Shanteau begins “Psychological Characteristics of Expert Decision Makers” by citing quite a number of studies showing that experts are frequently both inaccurate (invalid results) and unreliable (imprecise results). Moreover, studies support the surprising findings that expertise is not enhanced by experience and that experts are frequently influenced by irrelevant considerations and frequently exclude relevant considerations. Even worse, a literature survey showed no findings to the contrary. The studies supporting these results come from a large number of diverse domains in which expertise is valuable: the paper mentions many of these.

Shanteau continues by listing 14 attributes of experts, some cognitive (perceptual and attentional abilities, differentiation between the relevant and the irrelevant, simplification of complexities, communication skills, knowledge of when to make an exception, selectivity in the choice of problems to tackle, adaptive decision strategies, knowledge of the subject area, creativity, automaticity of judgments, and inability to articulate their decision-making strategies) and the rest emotional (sense of responsibility for their judgments, projection of self-confidence, and tolerance of stress).

Each of these attributes is then discussed in detail. For example, an expert’s perceptual and attentional abilities allow him or her to recognize what novices do not even see. Simplification of complexities allows an expert to see things--like a chess board--in chunked or patterned terms. I would suggest that the creative and automatic aspects of expert judgment, both of which are difficult to explain, are right hemisphere functions [1]. This seems to be consistent with Shanteau’s descriptions as well as with research on the separate functions of the two hemispheres of the brain.

The paper continues with a discussion of the implications of these 14 attributes. Some experts are valued for their perceptual abilities (e.g., livestock judges), some for their purely cognitive abilities (e.g., auditors). Some are “substantive” experts who draw on a large knowledge base for their judgments (e.g., soil judges), others are “assessment” experts who must make inferences and predictions based on limited data (e.g., personnel selectors). Some, finally, are “passive” experts (e.g., accountants) who need not act on their findings, while others are “active” experts (e.g., business managers). The section concludes with a discussion of how experts might be trained in these 14 areas, and the difficulties of and alternatives to such training.

The final section makes an interesting point. Because experts are confident people who take responsibility, there is a large degree of unwillingness to work with expert systems. Some suggestions are made to lessen this resistance.

The paper is interesting, clearly written, comprehensive, and very well documented. It should interest knowledge engineers and, especially, cognitive scientists.