Step 2: Develop the Expert Model
As described earlier, an expert model summarizes and integrates expert knowledge on the topic of interest, typically using graphic decision-modeling representations, which provides an analytical framework for the design and analysis of later in-depth mental models research with key stakeholders. This framework facilitates later direct comparison between experts’ and stakeholders’ mental models of the topic. The model-creation process usually starts with a review of literature or relevant materials provided by the client, followed by informal, but in-depth, one-on-one interviews with a small number of experts or a workshop with a group of experts. Notes, recordings, or transcripts of the expert sessions are used as needed to support development of the expert model, though they are not typically formally coded and analyzed.
A common form of an expert model is an influence diagram, which represents knowledge in terms of variables and the relationships among them, as they relate to the outcomes of interest. The key influences or variables in the system are depicted as nodes, typically ovals, with descriptive titles and subtext and can depict a number of different types of influences, including the potential outcomes associated with the topic, the people whose decisions and behavior influence those outcomes, factors that “drive” or provide the basic context of the situation being studied, and the factors that influence these stakeholders’ decision making and behavior.
The nodes are positioned to create a logical flow, typically starting with driving or foundational contextual variables in the upper left corner and ultimate decisions or desired outcomes positioned in the lower right corner of the model. Relationships and primary direction of influence among the variables are illustrated with arrows with the node at the tail of an arrow “influencing” the variable at the arrow’s head. Additional graphic design concepts can be employed such as using color to semantically group similar nodes. This is particularly helpful in enabling users to “follow” nodes through various models that may present the variables at different levels of detail. Other techniques to represent expert knowledge may include logic models, decision trees, fault trees, and multiattribute matrices.
Expert models are created iteratively and final models are often validated by reviewing them with the experts who participated in their creation. Expert models can also be refined over the duration of a project to represent the most up-to-date knowledge on the topic as it evolves. Expert models provide the analytical framework for the design and analysis of in-depth stakeholder interviews, allowing direct comparison between experts’ and stakeholders’ mental models of the topic.
Expert models are typically not used as primary communications tools; however, they do provide a useful foundation upon which to engage experts and often key stakeholders in dialogue about the initiative. The system perspective enables all participants to understand all of the influences on the decision making of the focal stakeholders and to come to a shared understanding of the system drivers and desired outcomes.
