Making Social Sciences More Scientific: The Need for Predictive Models Book

List of Figures xiii

List of Tables xv

Part I The Limitations of Descriptive Methodology

1 Why Social Sciences Are Not Scientific Enough 3

2 Can Social Science Approaches Find the Law of Gravitation? 14

3 How to Construct Predictive Models: Simplicity and Nonabsurdity 23

4 Example of Model Building: Electoral Volatility 34

5Physicists Multiply, Social Scientists Add-Even When It Does Not Add Up 52

6 All Hypotheses Are Not Created Equal 71

7 Why Most Numbers Published in Social Sciences Are Dead on Arrival 82

Part II Quantitatively Predictive Logical Models

8 Forbidden Areas and Anchor Points 95

9 Geometric Means and Lognormal Distributions 120

10 Example of Interlocking Models: Party Sizes and Cabinet Duration 130

11 Beyond Constraint-Based Models: Communication Channels and Growth Rates 139

12 Why we should Shift to Symmetric Regression 154

13 All Indices Are Not Created Equal 176

Part III Synthesis of Predictive and Descriptive Approaches

14 From Descriptive to Predictive Approaches 187

15 Recommendations for Better Regression 199

16 Converting from Descriptive Analysis to Predictive Models 215

17 Are Electoral Studies a Rosetta Stone for Parts of Social Sciences? 225

18 Beyond Regression: The Need for Predictive Models 236

References 241

Index 249