“…Discussions of these methods in the published literature emphasize the importance-and difficulty-of identifying true experts who are able to effectively articulate the knowledge requirements and solution strategies associated with task performance, (cf., Hall, Gott & Pokomy, 1995;Hoffman, Shadbolt, Burton & Klein, 1995;Crandall, Klein, Militello & Wolf, 1994) As noted by RSG-27, the more effective methods of cognitive task analysis organize subsequent interviews and data collection around a knowledge representation framework that is appropriate for the specific task. Here, a variety of approaches can be employed, including the use of annotated goal-method graphs (DuBois & Shalin, 1995), precept-action-result-interpretation structures (Hall, Gott & Pokorny, 1995), mental models of tasks and their context (Crandall, Klein, Militelo & Wolf, 1994), procedural and conceptual knowledge ontologies (Benysh, Koubek & Calvez), and declarative and procedural knowledge ontologies (Williams, Hultman & Graesser, 1998). In addition to the methods employed during the preliminary phase, more formal methods of knowledge elicitation described in the literature include structured interviews, controlled observation of task performance, verbal protocol analysis (thinking out loud), withholding specific information to assess its impact, formal decomposition of critical incidents, and psychological scaling that employs multivariate statistical analysis of pair-wise comparisons.…”