The initial aims of the project were to identify characteristics of the command process at Higher Command levels in the Royal Navy so that systems design could include decision aids at this important focal point"in the overall system (a naval Task Force). A prerequisite for formulating recommendations is an understanding "of the Task Force Commander's role and decision process. Consequently, an attempt was made early on to structure the Task Force command task. It became apparent that, without special measures, such a description could not be acquired. This thesis, therefore, is the story of a project about collecting data and informa tion, using it and then interpret ting it for the aboye purposes. Since there was little known about how to achieve an understanding of senior decision making, especially in potentially unstructured areas like the naval Higher Command function in a conflict environment, appropriate measures were developed to do so. The overall methodology designed consisted of: a) basic research; b) interview techniques; c) scientific gaming procedure; and d) a data collection-in-action regime. The methodology sought to use 'laboratory' techniques initially to acquire enough insight to then mount a study of clients "in action". It was considered necessary to supplement the former types of methods (interviews and gaming activities) with versions actually involving the clients when performing their role (in the form of a structured self-report study). The overall methodology was rooted in the systems ideas of Checkland (1981) and Bowen's views on the OR process (1981 and 1984) . Both authors react against the traditional prescriptive, normative approach of text book OR practitioners. The latter pursui t ignores messy, human aspects of organisational life and requires that a problem situation is well understood so that applied mathematical techniques can be used to formulate and then , solve' a problem. Most high-level decision problems, though, cannot be so formulated. This thesis is an attempt to formulate and understand high-level decision problems in a different way, using different techniques, but with a similar aim of arriving at useful and meaningful decision support recommendations. The methods to be described should start to fill the gap that exists at the moment in the OR repertoire of methodology catering for such needs. It was not possible to implement the overall methodology in the naval context. The industrial phase of the research was therefore initiated so that all of the phases could be tested in an albeit limited programme. Some preliminary insights and findings emerged in both contexts. Essentially, high-level decision makers approach their tasks differently. They have differing priorities also - as revealed by the interview and gaming phases. However, the data collection-in-action study (implemented in an industrial context only) revealed that what happens in reality (when they are at work) does not fully reflect the pattern of priorities revealed in the laboratory studies. Consequently, it is necessary to use other means to arrive at a complete picture of their decision making process. The overall methodology includes the interview and gaming phases because they are necessary to acquire enough insights to mount a data collection-in-action study later on. Also, they have other research and training uses. The ~hesis also includes the use of an analysis technique (based on Hogberg 1985) which assists with the appreciation of high-level decision making problems. The technique forms the basis of a proposed decision support system for both military and industrial contexts.