <p><i><b>Ivan Ermakoff</b></i>, Professor of Sociology, University of Wisconsin-Madison, USA</p> <p>Agent-based models are a uniquely powerful tool for understanding how patterns in society may arise in often surprising and counter-intuitive ways. This book offers a strong and deeply reflected argument for how ABM&rsquo;s can do much more: add to actual empirical explanation. The work is of great value to all social scientists interested in learning how computational modelling can help unraveling the complexity of the real social world.</p> <p><i><b>Andreas Flache</b></i>, Professor of Sociology at the University of Groningen, Netherlands</p> <p><i>Agent-based Models and Causal Inference&nbsp;</i>is an important and much-needed contribution to sociology and computational social science. The book provides a rigorous new contribution to current understandings of the foundation of causal inference and justification in the social sciences. It provides a powerful and cogent alternative to standard statistical causal-modeling approaches to causation. Especially valuable is Manzo&rsquo;s careful analysis of the conditions under which an agent-based simulation is relevant to causal inference. The book represents an exceptional&nbsp;contribution to sociology, the philosophy of social science, and the epistemology of simulations and models.</p> <p><i><b>Daniel Little</b></i>, Professor of philosophy,&nbsp;University of Michigan, USA</p> <hr /> <p><i>Agent-based Models and Causal Inference</i> delivers an insightful investigation into the conditions under which different quantitative methods can legitimately hold to be able to establish causal claims. The book compares agent-based computational methods with randomized experiments, instrumental variables, and various types of causal graphs.</p> <p>Organized in two parts, <i>Agent-based Models and Causal Inference</i> connects the literature from various fields, including causality, social mechanisms, statistical and experimental methods for causal inference, and agent-based computation models to help show that causality means different things within different methods for causal analysis, and that persuasive causal claims can only be built at the intersection of these various methods.</p> <p>Readers will also benefit from the inclusion of:</p> <ul> <li>A thorough comparison between agent-based computation models to randomized experiments, instrumental variables, and several types of causal graphs</li> <li>A compelling argument that observational and experimental methods are not qualitatively superior to simulation-based methods in their ability to establish causal claims</li> <li>Practical discussions of how statistical, experimental and computational methods can be combined to produce reliable causal inferences</li> </ul> <p>Perfect for academic social scientists and scholars in the fields of computational social science, philosophy, statistics, experimental design, and ecology, <i>Agent-based Models and Causal Inference</i> will also earn a place in the libraries of PhD students seeking a one-stop reference on the issue of causal inference in agent-based computational models.</p>
<p>List of Acronyms xi</p> <p>List of Tables xii</p> <p>Preface xiii</p> <p>The Book in a Nutshell xvii</p> <p><b>Introduction&nbsp;</b><b>1</b></p> <p>1 The Book&rsquo;s Question&nbsp;3</p> <p>2 The Book&rsquo;s Structure&nbsp;6</p> <p><b>Part I: Conceptual and Methodological Clarifications&nbsp;</b><b>9</b></p> <p><b>1 The Diversity of Views on Causality and Mechanisms&nbsp;</b><b>11</b></p> <p>1.1 Causal Inference&nbsp;11</p> <p>1.2&nbsp;Dependence&nbsp;and&nbsp;Production&nbsp;Accounts of Causality&nbsp;13</p> <p>1.3&nbsp;Horizontal&nbsp;and&nbsp;Vertical&nbsp;Accounts of Mechanisms&nbsp;17</p> <p>1.3.1 Vertical&nbsp;versus&nbsp;Horizontal View&nbsp;19</p> <p>1.3.2 Horizontal&nbsp;versus&nbsp;Vertical View&nbsp;21</p> <p>1.4 Causality and Mechanism Accounts, and ABM&rsquo;s Perception&nbsp;22</p> <p><b>2 Agent-based Models and the Vertical View on Mechanism&nbsp;</b><b>25</b></p> <p>2.1 ABMs and Object-oriented Programming&nbsp;26</p> <p>2.2 ABMs and Heterogeneity&nbsp;27</p> <p>2.3 ABMs and Micro-foundations&nbsp;28</p> <p>2.4 ABMs and Interdependence&nbsp;28</p> <p>2.5 ABMs and Time&nbsp;29</p> <p>2.6 ABMs and Multi-level Settings&nbsp;30</p> <p>2.7 Variables within Statistical Methods and ABMs&nbsp;31</p> <p><b>3 The Diversity of Agent-based Models&nbsp;</b><b>33</b></p> <p>3.1 Abstract&nbsp;versus&nbsp;Data-driven ABMs: An Old Opposition&nbsp;34</p> <p>3.2 Abstract&nbsp;versus&nbsp;Data-driven ABMs: Recent Trends&nbsp;36</p> <p>3.3 Theoretical, Input, and Output Realism&nbsp;38</p> <p>3.4 Different Paths to More Realistic