Date of Award


Document Type

Campus Access Dissertation


Moore School of Business

First Advisor

Manoj Malhotra


The economic and social costs of operational failures have become staggering. Reducing the frequency and severity of operational failures is a potential source of significant competitive advantage. However, extant research related to operational failures, frequently termed as accident research, is predominantly found in other disciplines and literature streams. It is also mostly associated with major disasters and employs case based methodology. Therefore many authors have called for new empirical research related to causality of less severe, routine operational failures in order to validate and extend accident theory. This dissertation begins to extend current theory and understanding through evaluating extant accident causality literature from an Operations Management (OM) perspective, validating theoretical postulates emerging from that literature, and empirically examining OM causal factor relationships that may exist in routinely occurring adverse operational events.

In the first essay, a generic adverse event model is developed and presented. It examines both OM triggering factors, as well as their human error antecedents. Using data from the US Department of Energy, the significance of these factors and their relative influence are analyzed for two different severity levels and three different types of human error. Results show that human errors serve as significant causal factors in over 70% of all adverse operational events. In addition, events triggered by human error tend to be more severe than events triggered by equipment failures or external phenomena such as storms and power outages. We also show that higher level cognitive errors in the skill-rule-knowledge error taxonomy are positively related to event severity.

In Essay 2, we explore the antecedents of human errors in greater detail. We find that the consensus OM antecedents of human error have differing levels of influence on different types of errors. In particular, management and engineering factors play a multi-dimensional role in how they influence human error. The breath of their influence extends from the design of the operations system, its assets and its processes to its direction, control, and change. In addition, other production as well as support and linking processes are dependent on information from the engineering and management functions.

Building on the results of Essays 1 and 2, we evaluate in Essay 3 the effects of five sub-dimensions of the management and engineering antecedent factors on different forms of errors. These analyses reveal important differences in the role played by antecedent factors in driving operational failures. For example, issues with design documentation and change management appear to significantly influence rule and knowledge based errors, but not skill based errors. Training problems also appear to be more influential in causing rule based errors than skill or knowledge based errors. These differences are likely driven by differences in the level and type of information and decision support needed by workers involved in higher levels of cognitive human performance.

Collectively, the results in both essays provide important empirical validation and fine grained interpretation of consensus beliefs emerging from years of disaster research in multiple disciplines. They also significantly extend theory by demonstrating the relative strengths of consensus factors that influence severity and human error in operational failures.