consortiums

PROJECT 6

SELECTION AND CLASSIFICATION

José H. Dulá ,
School of Business Administration
University of Mississippi

Donna Retzlaff-Roberts
University of Memphis

Proposed Grant $85,939

Abstract:
Evaluation of military personnel is a difficult task. Their training, preparation, education, experience, and skills, frequently make them unique and resistant to classification. Evaluations based on fixed inflexible performance criteria are likely to bog down

in disagreements about appropriate number of attributes to include and the relative merits of one attribute over another. For example, too many criteria and the model will not be discriminating enough; but as criteria are reduced the problem arises as to what to include and what to exclude to make the model discriminating and equitable. As attributes are culled, the scheme becomes increasingly less acceptable to more and more people. Also, such schemes may become counterproductive since the individuals have the capacity to understand and adapt to these criteria making themselves, in turn, inflexible and less effective to the organization.

This project proposes a new methodology for evaluation based on nonparametric frontier analysis to evaluate and compare a group of entities sharing a common list of performance criteria. The methodology is based on Data Envelopment Analysis (DEA) but is much more in-depth and comprehensive. The approach we propose involves numerous DEA analyses for multiple subsets of the performance criteria. The information that emerges from such a study is richer and more complex than from traditional DEA. Additionally, assessments and classifications about entities’ performance are more accurate and robust and the role of the performance criteria will be better understood. It is ideally suited in evaluations of complex, sophisticated, entities such as highly trained and specialized professionals, units, or ships within a specific category. Today's computing technology makes this methodology feasible.