Printer Friendly

The usefulness of behavioral laboratory experiments in supply chain management research.


The advent of the field of behavioral operations has brought behavioral laboratory experiments into the mainstream of methodologies used in the field of supply chain management. The term behavioral refers to the explicit study of human behavior and decision making in relation to what is rational. Laboratory experiments, as opposed to field experiments, involve the use of simulated treatments in an artificial environment. This article briefly discusses the usefulness of this methodology.

Behavioral laboratory experiments can be classified as an empirical method since they rely on the observation of behavior and on the analysis of data. As such, they are a method of testing theory. To some extent, they also serve as a framework for generating observations that can lead to further theoretical developments. Within the spectrum of empirical methods, they emphasize internal over external validity, i.e., they allow for a stricter correspondence between treatment and theoretical context while often sacrificing real-world representativeness of the underlying phenomenon (Campbell 1957). Behavioral experiments thereby serve as one element within the process of triangulation (Boyer and Swink 2008). Results from experiments do not prove a theory, but, when interpreted together with surveys, field studies and archival research, they further strengthen (or weaken) the confidence we can have in the validity of a theory. The distinctive advantages of behavioral laboratory experiments can be summarized under the terms of control, efficiency and responsiveness.

The key advantage of experiments is their control (Highhouse 2009). Any empirical test of a causal relationship needs to assess the statistical correlation among independent and dependent variables, the temporal precedence among these variables, and exclude the possibility of spurious relationships. Experiments allow for the precise temporal sequencing of treatment and effects, and also allow for the randomization of spurious causes such that they do not correlate with the treatment. They also allow for the effective use of control treatments, such as placebos, that create an adequate reference group to measure substantive effects. Further, since the context of an experiment can be precisely described and influenced by the experimenter, experiments can lead to very clear theoretical predictions and the precise establishment of a normative benchmark. Other empirical methods, while having more external validity, often lack some of these qualities. Survey and archival studies can never completely exclude the possibility of omitted variables. Field experiments rarely allow researchers to observe the change in only one variable at a time. Field studies seldom allow a complete enough characterization of a decision environment to clearly establish behavior as biased or rational.

Laboratory experiments are also highly efficient. While they require institutional investments, such as the construction of behavioral laboratory space and the assembly and maintenance of a standardized subject pool, the marginal cost of running a laboratory experiment is low (although there certainly are exceptions, such as experiments involving f-MRI imaging) compared with the cost involved in buying large databases or collecting survey data from multiple companies. Further, laboratory experiments require relatively little time to run compared with the time it takes to complete a large-scale survey data collection. These two attributes, low cost and high speed, make experiments a less risky approach to research, and a less risky route to take for doctoral students in their path to graduation.

This efficiency creates responsiveness, i.e., the ability to utilize a feedback learning cycle in research. Many other empirical methods, such as survey or archival research, are slow, expensive or sometimes even unique. Surveys can take a long time to come to fruition, and are often limited in their population. Archival datasets often exist only once, and a similar dataset cannot be obtained. This requires that researchers adhere with strict discipline to fully specifying theory before running empirical tests. If theoretical results do not work out as predicted, researchers who employ a survey method are very limited in what they can do as a result. If they further explore the reasons why a particular hypothesis was not supported within their data, technically, they cannot use the same data to re-test a modified version of their theory. In experiments, however, researchers can more easily learn from their experiments, and design and conduct a new experiment as a result. This enables a better feedback/learning approach to empirical research. As Smith (1976, p. 274) writes: "(...) one develops a model, which is then tested with the only body of field data that exists. The results of the test turn out to be ambiguous or call for improvements, and one is tempted to now modify the model in ways suggested by the data to improve the fit. Any test of significance now becomes hopelessly confused if one attempts to apply it to the same data. (...) The fact that one can always run a new experiment means that it is never tautological to modify the model in ways suggested by the results of the last experiment."

In summary, there are good reasons why we should welcome behavioral laboratory experiments as one valid research method into the repertoire used in supply chain management research. Criticisms against the method are often directed at the simulated context of experiments (i.e., they are far removed from the reality of modern organizations) or the nonrepresentative type of participants used in these experiments (i.e., undergraduate students are not supply chain managers). Such criticisms can be addressed by either simulating what is relevant about the organizational context within the experiment, or by recruiting more relevant participants (i.e., MBA students, managers) if one believes that the underlying psychological mechanisms require it. From a broader perspective, behavioral experiments are rarely the final step in the process of triangulation, but rather serve as an early and efficient test of theoretical ideas that can lead to a refinement and strengthening of theory.


Boyer, K. and M. Swink. "Empirical Elephants--Why Multiple Methods Are Essential to Quality Research in Operations and Supply Chain Management," Journal of Operations Management, (26), 2008, pp. 337-348.

Campbell, D.T. "Factors Relevant to the Validity of Experiments in Social Settings," Psychological Bulletin, (54:4), 1957, pp. 297-312.

Highhouse, S. "Designing Experiments that Generalize," Organizational Research Methods, (12:3), 2009, pp. 554-566.

Smith, V.L. "Experimental Economics: Induced Value Theory," The American Economic Review, (66:2), 1976, pp. 274-279.


University of Minnesota

Enno Siemsen (Ph.D., University of North Carolina at Chapel Hill) is an assistant professor in the Carlson School of Management at the University of Minnesota in Minneapolis, Minnesota. His research focuses on process improvement, new product development and forecasting. Dr. Siemsen's work has been published in many peer-reviewed publications including Management Science, Manufacturing and Service Operations Management, and the Journal of Operations Management.
COPYRIGHT 2011 National Association of Purchasing Management, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2011 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Siemsen, Enno
Publication:Journal of Supply Chain Management
Article Type:Report
Geographic Code:1USA
Date:Jul 1, 2011
Previous Article:The vignette in a scenario-based role-playing experiment.
Next Article:Questions to consider when selecting student samples.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters