Agent-Based Models and Causal Inference - Gianluca Manzo

. .

Résumé :

List of Acronyms xi

List of Tables xii

Preface xiii

The Book in a Nutshell xvii

Introduction?1

1 The Book's Question?3

2 The Book's Structure?6

Part I: Conceptual and Methodological Clarifications?9

1 The Diversity of Views on Causality and Mechanisms?11

1.1 Causal Inference?11

1.2?Dependence?and?Production?Accounts of Causality?13

1.3?Horizontal?and?Vertical?Accounts of Mechanisms?17

1.3.1 Vertical?versus?Horizontal View?19

1.3.2 Horizontal?versus?Vertical View?21

1.4 Causality and Mechanism Accounts, and ABM's Perception?22

2 Agent-based Models and the Vertical View on Mechanism?25

2.1 ABMs and Object-oriented Programming?26

2.2 ABMs and Heterogeneity?27

2.3 ABMs and Micro-foundations?28

2.4 ABMs and Interdependence?28

2.5 ABMs and Time?29

2.6 ABMs and Multi-level Settings?30

2.7 Variables within Statistical Methods and ABMs?31

3 The Diversity of Agent-based Models?33

3.1 Abstract?versus?Data-driven ABMs: An Old Opposition?34

3.2 Abstract?versus?Data-driven ABMs: Recent Trends?36

3.3 Theoretical, Input, and Output Realism?38

3.4 Different Paths to More Realistic ABMs?40

3.4.1 Theoretically Blind Data-driven ABMs?41

3.4.2 Theoretically Informed Data-driven ABMs?45

Part 2: Data and Arguments in Causal Inference?49

4 Agent-based Models and Causal Inference?51

4.1 ABMs as Inferential Devices?52

4.1.1 The Role of Theoretical Realism?52

4.1.2 The Role of Output Realism and Empirical Validation?54

4.1.3 The Role of Input Realism and Empirical Calibration?55

4.1.4?In Principle?Conditions for Causally Relevant ABMs?57

4.1.5 Can Data-driven ABMs Produce Information?on Their Own??58

4.2?In Practice?Limitations?59

4.2.1 ABMs' Granularity and Data Availability?59

4.2.2 ABM's Granularity and Data Embeddedness?61

4.3?From-Within-the-Method?Reliability Tools?62

4.3.1 Sensitivity Analysis?64

4.3.2 Robustness Analysis?65

4.3.3 Dispersion Analysis?65

4.3.4 Model Analysis?66

5 Causal Inference in Experimental and Observational Methods?69

5.1 Causal Inference: Cautionary Tales?71

5.2?In Practice?Untestable Assumptions?73

5.2.1 RCTs and Heterogeneity?73

5.2.2 IVs and the Relevance Condition?74

5.2.3 DAGs, Causal Discovery Algorithms and Graph Indistinguishability?76

5.3?In Principle?Untestable Assumptions?79

5.3.1 RCTs and Stable Unit Treatment Value Assumption (SUTVA)?79

5.3.2 IVs and the Exclusion Condition?81

5.3.3 DAGs and Strategies for Causal Identification?83

5.3.3.1 DAGs and the Backdoor Criterion?83

5.3.3.2 DAGs and the Front Door Criterion?84

5.4 Are ABMs, Experimental and Observational Methods Fundamentally Similar??85

5.4.1 Objection 1: ABM Lacks Formal Assumptions?86

5.4.2 Objection 2: ABM Lacks Materiality?89

5.4.3 Objection 3: ABMs Lack Robustness?91

5.5 A Common Logic: Abduction?94

6 Method Diversity and Causal Inference?95

6.1 Causal Pluralism, Causal Exclusivism, and Evidential Pluralism?97

6.2 A Pragmatist Account of Evidence?99

6.3 Evidential Pluralism and Coherentism?101

6.4 When is Diverse Evidence Most Relevant??104

6.5 Examples of Method Synergies?106

6.5.1 Obesity: ABMs and Regression Models?106

6.5.2 Network Properties: ABMs and SIENA Models?109

6.5.3 HIV prevalence: ABMs and RCTs?111

6.5.4 HIV treatment...

Biographie:

Gianluca Manzo 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.