The Explanatory Power of Models - Method for Social Science

In this review of The Explanatory Power of Models, the reviewer finds a focused methodological text aimed at researchers who want to close the gap between empirical work and theory. The single biggest reason to buy is its systematic treatment of how to construct models that improve explanatory power across disciplines; the book works through statistical, mathematical, conceptual and computational modelling with the explicit goal of restoring classical induction as a methodological tool. It reads like a careful, academic review of modelling practice rather than a textbook for beginners.

Key Features

• Method-driven approach: Presents a progressive method for constructing models that is intended to bridge empirical and theoretical research and improve explanatory power.
• Cross-disciplinary examples: Examines modelling practices from statistics, mathematics, diagrams, machines and simulations to illustrate transferable techniques.
• Reverse engineering inspiration: Uses reverse engineering as a guiding principle to extract structure and causal insight from complex empirical systems.
• Critique of covering-law approach: Abandons the standard covering law model and argues for a restored role for classical induction in social science explanation.
• Attention to computation: Includes discussion of computer simulations and artificial neural networks to connect traditional modelling with modern tools.

Who It's For

This book is best for social scientists, advanced graduate students and researchers in applied statistics or computational social science who are comfortable with formal models and want a methodological framework to make models more explanatory. It helps readers who already use statistical or mathematical models and seek ways to link those tools to broader theory.

Those seeking an introductory textbook, a how-to coding guide, or step-by-step statistical instruction should look elsewhere; the treatment is conceptual and methodological rather than a beginner's manual or a source of practical software tutorials.

Pros & Cons

Pros

• Thorough, method-focused analysis that offers a clear framework to improve explanatory power in models.
• Draws on a wide range of modelling practices, including diagrams, maps, machines and simulations, for concrete comparison.
• Engages with modern computational topics such as artificial neural networks to keep the discussion current.

Cons

• The book is conceptual and assumes prior familiarity with statistical and mathematical models, so it may not suit readers wanting hands-on tutorials.

Specifications

Our Verdict

The Explanatory Power of Models is a valuable methodological contribution for researchers who want to make their models more explanatory rather than merely predictive. It is good value for scholars seeking a rigorous, cross-disciplinary framework that links empirical modelling practices to theory construction, though it is best suited to readers with some prior exposure to formal models.

Frequently Asked Questions

Does this book cover computational methods?
Yes. The text includes discussion of computer simulations and artificial neural networks alongside traditional modelling techniques.

Is the book suitable for beginners?
Not ideal for true beginners; it assumes familiarity with statistical and mathematical modelling and focuses on methodology rather than step-by-step instruction.

What methodological stance does the author take?
The author abandons the standard covering law approach, favors classical induction, and draws inspiration from reverse engineering to build explanatory models.