|
|
 |
|
Jane Booker with Los Alamos National Laboratory |
|
Date: October 5, 2004
Time: 10:30am
Location: NIA, Rm 404
Speaker: Jane Booker with the Los Alamos National Laboratory
Subject: "Eliciting and Analyzing Expert Knowledge: A Whirlwind Overview"
Additional Information: Webcast
The use of expert knowledge (also known as expert judgment or expert opinion) has origins in the 1970s in nuclear reactor probabilistic risk assessment (PRA) studies and decision analysis, where data were sparse or lacking. Since then, research, development and implementation of methods for eliciting and analyzing the knowledge and experience of experts have been formalized and applied in other areas such as reliability and certification. These methods draw from disciplines including cognitive psychology, decision analysis, statistics, sociology, cultural anthropology, and knowledge acquisition. This presentation is an overview elicitation guidelines, methods and techniques employing ethnographic techniques, bias minimization, knowledge acquisition, and formal elicitation procedures. Once knowledge is elicited, it must be combined with other information, involving the use of analysis and uncertainty characterization methods.
This presentation includes descriptions and philosophy of methods for eliciting and analyzing expert knowledge and provides guidelines and steps for how to carefully interview experts and analyze their knowledge. The formal methods add rigor, provide defensibility, and increase the ability to update the information as the state of knowledge changes.
While expert knowledge is not a replacement for test, experimental or observational data, it contains valuable information necessary for analysis, modeling, and decision making. Like data, it must be meticulously gathered and can be combined with other information or data. Also like data, it is subject to uncertainties, which must be estimated and propagated in aggregation and modeling. Other similarities and differences of expert knowledge and data are provided.
Research into methods for eliciting and analyzing expert knowledge is ongoing. Some of these new areas include information integration methodology development for combining expert knowledge with other data/information, and the use of general information theories for uncertainty quantification.
|
|
|
Home