Cognitie, Creier, Comportament / Cognition, Brain, Behavior Vol. IX(3), 567-574, 2005 ©Romanian Association of Cognitive Sciences STUDYING COMMUNICATION NETWORKS WITH AGNA 2.1 Marius I. BENTA * University College Cork, Ireland ABSTRACT This paper gives a general description of the functionality and use of Version 2.1 of an application employed in the study of communication networks called Applied Graph and Network Analysis. Along with some elementary notions related to this field, the graphical user interface and the main tools available in Agna are presented. Various computational commands, such as network transformations or analyses, are described in the second part of the article. KEY-WORDS: communication networks, social networks, behavior analysis software Social Networks Analysis (henceforth SNA) is a powerful methodology that has known tremendous developments in the past decades (for a general-scope introduction to SNA, see Hanneman, 2005; Watts, 2003; Barabasi, 2002; Scott, 2000; Wasserman, 1994; Hoffman, 1992; Knoke, 1982). It has emerged from the studies in the social psychology of small groups done by Alex Bavelas, Harold Leavitt, Harrison White, and Claude Flament in the 50s and 60s coupled with the techniques employed in the sociometry school. More recently, SNA techniques have begun to be applied to studying non-human social behavior, particularly social structures and communication networks in mammals (see, for example, Newman, 2004; Lusseau, 2003; McComb, 2001; Wells, 1987). The main assumption of the SNA framework is that the systemic and sub- systemic properties of a communication group—be it a human, an animal, an institutional, or an economic relational structure—depend to a large extent on the topological and quantitative descriptors thereof. In other words, the shape of a group communication structure determines such attributes of the group as the efficiency in performing a given task, the degree of moral satisfaction of its members, or the chances of a given member to reaching a leadership position. SNA has developed its specific terminology largely borrowed from graph theory. Accordingly, a network is a set of nodes (or actors) connected among them * Corresponding address: E-mail: imbenta@yahoo.co.uk