One of the most significant hazards workers face in the open cut mining sector is the potential for mining vehicle accidents; vehicle collisions result in 10 to 20 deaths annually within Australia. Mining vehicle interactions rely on operator decision making with no higher order controls to manage hazards associated with accidents. Hazards relate to the size and visibility of the vehicles, as well as the time they need to move around each other. Based on existing naturalistic decision making (NDM) theory, the research proposed two questions to explore the relationship between shared operator decision making and safe vehicle interactions:
1. How do operators make decisions during vehicle interactions?
2. How can shared operator decision making be influenced for safer vehicle interactions?
An inductive approach was used to investigate the phenomenon which provided deep insight into individual operator decision making, by comparing individual operator decision making, and shared operator decision making. As a multimethod approach, naturalistic observations of mining vehicle operators and analysis of historical data sets (e.g. incident statements, historical focus group data and company procedures) were used to understand both normal interactions and accidents.
Answering the first research question, the recognition primed decision (RPD) model was used as a theoretical frame to describe operator decision making for each vehicle interaction as the various dimensions of the model aligned closely with aspects of mining vehicle interactions. This study found that normal interactions occur when operators share decision making, linked to the same understanding of cues, interaction patterns and mental simulations, resulting in expected operator action scripts. Vehicle collisions are more likely to result when there is a breakdown of shared decision making, resulting in vehicle collision pathways which are either not recognised by one or both operators, or recognised too late for adjustment as operators run out of time to take evasive action.
Answering the second research question, thematic analysis extended current NDM literature, identifying the four themes which influence shared decision making:
1. The timing of cues is important, however, as most interaction cues are passive, they may go unnoticed. Where cues are not adequately specified, there is a greater likelihood that operators will use the wrong cues to make only partially informed decisions. This emphasizes the need to ensure cues are contextually simple, salient and timely (e.g. specific light signals, defining the boundaries of work areas, triggers for the use of radio communication).
2. Collectively, if pairs of interacting operators imagine the same patterns from the cues they notice, it is more likely they share mental models. This emphasizes the need for operators to develop homogeneity of expert mental models, an understanding of common tasks and communication (e.g. training which reinforces routine interaction patterns).
3. Uncertainty reduces operators' confidence in a situation, driving them to imagine how interaction sequences may go wrong before they act, so they may slow down to obtain more information. This emphasizes the need for operators to agree on how a shared goal will allow the coordination of interdependent activities (e.g. formal and informal discussion on the meaning of cues and patterns).
4. As the complexity of interactions increase it is harder for operators to maintain cue-action relationships. Misunderstandings are problematic when operators are working closely together, and space and time are limited. This emphasizes the need to increase the adaptive capacity of situations to make them more resilient (e.g. by identifying and controlling 'choke points' by limiting vehicle speed, or the number of vehicles in an area).
Contrasting current approaches within industry, the research demonstrates that the position of social constructionism provides different insights to that of positivism. Specifically, the research makes a number of contributions to current NDM literature by investigating shared decision making. Contrary to the current NDM research, which tends to assume that interacting decision makers act as individuals in isolation, the research shows that four additional theoretical frameworks are required to understand how decision makers collaborate to achieve common goals.
1. Team situational awareness provides a stronger position, highlighting the need for shared displays which provide decision makers with the right information from their own perspectives.
2. Analysis implies that decision makers in a common environment largely have shared mental models, but that the alignment of mental models in a temporal sequence can result in misunderstandings.
3. Sensemaking is more complex in shared decision making and requires decision makers to constantly search for uncertainty, cross-checking each other's actions to ensure they have the same goal.
4. Human error is inherent to shared decision making, resulting from the need to make approximate adjustments toward common goals. The environment in which shared decision making takes place requires an appropriate degree of adaptive capacity.
The research includes three limitations. Firstly, the interpretation of data was restricted to the researcher's perspective of naturalistic observations and historical document analysis. Secondly, social reactivity may have influenced operator behaviours during naturalistic observations. Thirdly, the accuracy of the organisation's self-reporting. The research included reflexivity, member checking and triangulation to reduce the risk of these limitations.
The research offers considerable insight into current NDM literature, finding that additional prerequisites must be satisfied for decision makers to successfully achieve common goals. The research provides theoretical and practical contributions for both researchers and practitioners to positively influence shared decision making in high risk operational environments.
ii Declaration I certify that except where due acknowledgement has been made, the work is that of the author alone; the work has not been submitted previously, in whole or in part, to qualify for any other academic award; the content of the thesis is the result of work which has been carried out since the official commencement date of the approved research program; any editorial work, paid or unpaid, carried out by a third party is acknowledged; and, ethics procedures and guidelines have been followed.