A comparison between singular value decomposition and eigenvector method in group decision making

Abstract

In any group decision making exercise, the issues of having justified and consistent decisions are always of major concerns, especially when the group members come from diverse backgrounds, selection expectations and priorities, and there are a number of selection criteria to consider. This paper looks at these issues in the faculty member selection domain where the objective is to select the best candidate to be the new faculty member. Two hypotheses are tested using two-way ANOVA; overall candidates’ ranking depends on their own performance and competencies, and overall candidates’ ranking does not depend on individual selection committee members’ different roles. Selection committee members were treated as treatment factor while the candidates as the blocking factor. Singular value decomposition (SVD) and eigenvector method (EM), two different methods in analytic hierarchy process (AHP) were used to obtain the rankings. The statistical tests showed that both SVD and EM resulted in the same conclusion, that is the overall candidates’ ranking does not depend on the difference in the selection committee members’ selection criteria preferences, but on the individual candidates’ performance and competencies. Hence, all parties can be assured of an unbiased evaluations and the selection process itself is conducted in the most professional manner