Dialogue processes for generating decision alternatives

The practice of Decision Analysis (DA) rests on two pillars: mathematical modeling, and facilitated dialogue. While the mathematical modeling techniques are based on a well developed, normative theory, the dialogue facilitation techniques, in comparison, are not. To help address this disparity, I model a particularly critical phase of DA dialogue---preliminary conversations in which the analyst frames the decision and develops an initial set of decision alternatives. Based on discourse analysis, field observations, and literature reviews, I identify dialogue processes that are pivotal in such conversations.; Based on the dialogue processes I identify, and drawing on prior work in dialogue-game theory as well as general morphological analysis, I develop a computational dialogue model in which participants introduce, develop, and evaluate decision alternatives. This computational model constitutes a scheme for encoding the flow of real-world dialogues, and provides a basis for graphically visualizing important aspects of dialogue processes. Analysis of the interaction of several teams performing a conceptual design task shows this model captures key aspects of team interaction related to effective decision-making. Analysis also shows that the effectiveness of these teams can be statistically predicted based on quantitative, model-based metrics of their dialogue processes.; The dialogue processes captured in the model provide the underpinnings of a "dialogue tool kit" consisting of schemas and methods that can help DA practitioners develop a more satisfactory set of decision alternatives by recognizing and eliciting key aspects of a decision-frame, challenging and shifting the frame, and expanding the range of prospects considered prior to formal decision modeling.