Describing the Representation of Decision Problems: An Application of Multidimensional Scaling and Cluster Analysis.
Abstract
This study describes important representations for an example of a common class of decision problems--facing a shortage of a commodity. Describing potential problem representations is important because decision problems are typically 'ill-structured,' and a decision maker's representation of a problem is not obvious to the experimenter. Describing the dimensions along which a group of subjects judged the similarity of potential solutions to a problem should give insight into various ways in which the problem may be represented. Multidimensional scaling and cluster analysis were used to analyze the similarity of 43 acts suggested to solve the parking problem at the University of Oklahoma. Hierarchical cluster analysis was used to analyze the similarity judgments to examine neighborhoods of acts in the three-dimensional space to determine whether an alternative interpretation of the relationships between acts might be obtained. Seven clusters were identified. Four clusters were specific instances of a more general category 'increase the amount of space available'. Another cluster was the category 'involves alternate forms of transportation'. Two other clusters involved rescheduling activities and enforcing current parking regulations more strictly. The three dimensions derived from multidimensional scaling and the set of clusters obtained from cluster analysis seem to describe alternative strategies for solving the parking problem from which individual decision makers might sample when representing the problem.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 15, 1981
- Accession Number
- ADA110175
Entities
People
- Carol A. Manning
Organizations
- University of Oklahoma