El análisis composicional (CoDa) como herramienta para el mapeo de la teoría de la agenda-setting. Estudio de caso de la opinión sobre política en el sistema mediático español, dentro del período de las Elecciones Generales de 2015

Since researchers Maxwell McCombs and Donald Shaw published the seminal work of Chapel Hill on agenda-setting theory (AS) at the beginning of the 1970s, its academics have developed their investigations using two methodological tools: a first phase of content analysis, used to study the news coverage and, therefore, the prominence or salience of certain topics or attributes in the agenda; and a second study of statistical nature, based on the verification of the level of correlation between the issues, their senders or the global of the different agendas. Content analysis has generally focused on the relative or comparative importance of some topics over others. It is assumed that some media or social actors are busier than others in certain topics, which has led to a general study of the relative volume of information, to the detriment of the detailed content analysis. Such differences (or similarities) have been traditionally evaluated from the comparison of the frequencies of their main contents, which, by means of Spearman correlations between the senders, have fostered an analysis on the hierarchical order, disregarding the relative or absolute and, therefore, any other information. In view of this need, and to resolve these procedures when the interest lies in the relative importance of the parts of a whole, the thesis deals with the analysis of Compositional Data (CoDa), founded by Aitchinson in 1982, along with the graphic representation through biplots, promoted by Gabriel in 1971, for the study and mapping of the AS theory. The interest in CoDa lies in the proportion of each part or component, since the absolute quantities are irrelevant and only expose the sample size. A very similar problem in all aspects that is established in the study of the AS, where the absolute data are irrelevant and mostly point to the content, the intent of the content, the popularity of the issuer, etc. In this line, only the proportions of each content category or the relative size of one type of content over another are 3/3 truly informative. According to these premises, the present doctoral thesis is structured in five articles that distinguish the originality of the study in presenting the CoDa-biplot as a tool for mapping and studying the AS theory. Using the period of the 2015 General Elections in Spain, the thesis is presented as a thematic unit that addresses, throughout five publications, the mapping and compositional study of the theory of AS of opinion on politics in the Spanish media system (media agenda), as well as political communications on Twitter of the main Spanish parties (political agenda). Likewise, the aggregate study of both agendas (the media and politics) allows us to approach the analysis and representation of the intermediate AS ​
​L'accés als continguts d'aquesta tesi queda condicionat a l'acceptació de les condicions d'ús establertes per la següent llicència Creative Commons: http://creativecommons.org/licenses/by-nc-sa/4.0/