Towards Bayesian Model-Based Demography: Agency, Complexity and Uncertainty in Migration Studies

Description

This open access book presents a ground-breaking approach to developing micro-foundations for demography and migration studies. It offers a unique and novel methodology for creating empirically grounded agent-based models of international migration – one of the most uncertain population processes and a top-priority policy area. The book discusses in detail the process of building a simulation model of migration, based on a population of intelligent, cognitive agents, their networks and institutions, all interacting with one another. The proposed model-based approach integrates behavioural and social theory with formal modelling, by embedding the interdisciplinary modelling process within a wider inductive framework based on the Bayesian statistical reasoning. Principles of uncertainty quantification are used to devise innovative computer-based simulations, and to learn about modelling the simulated individuals and the way they make decisions. The identified knowledge gaps are subsequently filled with information from dedicated laboratory experiments on cognitive aspects of human decision-making under uncertainty. In this way, the models are built iteratively, from the bottom up, filling an important epistemological gap in migration studies, and social sciences more broadly. Foreword Bayesian Epistemic Probability Agent-Based or Model-Based Demography? Conclusion References About the Authors Lead Author Contributors Acknowledgements Contents List of Boxes List of Figures List of Tables Part I: Preliminaries Chapter 1: Introduction 1.1 Why Bayesian Model-Based Approaches for Studying Migration? 1.2 Aims and Scope of the Book 1.3 Structure of the Book 1.4 Intended Audience and Different Paths Through the Book Chapter 2: Uncertainty and Complexity: Towards Model-Based Demography 2.1 Uncertainty and Complexity in Demography and Migration 2.2 High Uncertainty and Impact: Why Model Asylum Migration? 2.3 Shifting Paradigm: Description, Prediction, Explanation 2.4 Towards Micro-foundations in Migration Modelling 2.5 Philosophical Foundations: Inductive, Deductive and Abductive Approaches 2.6 Model-Based Demography as a Research Programme Part II: Elements of the Modelling Process Chapter 3: Principles and State of the Art of Agent-Based Migration Modelling 3.1 The Role of Models in Studying Complex Systems 3.1.1 What Can a Model Do? 3.1.2 Not ‘the Model of’, but ‘a Model to’ 3.1.3 Complications 3.2 Complex Social Phenomena and Agent-Based Models 3.2.1 Modelling Migration 3.2.2 Uncertainty 3.3 Agent-Based Models of Migration: Introducing the Routes and Rumours Model 3.3.1 Research Questions 3.3.2 Space and Topology 3.3.3 Decision-Making Mechanisms 3.3.4 Social Interactions and Information Exchange 3.4 A Note on Model Implementation 3.5 Knowledge Gaps in Existing Migration Models Chapter 4: Building a Knowledge Base for the Model 4.1 Key Conceptual Challenges of Measuring Asylum Migration and Its Drivers 4.2 Case Study: Syrian Asylum Migration to Europe 2011–19 4.3 Data Overview: Process and Context 4.3.1 Key Dimensions of Migration Data 4.3.2 Process-Related Data 4.3.3 Contextual Data 4.4 Quality Assessment Framework for Migration Data 4.4.1 Existing Frameworks 4.4.2 Proposed Dimensions of Data Assessment: Example of Syrian Asylum Migration 4.5 The Uses of Data in Simulation Modelling 4.6 Towards Better Migration Data: A General Reflection Chapter 5: Uncertainty Quantification, Model Calibration and Sensitivity 5.1 Bayesian Uncertainty Quantification: Key Principles 5.2 Preliminaries of Statistical Experimental Design 5.3 Analysis of Experiments: Response Surfaces and Meta-Modelling 5.4 Uncertainty and Sensitivity Analysis 5.5 Bayesian Methods for Model Calibration Chapter 6: The Boundaries of Cognition and Decision Making 6.1 The Role of Individual-Level Empirical Evidence in Agent-Based Models 6.2 Prospect Theory and Discrete Choice 6.3 Eliciting Subjective Probabilities 6.4 Conjoint Analysis of Migration Drivers 6.5 Design, Implementation, and Limitations of Psychological Experiments for Agent-Based Models 6.6 Immersive Decision Making in the Experimental Context Chapter 7: Agent-Based Modelling and Simulation with Domain-Specific Languages 7.1 Introduction 7.2 Domain-Specific Languages for Modelling 7.2.1 Requirements 7.2.2 The Modelling Language for Linked Lives (ML3) 7.2.3 Discussion 7.3 Model Execution 7.3.1 Execution of ML3 Models 7.3.2 Discussion 7.4 Domain-Specific Languages for Simulation Experiments 7.4.1 Basics 7.4.2 Complex Experiments 7.4.3 Reproducibility 7.4.4 Related Work 7.4.5 Discussion 7.5 Managing the Model’s Context 7.6 Conclusion Part III: Model Results, Applications, and Reflections Chapter 8: Towards More Realistic Models 8.1 Integrating the Five Building Blocks of the Modelling Process 8.2 Risk and Rumours: Motivation and Model Description 8.3 Uncertainty, Sensitivity, and Areas for Data Collection 8.4 Risk and Rumours with Reality: Adding Empirical Calibration 8.5 Reflections on the Model Building and Implementation Chapter 9: Bayesian Model-Based Approach: Impact on Science and Policy 9.1 Bayesian Model-Based Migration Studies: Evaluation and Perspectives 9.2 Advancing the Model-Based Agenda Across Scientific Disciplines 9.3 Policy Impact: Scenario Analysis, Foresight, Stress Testing, and Planning 9.3.1 Early Warnings and Stress Testing 9.3.2 Forecasting and Scenarios 9.3.3 Assessing Policy Interventions 9.4 Towards a Blueprint for Model-Based Policy and Decision Support Chapter 10: Open Science, Replicability, and Transparency in Modelling 10.1 The Replication Crisis and Questionable Research Practices 10.2 Open Science and Improving Research Practices 10.3 Implications for Modellers Chapter 11: Conclusions: Towards a Bayesian Modelling Process 11.1 Bayesian Model-Based Population Studies: Moving the Boundaries 11.2 Limitations and Lessons Learned: Barriers and Trade-Offs 11.3 Towards Model-Based Social Enquiries: The Way Forward Appendices: Supporting Information Appendix A. Architecture of the Migrant Route Formation Models A1. Model Description A2. Processes A3. Illustration Appendix B. Meta-Information on Data Sources on Syrian Migration into Europe B1. Selected Key Sources of Data on Syrian Migration into Europe B2. Supplementary General Sources on Migration Processes, Drivers or Features Appendix C. Uncertainty and Sensitivity Analysis: Sample Output Appendix D. Experiments: Design, Protocols, and Ethical Aspects D.1. Prospect Theory and Discrete Choice Experiment D.2. Eliciting Subjective Probabilities D.3. Conjoint Analysis of Migration Drivers Appendix E. Provenance Description of the Route Formation Models Glossary References Index