Mental Models Theory and Military Decision-Marking: A Pilot Experimental Model

2003 
Abstract : This paper examines a recent theory on human inductive reasoning termed Mental Model theory. The theory predicts problems for decision makers arising from basic cognitive processes. For example, theorists argue that subtle cognitive processes described by mental models may have been instrumental in some catastrophic decisions in industry (e.g., the Chernobyl nuclear reactor meltdown) and in the military (e.g., the USS Vincennes incident). In particular, construction of the mental models used when making critical decisions is vulnerable to both problem complexity and logically conflicting (false) information. These vulnerabilities occur because of the limited capacity of human working memory. The capacity limitation restricts both the type of information that is represented in mental models and the number of component models. This project aimed to further investigate Mental Model theory for the benefit of future experimental studies in applied Military settings and the design of decision support tools. An experiment was conducted to test the central predictions of the theory. Twelve subjects (6 military, 6 nonmilitary) were presented with a set of 16 problems designed to manipulate complexity, the problem domain, and the requirement for false information. Comparisons also were made between the performance of the military and nonmilitary groups. A comparison of the number of problems answered correctly showed that the vulnerability to premise falsity extended to the military domain for both military and nonmilitary subjects. Problem complexity, however, was found to have little impact on the number of correct responses. It also was found that subjects responded more quickly when their logic required false premises and more slowly when the question addressed was complex. (6 figures, 9 refs.)
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    6
    References
    2
    Citations
    NaN
    KQI
    []