Factor Analysis and Methods of Supplier Selection

Abstract

We discuss in this paper the decision making in choosing the best alternative from some available options based on possibly a large number of selection criteria. This multi-criteria decision problem typically arises in supplier selection in supply chain management. Recently, there has been an increasing interest in the applications of dimensional reduction methods such as factor analysis to such decision processes. They have been widely applied in conjunction with some classical methods such as AHP to create a hierarchical structure and identify the underlying factors or constructs. There are, however, a number of inherent issues and difficulties which have not been adequately addressed in the literature. For instance, there may be some criteria which load significantly on more than one factor, creating considerable difficulties in categorizing the criteria into mutually exclusive groups. More importantly, it is seen in this paper that it is not always sensible to determine the importance of an identified factor according to its amount of shared common variance or explained variation. Similarly, attempts to routinely determine the local relative weight (within a factor) of importance of a criterion based on its factor loading or correlation with the factor may also lead to results markedly different from those based on the views or judgement of the practitioner or expert. To circumvent these difficulties, a simple, practical and easily implemented procedure is proposed. Although factor analysis is employed, it merely serves as a means of facilitating the direct rating of importance of each criterion, alleviating many of the difficulties of the classical factor analysis approach. Two examples are given to illustrate the proposed method and illustrate some potential problems of current approaches in the literature.