A structured approach to extract strategic objective categories from textual sources

Abstract

This paper employs a structured approach to extract strategic decision maker objectives from textual sources as part of a problem structuring activity for constructing decision-aiding models based on historical decisions. It focuses on textual sources as access to original decision makers cannot be guaranteed. The approach combines multiple methods to first identify a comprehensive set of possible decision maker objectives for historical decisions and then map them to a smaller and more manageable set of objective categories. The usefulness and consistency of these objective categories is tested by applying them to an analysis of Australian post-cold war strategic decisions regarding the employment of military force. These objective categories provide an auditable and useful means to structure interviews with decision makers that will inform the modelling process. Objectives were first identified in primary source texts using a method based on directed content analysis (DCA). These objectives were then vetted for possible redundancy using a similarity measure based on Euclidean distance. Next these objectives were structured using value focused thinking into a manageable set of objective categories. Finally, the usefulness and consistency of these objective categories was assessed by comparing analyses using these categories of two Australian strategic decisions using two data sources. This initial problem structuring activity builds on a previous analysis of strategic decisions on East Timor during 1999 (Coutts, 2010) and will be followed by more detailed studies that will include interviews that lead to detailed modelling of the objectives, influences and reasoning behind these decisions. Such decisions are studied by analysts in order to understand the implications of different strategies, improve strategic models and thus better inform future decisions (Auerswald, 2004; Fredrickson and Mitchell, 1984). Employing a structured approach to study these decisions will complement existing research conducted through the political science and international relations disciplines through the concept of methodological triangulation (Webb et al., 1966). However, challenges exist in structuring a study of historical strategic decisions. When structuring a current decision support problem, analysts can rely on reasonable access to stakeholders and/or decision makers due to the perceived immediacy of the benefits of participation. However such access is often more constrained for historical decisions where there is no immediate pressure on the decision maker to participate and instead the analytical challenge is to structure the study through other means. Of particular concern is the need to structure, and provide rigour to, the design of rare interviews with busy strategic decision makers. Hence the intent of this paper is to produce a manageable and defendable set of objective categories to inform interview design.