Preference-neutral attribute weights in the journal-ranking problem

Abstract

A linear programming model is proposed for assigning linear attribute weights in the journal-ranking problem. The constraints in the model are derived solely from any quasi-dominance relations that can be established between the journals. The objective function of the model minimizes the maximum difference between the implied valuations for the pair of journals that define a constraint. In the sense that personal inputs are not introduced, the derived weights are preference neutral. The feasibility of the procedure is demonstrated for two sets of data. By considering various random samples of journals from the larger data set, it is shown that large differences can emerge in the attribute weights and in the journal rankings from different samples of journals, even when the sample sizes are large relative to the population size.

Keywords:

• education
• performance measurement
• linear programming
• multi-objective

Acknowledgements

The first draft of this paper was written during my tenure as Adjunct Professor in the IDS Department at San Diego State University, where I enjoyed the considerable support and collegiality of the department and of the College of Business Administration faculty. The paper was revised into the present draft while I served as Visiting Professor in the EF Department of the City University of Hong Kong. Earlier financial support for the project in the form of a Summer Research Grant from the Warrington College of Business Administration of the University of Florida is also gratefully acknowledged. Mr. Kerim Yalim provided invaluable computer programming and computational assistance, and Professor Joan Donohue was kind enough to make available to me the appropriate formulae for computing the attribute scores that are used in the primary illustration. I, however, bear sole responsibility for the product that has resulted from all of this help.