ABMs&nbsp;40</p> <p>3.4.1 &ldquo;Theoretically Blind&rdquo; Data-driven ABMs&nbsp;41</p> <p>3.4.2 &ldquo;Theoretically Informed&rdquo; Data-driven ABMs&nbsp;45</p> <p><b>Part 2: Data and Arguments in Causal Inference&nbsp;</b><b>49</b></p> <p><b>4 Agent-based Models and Causal Inference&nbsp;</b><b>51</b></p> <p>4.1 ABMs as Inferential Devices&nbsp;52</p> <p>4.1.1 The Role of &ldquo;Theoretical Realism&rdquo;&nbsp;52</p> <p>4.1.2 The Role of &ldquo;Output Realism&rdquo; and Empirical Validation&nbsp;54</p> <p>4.1.3 The Role of &ldquo;Input Realism&rdquo; and Empirical Calibration&nbsp;55</p> <p>4.1.4&nbsp;In Principle&nbsp;Conditions for Causally Relevant ABMs&nbsp;57</p> <p>4.1.5 Can Data-driven ABMs Produce Information&nbsp;on Their Own?&nbsp;58</p> <p>4.2&nbsp;In Practice&nbsp;Limitations&nbsp;59</p> <p>4.2.1 ABMs&rsquo; Granularity and Data Availability&nbsp;59</p> <p>4.2.2 ABM&rsquo;s Granularity and Data Embeddedness&nbsp;61</p> <p>4.3&nbsp;From-Within-the-Method&nbsp;Reliability Tools&nbsp;62</p> <p>4.3.1 Sensitivity Analysis&nbsp;64</p> <p>4.3.2 Robustness Analysis&nbsp;65</p> <p>4.3.3 Dispersion Analysis&nbsp;65</p> <p>4.3.4 Model Analysis&nbsp;66</p> <p><b>5 Causal Inference in Experimental and Observational Methods&nbsp;</b><b>69</b></p> <p>5.1 Causal Inference: Cautionary Tales&nbsp;71</p> <p>5.2&nbsp;In Practice&nbsp;Untestable Assumptions&nbsp;73</p> <p>5.2.1 RCTs and Heterogeneity&nbsp;73</p> <p>5.2.2 IVs and the &ldquo;Relevance&rdquo; Condition&nbsp;74</p> <p>5.2.3 DAGs, Causal Discovery Algorithms and Graph Indistinguishability&nbsp;76</p> <p>5.3&nbsp;In Principle&nbsp;Untestable Assumptions&nbsp;79</p> <p>5.3.1 RCTs and &ldquo;Stable Unit Treatment Value Assumption&rdquo; (SUTVA)&nbsp;79</p> <p>5.3.2 IVs and the &ldquo;Exclusion&rdquo; Condition&nbsp;81</p> <p>5.3.3 DAGs and Strategies for Causal Identification&nbsp;83</p> <p>5.3.3.1 DAGs and the &ldquo;Backdoor&rdquo; Criterion&nbsp;83</p> <p>5.3.3.2 DAGs and the &ldquo;Front Door&rdquo; Criterion&nbsp;84</p> <p>5.4 Are ABMs, Experimental and Observational Methods Fundamentally Similar?&nbsp;85</p> <p>5.4.1 Objection 1: ABM Lacks &ldquo;Formal&rdquo; Assumptions&nbsp;86</p> <p>5.4.2 Objection 2: ABM Lacks &ldquo;Materiality&rdquo;&nbsp;89</p> <p>5.4.3 Objection 3: ABMs Lack &ldquo;Robustness&rdquo;&nbsp;91</p> <p>5.5 A Common Logic: &ldquo;Abduction&rdquo;&nbsp;94</p> <p><b>6 Method Diversity and Causal Inference&nbsp;</b><b>95</b></p> <p>6.1 Causal Pluralism, Causal Exclusivism, and Evidential Pluralism&nbsp;97</p> <p>6.2 A Pragmatist Account of Evidence&nbsp;99</p> <p>6.3 Evidential Pluralism and &ldquo;Coherentism&rdquo;&nbsp;101</p> <p>6.4 When is Diverse Evidence Most Relevant?&nbsp;104</p> <p>6.5 Examples of Method Synergies&nbsp;106</p> <p>6.5.1 Obesity: ABMs and Regression Models&nbsp;106</p> <p>6.5.2 Network Properties: ABMs and SIENA Models&nbsp;109</p> <p>6.5.3 HIV prevalence: ABMs and RCTs&nbsp;111</p> <p>6.5.4 HIV treatments: ABMs and DAG-based identification strategies&nbsp;113</p> <p><b>Coda&nbsp;</b><b>115</b></p> <p>1 Possible Objections&nbsp;116</p> <p>1.1 Causation is Not Constitution&nbsp;117</p> <p>1.2 Lack of a Specific Research Strategy&nbsp;118</p> <p>1.3 A Limited Methodological Spectrum&nbsp;119</p> <p>2 Summary&nbsp;121</p> <p>References 127</p> <p>Index 149&nbsp;&nbsp;</p>
<p><b>Gianluca Manzo</b> is a professor of sociology at Sorbonne University and a fellow of the European Academy of Sociology. He has held various positions at institutions across the world including Nuffield College, Columbia University, the European University Institute (EUI), and the Universities of Oslo, Barcelona, Cologne, and Trento.</p>
<p><b>Agent-based Models and Causal Inference</b></p> <p>Scholars of causal inference have given little credence to the possibility that ABMs could be an important tool in warranting causal claims. Manzo’s book makes a convincing case that this is a mistake. The book starts by describing the impressive progress that ABMs have made as a credible methodology in the last several decades. It then goes on to compare the inferential threats to ABMs versus the traditional methods of RCTs, regression, and instrumental variables showing that they have a common vulnerability of being based on untestable assumptions. The book concludes by looking at four examples where an analysis based on ABMs complements and augments the evidence for specific causal claims provided by other methods. Manzo has done a most convincing job of showing that ABMs can be an important resource in any researcher’s tool kit.