The Handbook of Behavioral Operations

2Econometrics for Experiments

Kyle Hyndman¹ and Matthew Embrey²

¹ Naveen Jindal School of Management, University of Texas at Dallas, Richardson, TX, USA

² Department of Economics, University of Sussex, Falmer, Brighton, UK

2.1 Introduction

The use of experimental methods in operations management has undergone tremendous growth in recent years. Unsurprisingly, this has led to heterogeneity in the techniques used to analyze the experimental data. Because of this, it is a good time to take stock of the methods and tools that have been commonly used to date. It is our hope to lay out and explain the current best practices in econometric analysis of experimental data. As will become clear, in some areas there are competing schools of thought. The most prominent example concerns the issue of what constitutes an independent observation and how to deal with dependence across observations. We endeavor to explain these competing approaches, including their strengths and weaknesses, which are often overlooked. Throughout the chapter, we make an effort to illustrate the issues under discussion as they relate to both published and current research in the field. Several authors have graciously made their data available to us, which allows us to illustrate certain tests or methods of analysis that may not have been included in the original papers.

This chapter is divided, roughly, in two parts. In the first part, which consists of Sections 2.2–2.4, we cover the basics of experimental design ...