A systematic method of analyzing the colorimetric behavior of a set of observers is proposed. The differences between each observer and a standard one are analyzed with different procedures in two color-representation Systems. It is apparent that these differences strongly depend on the color-representation system in which they are analyzed. Results obtained in this work indicate that comparison between two observers should be carried out by applying an optimized operator. This operator minimizes the differences between the color-representation systems associated with the observers that are compared. The proposed method should be applied when color-matching properties of a set of observers, or when color matching obtained with different colorimetric instruments, are compared.