Measurement and Moderation: Finding the Boundary Conditions in Logistics and Supply Chain Research

• Author: Thomas J. Goldsby,Carl Marcus Wallenburg,Jason W. Miller,A. Michael Knemeyer
• DOI: http://doi.org/10.1111/jbl.12013
• Published date: 01 June 2013
• Date: 01 June 2013

Measurement and Moderation: Finding the Boundary Conditions

in Logistics and Supply Chain Research

Thomas J. Goldsby

1

, A. Michael Knemeyer

1

, Jason W. Miller

1

, and Carl Marcus Wallenburg

2

1

The Ohio State University

2

Kühne Institute for Logistics Management, WHU –Otto Beisheim School of Management

Amoderator is any variable that affects the strength of a relationship between a predictor and an outcome variable. While simple in concept,

the application of moderation analysis can yield profound implications to research conducted in logistics and supply chain management.

Moderation analysis illuminates boundary conditions to purported relationships, providing a deeper perspective on what may, to date, represent

generalizable findings and commonly held beliefs in the field. Such findings prove interesting and enrich our theories. Further, moderation relies

on precise measurement of theoretical constructs in order to avoid attenuation of statistical tests and detect interaction effects. This thought lead-

ership piece seeks to: (1) assert the value of moderation analysis and encourage a more prominent place in our survey-based research projects,

(2) provide best practice approaches for using this type of analysis in pursuit of greater depth and clarity in our research, and (3) provide seeds

for potential research projects that could benefit from the use of this type of analysis. Guidance is also provided for reviewers who assess manu-

scripts featuring moderation.

Keywords: moderation analysis; interactions; empirical methods; measurement; boundary conditions

INTRODUCTION

As the logistics and supply chain management (SCM) disciplines

mature, increased emphasis will be placed on establishing the

boundary conditions of our theories through moderation analysis

(Fawcett and Waller 2011). Identifying limits and boundary con-

ditions improves the precision of our theorizing (Edwards and

Berry 2010), increases the validity of our findings (Leavitt et al.

2010), and helps ensure that we utilize nontautological theories

given that all falsifiable theories must have their limits (Popper

1959; Gray and Cooper 2010). Hall and Rosenthal (1991) assert

that “if we want to know how well we are doing in the biologi-

cal, psychological, and social sciences, an index that will serve

us well is how far we have advanced in our understanding of the

moderator variables of our field”(p. 447). Simply stated, the

exclusion of important moderator variables can lead to overgen-

eralizations that fail to illuminate the boundary conditions under

which purported relationships exist.

In order to leverage moderation analysis effectively, it is criti-

cal to understand the connection between moderation and mea-

surement of our theoretical constructs. Moderation and

measurement are tightly intertwined in that (a) measurement error

can result in extreme attenuation of statistical tests for modera-

tion when regression approaches are used (Aiken and West

1991) and (b) common method variance (CMV) severely deflates

interaction effects (Siemsen et al. 2010). In addition to illustrat-

ing how intimately moderation analysis is linked with measure-

ment, the objectives of this article are: (1) to assert the value of

moderation analysis and encourage a more prominent place in

our disciplines’survey-based research projects, (2) to provide

best practice approaches for using this type of analysis in pursuit

of greater depth and clarity in our research, and (3) to provide

some seeds for potential research projects that could benefit from

the use of this type of analysis.

MODERATION 101

Amoderator is any variable that affects the strength of a rela-

tionship between a predictor and an outcome variable. Imagine

one is looking at how a certain logistics practice influences firm

performance. In many instances, it is likely that extant theory

provides a logical rationale for why at least one variable, such as

the level of industry competition or characteristics of the prod-

ucts being sold, should affect the strength and/or direction of the

relationship between the focal logistics practice and firm perfor-

mance. In fact, some moderator variables may even lead to an

observed negative effect of the investigated logistics practices on

performance. Identifying such boundary conditions provides a

deeper perspective of the focal relationship and enriches our the-

ories (Fawcett and Waller 2011), and it helps researchers evalu-

ate the robustness of their results (Maloni and Carter 2006;

Goldsby and Autry 2011). As noted by Leavitt et al. (2010) “by

predicting interaction

1

effects…our confidence in a theory…is

bolstered”(p. 660). It is this deeper perspective that we, as

researchers, should be striving to provide to our constituents

(Knemeyer and Naylor 2011). Furthermore, moderation analysis

allows researchers to discover unanticipated contingencies

between variables, which can challenge commonly held beliefs,

one example of what Davis (1971) terms “interesting”research

Corresponding author:

Thomas J. Goldsby, Department of Marketing and Logistics, Fisher

College of Business, The Ohio State University, Fisher Hall, 2100 Neil

Avenue,Columbus, OH43210, USA;E-mail: goldsby_2@fisher.osu.edu

1

“Moderation”and “interaction”can be treated as synonyms

given that theorizing why the moderator variable changes the

strength and/or direction of the focal variable’s relationship with

the outcome implies that an interaction exists between the focal

predictor and moderator.

Journal of Business Logistics, 2013, 34(2): 109–116

© Council of Supply Chain Management Professionals