Comparison of alternative scoring techniques when assessing decision maker's multi-objective preferences in natural resource management

Abstract

A popular way to assess a decision maker's preferences in multi-objective natural resource management is to ask the decision maker to compare alternative management plans pairwisely in the ratio scale. Several numerical scoring techniques have been proposed for the ratio scale comparisons, but the performance of the alternative techniques is not fully understood. At the same time, the choice of scoring technique potentially impacts the description of the decision maker's preferences and therefore the actual management decisions. In this paper, the regression model for the ratio scale pairwise comparisons is used, and the differences between the alternative scoring techniques are studied based on two different viewpoints. The first idea is to interpret the scoring techniques as fixed models and then compare them under fixed numerical scaling. The applicability of several potential decision-making strategies is discussed in this context. The second idea is to parameterise the scoring techniques and then compare them under optimal scaling. This was possible in an artificial experiment, where the true values of the alternatives are known. The results supported the use of a geometric scoring technique. Also, the importance of assessing an appropriate scale parameter of the geometric progression is pointed out.