Co-occurrence Network from Semantic Differential Data

The aim of this paper is to propose a method to transform semantic differential data in a network structure, whose graph representation is interpreted as an empirical adjectival graph. The derived graph is constituted by the adjectives of the semantic differential task as nodes, two nodes are linked depending on the scoring assigned by the set of respondents. Semantic differential data are handled by means of a peculiar coding that induces a weighted adjacency matrix. This relational data structure allows us to realize, through a graph representation of the adjectives network, an adjectival graph of the concept under study. The weighting system is given by the co-occurrence of respondents’ scoring. In this way, the cohesive part of the network (the core) is constituted by the set of the adjectives chosen by the most part of respondents, whereas those signifiers (adjective-nodes) less tied to the underlying concept, will locate in the peripheral part of the network. The proposed approach to adjectival graph aims at defining the concept-stimulus by looking at the edges between the various adjectives that characterize it. A case study is presented to show the significance of the proposed approach, while simulations will aid to validate the results.