Avoiding epistemological silos and empirical elephants in OM: How to combine empirical and simulation methods?

• Author: Fabian J. Sting,Aravind Chandrasekaran,Kevin Linderman
• Date: 01 November 2018
• DOI: http://doi.org/10.1016/j.jom.2018.11.003
• Published date: 01 November 2018

Contents lists available at ScienceDirect

Journal of Operations Management

journal homepage: www.elsevier.com/locate/jom

Avoiding epistemological silos and empirical elephants in OM: How to

combine empirical and simulation methods?

Aravind Chandrasekaran

a

, Kevin Linderman

b,∗

, Fabian J. Sting

c

a

Center for Operational Excellence, The Ohio State University, USA

b

Curtis L. Carlson Professor in Supply Chain and Operations, University of Minnesota, USA

c

Chaired Professor of Supply Chain Management, Strategy and Innovation, University of Cologne, Erasmus University Rotterdam, Germany

ARTICLE INFO

Accepted by: S. de Treville

ABSTRACT

Progression of knowledge in operations management (OM) relies on researchers building and testing theories

using data from practice. However, standalone empirical research designs have inherent limitations and may not

adequately capture complex OM problems. This may result in researchers narrowing the scope of the problems

that they create epistemological silos and study empirical elephants. In this introductory article to the special

issue, we look at the ways simulation methods can augment and answer questions that are not addressed through

traditional empirical methods. We offer a framework regarding how and when to use simulation methods for a

given research objectives and design. We conclude by discussing the contingencies of using these methods as

well as the role of reviewers and editors in evaluating papers that involve such methods.

1. Introduction

Scientific progress in any management discipline relies on re-

searchers' efforts to build new theories and to test existing theories by

engaging with practice (Van de Ven and Johnson, 2006). Empirical

research in operations management (OM) has been built on this prin-

ciple (Flynn et al., 1990), with OM scholars employing empirical

methods such as case studies (Wu and Choi, 2005;Su et al., 2014) to

develop new theories and psychometric (Chandrasekaran et al., 2012)

and econometric analysis (Dhanorkar et al., 2017) to test these theories.

These methods have helped advance the knowledge in our field. Al-

though we have made considerable progress in research methods in

recent decades, a vast majority of OM research is still methodologically

isolated from neighboring disciplines and topically distant from prac-

tice, which brings into question our relevancy (Sodhi and Tang, 2014).

For instance, many empirical studies only focus on simple linear re-

lationships, often obtained from cross-sectional data that may not fully

represent real-world challenges. In this vein, Boyer and Swink (2008)

warn that scholars run the risk of studying “empirical elephants”—a

metaphor drawing from an ancient parable of five blind men trying to

describe an elephant by touching only one portion of the animal. In-

deed, researchers cannot fully understand practical phenomena if they

do not recognize the limitations endemic to their research methods and

utilize the full range of methodologies available to them. Empirical

research faces a number of challenges such as data availability and

reliability, unobserved variables, and the ability to analyze dynamic

phenomena that take place over an extended time horizon. Researchers

may choose to ignore such questions or simplify them, but to do so

means foregoing studies that address the most relevant and con-

temporary problems faced by operations managers.

Simulation models hold great potential for understanding problems

beyond the scope of traditional empirical research methods. They in-

volve the application of computer software to represent decision-

making processes, systems, and agents (Law, 2015). Simulations make

it possible to capture wide sets of variables, interconnections, and

boundary conditions; whose complexity goes beyond what can be

captured by either closed-form analytical models or conceptual models

(Keys and Wolfe, 1990), thus are able to capture the rich, non-linear,

and dynamic features typical of operational environments. The purpose

of this special issue is to demonstrate to OM researchers how combining

simulations with empirical methods can help develop deeper insights

and overcome some the inherent limitations in single-method studies.

The use of simulation is not new to OM or the readers of the Journal

of Operations Management (JOM). In fact, the first few issues of JOM

published several impactful simulation articles (e.g., Flowers, 1980;

Glaser and Hottenstein, 1982). In the past, JOM has also published a

special issue on the use of simulation models to study OM problems

(Shafer and Smunt, 2004). The broader management literature has also

used simulation models to study management problems. Some widely

used management theories have been developed from simulation

https://doi.org/10.1016/j.jom.2018.11.003

Received 4 September 2018; Received in revised form 28 November 2018; Accepted 28 November 2018

∗

Corresponding author.

E-mail address: klinderman@csom.umn.edu (K. Linderman).

Journal of Operations Management 63 (2018) 1–5

Available online 06 December 2018

0272-6963/ © 2018 Elsevier B.V. All rights reserved.

T