Zopounidis et al. used an ordinal regression analysis to assess an additive utility model and to obtain final ranking of a representative sample of commercial Greek banks. In this paper, we reanalyze their data by means of a new multivariate analysis method called Co-plot, a two-dimensional graphic display technique designed to analyze observations (e.g., 16 banks) and attributes (e.g., 7 attributes) simultaneously. The method produces three results: (1) similarity among the banks by the composite of all attributes involved (five clusters are obtained), (2) the structure of correlations among the attributes (three clusters are obtained), and (3) the mutual relationships between the banks and the attributes. The banks are mapped into a partial order according to their (increased) performance to obtain their rating. The final ranking obtained by the Co-plot method differed from that obtained by Zopounidis et al.
