In this paper we discuss comparative results about three experiments that have been conducted using a Multi-Criteria Group Decision Support Systems. The three experiments were conducted in France, in Canada and in Brazil. The two first experiments are already published in references [8] and [9]. We report in this paper, the experiment conducted in Brazil and we compare the results in the three countries. Master students in Brazil, used the developed system called GRUS, in order to see how the participants to a group decision making process feel comfortable using shared criteria but also private criteria. The used methodology was the following: three hypothesis are proposed which are: 1) there is a need to use private and common criteria in a collaborative decision making process; 2) the number of private and common criteria should be the same; 3) the use of GDSS remains difficult without a human facilitator. After each experiment the participants answered to a semi-directed questionnaire. The results are then analyzed to see if the three hypothesis are verified in each country. As main result of this study we can assume than two of the three hypothesis are verified but the third one is not or partially verified.