Determination of the relative importance of criteria when the number of people judging is a small sample
Technological and Economic Development of Economy
In the multicriterial decision aid (MCDA) process, specific information is given illustrating the preferences of a decision-maker or a group of stakeholders taking part in this process. One kind of information about the preferences is the relative importance of criteria. In the article it was put forward a proposition of settling the division of variability of value of a relative importance of criteria for the case when a group of people (one stakeholder's representatives) is involved in
... involved in constituting the so-called a small sample (n < 30). In the first place the homogeneity of the given information applying statistical-descriptive and positional measures was analysed. Then it was proposed to accept the confidence intervals (with the supposed value of the first kind of error) as a basis for determining the variability scope of the relative importance of individual criteria. The differences in the variability ol the individual criteria were determined according to the average arithmetical value calculated in the analysed test among those giving information. The application of the interval estimation constitutes the basis for carrying out of the sensitivity analysis of the final result obtained by the accepted MCDA method. In the analysed decision problem the Electre III method was used. Keywords: multiple-criteria decision aiding, relative importance of criteria, a group of stakeholder's representatives, a small sample, the homogeneity of assessment, interval of the variability.