Comparing Conceptual Structures: Consensus, Conflict, Correspondence and Contrast

Shaw & Gaines (1989) Knowledge Acquisition, 1, 341-363.

Article Summary

The authors propose the application of this knowledge acquisition model to domains where objective knowledge is not yet available. I wondered if a new or young domain, such as Human Factors, would be considered one in which objective knowledge is not yet readily available. The Human Factors discipline seems to be in a dynamic state of self-definition, and refinement. However, upon further reflection, I believe that almost any domain could benefit from further definition and refinement. If the primary sources of knowledge are the conceptual structures of individual experts in a new, or undefined domain, then in older and better defined domains/disciplines, the conceptual structures of experts could also yield interesting insight and definitions of the different constructed realities of a discipline.

The possibility of multiple experts in a domain either coming to consensus, conflict, correspondence and/or contrast about terminology and concepts is one that is becoming well known to me. Having taken numerous courses, from several different professors in Educational Psychology, on the topic of research design and methodology, one begins to appreciate and compare the various different viewpoints and perspectives that exist in one domain area.

The authors fail to operationally define what they consider to be an expert in a discipline. How does one identify an expert? Is it a person who has cognitive competence? One who can automaticaaly do things that non-experts cannot do? Or instead, is an expert one to whom many have attributed cognitive competence? This fits in with consensus theory, and expertise as an attribution. Is an expert one who is regarded as such by others (constructed reality)? And, upon what does the willingness of others to attribute expertise depend?

In a consideration of the implications of this model for the design of expert systems, will novices have to have some sort of extended knowledge base in order to benefit from an expert system? Will the expert system have utility for both a novice from within and from without the discipline area? Is this a consideration for expert system design? For example, would one have to be a novice with some medical background in order to derive useful information from a medical diagnosis expert system? Or, would such an expert system have utility for a laymen? The authors summarize the importance of addressing these questions with the recognition of possible conflicts between the experts and clients use of terminology, and the provision of a variety of corresponding concepts to improve the usability of a system.

In a discussion of conceptual systems, the authors refer to Kellys Personal Construct Psychology, and the notion of shared concepts and conceptual systems that experts may or may not have in common. The notion of conceptual systems that are so widely shared and so significant they they are construed as knowledge, seems to be related to Lincoln & Gubas (1985) consensus theory. This knowledge is treated as having an existence virtually independent of their carriers.

Exchange methodologies were developed to measure understanding and agreement either between two individuals, two roles, and/or two occasions. This methodology allows experts who may have differing points of view to share knowledge and experience using their own individual terminologies.

Socio Analysis is used to compare different conceptual systems in a domain. Different conceptual systems are compared for structure, and for similarities and differences. A socio analysis is a simple form of analogical reasoning.

The step by step description of the three phase process of using this knowledge elicitation model is quite helpful in synthesizing the parts of this article into a whole.

  • Step 1: Identify Domain (ex: attributes of an expert IT teacher).
  • Step 2: Select three experts (cognitive competence, or attributed competence)
  • Step 3: Group experts, and brainstorm a set of entities. (Or, individually, use WebGrid).
  • Step 4: Extract elicited entities from 3 grids for discussion and consolidation by group (ex: attributes of an expert IT teacher).
  • Step 5: Experts individually elicit attributes & values for the agreed upon entities.
  • Step 6: Exchange and Compare.
  • Step 7: Analyze Consensus and Conflict.

  • Return to CPSC 679: Michele Jacobsen's Stuff