<BR><b><i>Christopher Winship,</b> Diker-Tishman Professor of Sociology, Harvard University, USA</i> <p><i>Agent-based Models and Causal Inference</i> delivers an insightful investigation into the conditions under which different quantitative methods can legitimately hold to be able to establish causal claims. The book compares agent-based computational methods with randomized experiments, instrumental variables, and various types of causal graphs. <p>Organized in two parts, <i>Agent-based Models and Causal Inference</i> connects the literature from various fields, including causality, social mechanisms, statistical and experimental methods for causal inference, and agent-based computation models to help show that causality means different things within different methods for causal analysis, and that persuasive causal claims can only be built at the intersection of these various methods. <p>Readers will also benefit from the inclusion of: <ul><li>A thorough comparison between agent-based computation models to randomized experiments, instrumental variables, and several types of causal graphs</li> <li>A compelling argument that observational and experimental methods are not qualitatively superior to simulation-based methods in their ability to establish causal claims</li> <li>Practical discussions of how statistical, experimental and computational methods can be combined to produce reliable causal inferences </li></ul> <p>Perfect for academic social scientists and scholars in the fields of computational social science, philosophy, statistics, experimental design, and ecology, <i>Agent-based Models and Causal Inference</i> will also earn a place in the libraries of PhD students seeking a one-stop reference on the issue of causal inference in agent-based computational models.
<p>&nbsp;</p> <p>Scholars of causal inference have given little credence to the possibility that ABMs could be an important tool in warranting causal claims. Manzo&rsquo;s book makes a convincing case that this is mistake. The book starts by describing the impressive progress that ABMs have made as a credible methodology in the last several decades. It then goes on to compare the inferential threats to ABMs versus the traditional methods of&nbsp; RCTs, regression, and instrumental variables showing that they have common vulnerability of being based on untestable assumptions. The book concludes by looking at four examples where an analysis based on ABMs complements and augments the evidence for specific causal claims provided by other methods. Manzo has done a most convincing job of showing that ABMs can be an important in any researcher&rsquo;s tool kit.</p> <p>&nbsp;</p> <p>--Christopher Winship, professor of sociology at Harvard University</p> <p><br /><i>Agent-based Models and Causal Inference</i>&nbsp;is a first-rate contribution to the debate on, and practice of, causal claims. With exemplary rigor, systematic precision and pedagogic clarity, this book contrasts the assumptions about causality that undergird agent-based models, experimental methods, and statistically based observational methods, discusses the challenges these methods face as far as inferences go, and, in light of this discussion, elaborates the case for combining these methods&rsquo; respective strengths: a remarkable achievement. &nbsp;</p> <p>-- Ivan Ermakoff, professor of sociology at the University of Wisconsin-Madison</p> <p>Agent-based models are a uniquely powerful tool for understanding how patterns in society may arise in often surprising and counter-intuitive ways. This book offers a strong and deeply reflected argument for how ABM&rsquo;s can do much more: add to actual empirical explanation. The work is of great value to all social scientists interested in learning how computational modelling can help unraveling the complexity of the real social world.</p> <p>-- Andreas Flache, professor of sociology at the University of Groningen</p> <p><i>Agent-based Models and Causal Inference</i>&nbsp;is an important and much-needed contribution to sociology and computational social science. The book provides a rigorous new contribution to current understandings of the foundation of causal inference and justification in the social sciences. It provides a powerful and cogent alternative to standard statistical causal-modeling approaches to causation. Especially valuable is Manzo&rsquo;s careful analysis of the conditions under which an agent-based simulation is relevant to causal inference. The book represents an exceptional&nbsp;contribution to sociology, the philosophy of social science, and the epistemology of simulations and models.</p> <p>-- Daniel Little, professor of philosophy at the University of Michigan</p>

Diese Produkte könnten Sie auch interessieren:

Statistics for Microarrays
Statistics for Microarrays
von: Ernst Wit, John McClure
PDF ebook
84,99 €