Localización y visualización de las principales líneas de investigación a través del análisis de co-palabras y del análisis de redes sociales. Propuesta metodológica para la delimitación temática de dominios científicos

  1. CANTOS MATEOS, GISELA
Supervised by:
  1. Félix de Moya-Anegón Director
  2. María de los Ángeles Zulueta García Co-director
  3. Benjamín Vargas Quesada Co-director

Defence university: Universidad de Alcalá

Fecha de defensa: 15 October 2017

Committee:
  1. Purificación Moscoso Castro Chair
  2. Elena Corera Alvarez Secretary
  3. Antonio Perianes Rodríguez Committee member
Department:
  1. Filología, Comunicación y Documentación

Type: Thesis

Teseo: 528208 DIALNET lock_openTESEO editor

Abstract

The main objective of this doctoral thesis focuses on the design of a methodological proposal that allows the localization, identification and visualization of the main research lines of a given scientific domain, regardless of its amplitude or specificity thematic. This proposal aims to enable a thematic delimitation of both a general and a more specific character, which may be applicable for both gross and fine granularity studies. For this reason, this proposal applies to two domains clearly differentiated from the point of view of their temporal, geographical and chronological coverage. "Domain 1" is based on research from the Medline database on health and women during the period 1965-2005. On the other hand, the "Domain 2" focuses on Spanish stem cell research collected in the Science Citation Expanded database during the years 1997- 2012. Thus, the major challenges facing this ethodological proposal are that of descending to very specific and specialized levels of analysis, such as the identification of scientific subdisciplines, lines of research, sub-lines of research, etc., and the identification of the thematic aspects, both of a more consolidated nature and those of a more dynamic nature. The methodological proposal presented in this doctoral thesis focuses on the combination of several bibliometric techniques. First, it is based on the use of Co-words Analysis for the identification of the most representative terms from their joint appearance in the documents. Specifically, the indexing terms used by the bibliographic databases that have been selected in this study as sources of information have been used as units of analysis. That is, the descriptors MeSH in the case of Medline and the KeyWords Plus and Author Keywords in the case of the Science Citation Index Expanded. Secondly, bibliometric techniques have been combined from the analysis of social networks and from the representation and visualization of information, mainly using two specialized computer programs such as Pajek and VOSviewer. In the case of the first, it focuses on the generation of maps based on graphs that in this thesis have been simplified through the Pathfinder Networks pruning algorithm (PFNETs) and represented by the visualization algorithm Kamada Kawai. In the case of VOSviewer this one focuses on the generation of maps based on the distance through the visualization of similarities (VOS) and that uses different techniques of clustering in the graphical representation. Given the advantages and limitations involved in the use of one technique over the other, both have been used with the intention that they can be used in a complementary way. The results obtained in this doctoral thesis have been presented in the different publications that integrate it and are a consequence of the different phases of study that were proposed. Specifically, the results show that it has been possible to locate and visualise the main thematic aspects of the two scientific domains analysed. In the case of “Domain 1”, four maps were obtained according to the selected time periods (1965-1974, 1975-1984, 1985-1994 and 1995-2005). In them, it could be observed that the orientation of research in the early stages, essentially, focuses on those aspects of the health of women that must do with reproduction. However, as the years go by, research is moving towards other more heterogeneous and diverse areas such as the social, labour and psychological aspects that also affect the health of women. In the case of “Domain 2”, the results allowed a clear distinction between clinical and basic research. In addition, they allowed the localization and identification of up to four research lines such as: the therapeutic use of stem cells in haematological diseases; hematopoietic stem cells; the embryonic stem cells, specifically, the processes of expression and cell differentiation, as well as the processes and biotechnologies necessary to locate, produce, grow and, above all, to analyse in vitro embryonic stem cells; and finally, neural stem cells. The results obtained in this thesis allow to conclude that the proposed methodological proposal is adequate for locating, identifying and visualizing the main trends or lines of research that characterize scientific production, regardless of the breadth or specificity of the domain used and the coverage of the information sources used. Its methodological suitability for thematic delimitation makes it a tool with great potential for the development of new bibliometric studies in possible future research. However, despite this, this methodological proposal can also be improved. In the future, it may be interesting to use other analytical techniques such as the Author Co-citation Analysis (ACA) or the use of other terms such as titles, abstract, keywords, etc., as units of analysis.