Abstract
This research aimed to explore and support the reason-based decision making processes of experts and novices in a complex and uncertain domain: resolving labour disputes. Naturalistic Decision Making (NDM) has investigated the role of expertise in complex and uncertain domains that are often time pressured. NDM models typically focus on fast decisions while explaining the reasoning processes behind slower decisions less well. There is much research on expertise, experts’ reasoning on complex problems is less well understood. Therefore, this research aimed to look at experts’ reasoning in slower, reason-based decisions. The first empirical chapter examined how complex labour judgements were made by testing a Mental Model Theory (MMT) of probabilistic reasoning. This was followed by a second empirical chapter, in which participants’ (labour officers) thought processes were elicited using a think aloud protocol. Based on these findings, the thesis then progressed to develop a reasoning aid to support reasoning followed by an evaluation of any changes in reasoning processes and outcomes in the third empirical chapter. The final empirical chapter validated the efficiency of the reasoning aid. Six scenarios were developed to replicate typical labour cases and used in studies to assess reasoning processes on a realistic task. Participants for each study numbered 42, 22, 28 and 82 respectively. The data for Study 1 and 4 were analysed quantitatively, and the verbal protocols for Study 2 and 3 were analysed qualitatively. Verbal protocols were recorded and transcribed, then transcripts were coded based on participants’ reasoning processes. Differences between experienced and less-experienced officers were also tested. Study 1 provided mixed evidence of reasoning according to MMT, finding that experienced and less-experienced officers were not significantly different. In Study 2 the data were analysed using six higher-order codes proposed by Toulmin et al. (1979) and each protocol was drawn into an argument map. This showed that experienced officers drew more accurate conclusions, omitted less evidence and offered more justifications than less-experienced officers. The reasoning aid used in Study 3 improved less-experience officers’ reasoning such that conclusion accuracy was the same as that of experienced officers. However, Study 4 revealed that, while the reasoning aid had no impact on the reasoning processes, the level of experience had a significant effect. This research provides a good description of participants’ reason-based decision making. Toulmin's argument analysis approach provides a unique contribution to understanding reasoning in this realistic and complex task. Although, the reasoning aid reduces the differences between experienced and less-experienced officers, experience still plays a crucial role in ensuring correct outcomes.