Differences between expert and novice decision making as a function of presentation and elicitation modes

Abstract

The experiments described in this thesis examined the effects of presentation mode on students and staff making decisions, assuming that these groups represented experts and novices in the subject area of the decisions.
A pilot study provided suggestive evidence that it would be worth looking at hypothesis selection. Experiment 1 was designed to explore this issue. Subjects were asked to rate the likelihood of a number of categories, given some identifying features, using a computer-based design. Experts selected more hypotheses than novices in 3 different problems.
Experiment 3, presented equivalent problems to subjects, but used a paper-based format. Surprisingly the results were very different from Experiment 1, with all subjects having more hypotheses. This affected novices more than experts, they had more hypotheses than the experts.
Subsequent experiments attempted to isolate the factors responsible for this difference. Although the whole effect was not explained, it is clear that the subtle difference of the prior status of the hypotheses is important. In the original computer based design subjects were required to select which hypotheses were possible, and then to rate those possibilities. In the paper-based design, these two steps took place at the same time, so that they rated each hypothesis, marking it in a special way if it was ruled out. This difference was shown to have an effect. It is interesting that it should have a greater effect on novices than experts, they might be more susceptible to bias involved in the presentation medium.
The discussion elaborates on these issues, questioning the assumption that the subject groups did represent expertise groups. Alternative explanations for the presentation effects are also discussed. These issues are relevant to the field of knowledge elicitation. If minor changes in presentation mode can alter the responses of participants, these effects should be studied and understood.