Thematic and citation structure dynamics of Organic Food & Farming research

This paper analyses the Organic Food & Farming (OF&F) scientific domain dynamic throught a “progressive document co-citation analysis” based on peer-reviewed papers from Web of Science. The dataset of OF&F domain displayed an exponential growth and a thematic diversification pattern. Both dominant and marginal clusters in association with their main cited articles were identified. This study enables to pinpoint major themes addressed or emerging. It can feed further research work and projects, namely with the definition of information system and research policy. Introduction: need for research and research needs in OF&F The organic sector is increasingly considered by stakeholders as a tangible alternative to address current challenges faced by agriculture. This leads to a continuous expansion of both the knowledge base and the demand of synthesis in OF&F. Several authors suggested topics for research, whether as general themes (Watson et al., 2008; Rahmann et al., 2009) or specific fields (Kristiansen et al., 2006). Others addressed the specificity of organic research approaches (Lockeretz, 2000; Watson et al., 2008; Drinkwater, 2009). However, such reviews do not analyse evolutions in topics within the OF&F literature. This assessment is important as transnational projects and journals dedicated to OF&F develop. We propose, using scientometrics, to identify dynamic of specialties and associated research topics in OF&F domain. Material and methods The progressive document co-citation analysis (PDCA), developped inside the CiteSpace Software (Chen et al., 2010), allows to find the intellectual basis of specialties developed within the OF&F domain by mapping its internal citation structure. A co-citation link occurs when two documents are cited in the same citing documents. PDCA identifies high-density patterns (clusters) inside the co-citation network. Theses clusters gather the documents often cited together (intellectual basis) upon which authors build their specialties. The analysis is progressive since the dataset is divided into n-years slices, thus useful to interprete the clusters evolutions. As citation clusters are hierarchically organised, we iteratively tested various threshold parameters to produce different co-citation networks in order to detect a common robust structure emerging from the different resulting clustering. To better synthetise information, we gathered structurally and topically coherent clusters thanks to the CiteSpace automatic cluster labeling which detects specific cluster terms from citing documents (Chen et al., 2010). The Web of Science was used because it includes the cited references for each publication. The challenge was to build a query representing all the aspects of OF&F research (production modes, products, social dimensions) without inducing noises. Indeed, some terms related to OF&F are also used in other domains. After a first query, we extracted terms using “organic” to test their specificity in the OF&F domain. 1 UR 0767 Ecodéveloppement, INRA E-Mail : [email protected] We finally built a complex query, composed of 120 specific terms representing the domain and resulting in a dataset of 4401 journal articles from 1975 to 2009