Interpersonal Relationships in Group Interaction in CSCW Environments Yang Cao Computer Science Department,University of Saskatchewan Saskatoon, SK, S7N 5A9 Canada yac614@cs.usask.ca Golha Sharifi Computer Science Department,University of Saskatchewan Saskatoon, SK, S7N 5A9 Canada gos787@cs.usask.ca Julita Vassileva Computer Science Department,University of Saskatchewan Saskatoon, SK, S7N 5A9 Canada jiv@cs.usask.ca Yamini Upadrashta Computer Science Department,University of Saskatchewan Saskatoon, SK, S7N 5A9 Canada ysu156 @cs.usask.ca Abstract When designing computer-supported work systems (CSCW) it is important consider the evolution of attitudes among the users and how the achievement of user individual and collective goals influences third person. This is especially important in situations where users do not know each other in advance and all of their interactions are mediated through the environment, since the environment is instrumental in forming the relationships. Therefore our focus is not on modeling existing user attitudes and relationships, but on finding out how people actually develop attitudes of liking or disliking other people when interacting in a CSCW environment in a collaborative-competitive situation, how they change their attitudes towards others when they realize their attitudes towards themselves and how the design of the environment influences the emergent social fabric of the group. This paper describes a game designed specifically as a tool for this study. It discusses the preliminary experimental results and some of the consequences for the design of feedback mechanisms in CSCW systems. 1. Introduction The successful deployment of computer supported collaborative work systems (CSCW) requires taking into account social factors, like preexisting and developing attitudes, relationships between users, incentive mechanisms, organizational flows of control and information. There are examples of solid user communities that formed around pieces of technology (e.g. slashdot.com), but there are also hundreds of examples of failed ones. Therefore it is important to study the sociological aspects of cooperation, to discover and describe existing relationships among people, existing organizational structures (Artikis et al., 2002; Sierra & Noriega, 2002) and incentives for cooperative action (Golle et al., 2001) in the domain of the application and to incorporate appropriate mechanisms in the system’s architecture to accommodate them. However, modelling and adapting to existing user (organizational) roles is not sufficient. Empirical studies (Hummel & Schoeder, 1995) have found that “… important point lies in the selforganized dynamics of the lateral cooperation process itself, where countercooperative as well as cooperative situations may emerge”. The authors argued that “the success of CSCW applications is strongly influenced by the self-organizing dynamics of lateral cooperation”. Our goal is to find out how people develop attitude of liking or disliking other people when placed in a cooperative and competitive situation where their only interactions are through a computer-based environment. We use a specially designed mutli-player game environment with rules that require the users to express and modify explicitly their attitude to the other players. As in a real CSCW environment, the attitude of each player has indirect impact on his/her success, since only through cooperation can all players maximize their scores. However, uncertainty in the other