False Feedback in Economics
The Case for Replication
- Available for pre-order. Item will ship after July 22, 2021
This book investigates why economics makes less visible progress over time than scientific fields with a strong practical component, where interactions with physical technologies play a key role. The thesis of the book is that the main impediment to progress in economics is "false feedback", which it defines as the false result of an empirical study, such as empirical evidence produced by a statistical model that violates some of its assumptions. In contrast to scientific fields that work with physical technologies, false feedback is hard to recognize in economics. Economists thus have difficulties knowing where they stand in their inquiries, and false feedback will regularly lead them into the wrong directions.
The book searches for the reasons behind the emergence of false feedback. It thereby contributes to a wider discussion in the field of metascience about the practices of researchers when pursuing their daily businesses. The book thus offers a case study of metascience for the field of empirical economics.
The main strength of the book are the numerous smaller insights it provides throughout. The book delves into deep discussions of various theoretical issues, which it illustrates by many applied examples and a wide array of references, especially to philosophy of science. The book puts flesh on complicated and often abstract subjects, particularly when it comes to controversial topics such as p-hacking.
The reader gains an understanding of the main challenges present in empirical economic research and also of the possible solutions. The audience of the book are all applied researcher working with data, and, in particular, those who have sensed certain aspects of their research practice as problematic.
Table of Contents
1. Scientific Progress 2. Trial and Error 3. Conjectures and falsification 4. The garden of forking paths 5. The Duhem-Quine thesis 6. The detection of patterns 7. The illusion of true feedback 8. False feedback bubbles 9. The tree of knowledge 10. The locality of knowledge 11. Machine learning and sample splits 12. Practical experience 13. Robustness checks 14. Replication
Andrin Spescha is a Postdoctoral Researcher at ETH Zurich, KOF Swiss Economic Institute, Zurich, Switzerland. He received his PhD from ETH Zurich (Dr. sc. ETH) in 2018. Prior to this, he completed a Bachelor of Arts in Political Sciences and a Master of Arts in Economics at University of Zurich.