Behavioral Dynamics in Operations Management (BDOM)

Over the last several years there has been growing interest in the incorporation of behavioral theory and dynamics into common discussions of operations management. Such interest has been demonstrated by a growing number of operations management publications in high ranking non-traditionally OM outlets (eg. Journal of Applied Psychology), an increase in the publication of behaviorally focused work in OM outlets and an increase in special seminar series devoted to this interface (eg. Harvard's Behavioral Operations series, PennState's Behavioral Research in Operations conference, etc.).

What is obvious is the wide range of viewpoints as to the direction that new research aimed at merging elements of these two traditionally divergent fields should be. For example, should an emphasis be on incorporating artificial characterizations of human behavior that can be incorporated as explicit components of OR models, or should emphasis be placed on empirically investigating the interplay between OM recommendations/policies on subsequent human behavior? Most are of the opinion that there is no single 'correct' approach, but that a variety of approaches should be taken up in tandem.

A forthcoming article in the Journal of Operations Management, "Behavior in Operations Management: Assessing Recent Findings and Revisiting Old Assumptions", provides a framework for considering the application of behavioral theory towards behavioral-assumption clarifications across a variety of traditional operations management contexts from project management to procurement. It also emphasizes the severe gaps that still remain in the literature at this interface. The inadequacy of support for existing assumptions, and the attempts of models to nevertheless make claims based on such limitations remains problematic in terms of the true practicality of such models. This weakness is echoed in Sterman's prize winning lecture titled poignantly "All Models are Wrong".

Behavioral research in OM is certainly not new (as exemplified by Hill's 1982 work on scheduling heuristics), yet it remains underexamined and underutilized to the detriment of the entire OM community. To help further facilitate research interest at the Human Behavior and Operations interface, and in an attempt to appeal to both OR modelers and OM empiricists, the following integrated roadmap for exploration is posited:

Behavioral Explication / Refinement
Theorize and empirically test for the codifiable nature of specific aspects of human behavior (which are otherwise simply assumed by OR models) in operational work settings. Such explication will by necessity involve some level of empirical work, with the knowledge of existing math-theoretic assumptions in OR models. Ideally such work would involve the joint efforts of OM empiricists guided by OR modelers.

Robustness Confirmation / Validation
Confirm both existing and recent explications / codifications of human behavior (as applied in OR models) through retesting. Such work will not only ensure that the behavior can in fact be better modelled by specific codified forms than by alternate assumptions (which may have neglected its consideration entirely), but it will also allow for testing of the robustness of such explications across various OM settings and work scenarios. Mainly empirically focused work.

Explicit Integration of Behavior
Integrate empirically confirmed codified models into OR models to test for major deviations from past math-modelling findings. Identify new nuances or other areas for neeeded empirical clarification. Such an effort will likely be best promoted by OR modelers guided by the sincere interest of applying realistic codifications of behavior to better inform the operations community. It's findings will likely lead to further work in Behavioral Explication.

Mechanistic Enrichment
Integrate math-theoretic dynamics in a predictive sense with OB models of performance and subsequent empirical work (almost entirely based on assumed linear relationships of causality and association at this point). As with Behavior Explication this will necesarilly require joint efforts by both OR modelers familiar with mechanistic phenomena and empirical researchers familiar with both empirical methods and micro-behavioral theory. Such work would actually help clarify observed dynamics and filter out extraneous non-behavioral issues. It would ultimately lend to the sharpening of codifications of behavior.

John Sterman (MIT)