Decision Analysis Working Paper Abstract
Archive
WP020001
|
Title: Using a Bayesian Approach to Quantify Scale Compatibility
Bias
Authors: Richard M. Anderson and Benjamin F.
Hobbs, The Johns Hopkins University
Date:
Status: .Management Science, 48(12), Dec. 2002,
1555-1568
Weight selection is a crucial task in multiattribute decision analysis. Unfortunately,
the elicitation processes by which weights are obtained are susceptible not
only to random error but possibly to bias as well. We use Keeney and Raiffa’s
(1976) tradeoff weighting approach to elicit weight judgments from a group
of fisheries experts with management responsibility in the Lake Erie basin.
Then we use a Bayesian method to compute probability distributions of attribute
weights. In computing the Bayesian weights, our measurement model assumes
that the weight ratios produced by each respondent’s judgments are subject
to random error and an unknown scale compatibility bias. Ratios are log-transformed
and analyzed by a Bayesian linear model with a noninformative prior distribution.
Posterior distributions are then developed for the weights and the bias.
We estimate the compatibility bias for each person and in most cases it is
found to be significant and in the predicted direction, suggesting the importance
of its consideration in weight measurement models.
Our results also show that a simple heuristic procedure for assessing the
weights seems to be effective in eliminating the bias. Finally, the Bayesian
framework is shown to be useful for quantifying the value of additional information
about multiattribute weights.
Back
to the Decision Analysis Working Paper Index
