Ideologia: uma análise a partir de três formas de mensuração

• Ano de defesa: 2018
• Autor(a) principal: Oliveira, Crislâini Priscilla Nunes de Campos
• Orientador(a): Mundim, Pedro Santos
• Banca de defesa: Mundim, Pedro Santos, Botelho, João Carlos Amoroso, Nunes, Jordão Horta
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Goiás
• Programa de Pós-Graduação: Programa de Pós-graduação em Ciência Politica (FCS)
• Departamento: Faculdade de Ciências Sociais - FCS (RG)
• País: Brasil
• Palavras-chave em Português: Ideologia; Validade da medição; Pesquisa de opinião pública; Comportamento eleitoral
• Palavras-chave em Inglês: Ideology; Validity of measurement; Public opinion research; Electoral behavior
• Área do conhecimento CNPq: CIENCIAS HUMANAS::CIENCIA POLITICA
• Link de acesso: http://repositorio.bc.ufg.br/tede/handle/tede/9296

The purpose of this dissertation is to analyze three ways of measuring ideology from five models to answer the following question: what is the best way of measuring ideology taking into account cognitive schemas, values and behavior? The objective is to analyze and present the problems of operationalization of the concept in relation to the validity and reliability of the indicators. For this, we tried to correlate the vote of the Goianienses to the presidency of the republic and state government on the scale of classic ideological self-positioning, where 1 means to be left and 10 right; to a new configuration of the classic indicator of ideology from the solution proposed by Wood and Oliver (2012) that considers the level of schooling of the interviewees; and to an ideology index derived from a battery of questions and forced answers with questions on issues related to economic, moral, and social values; control variables. The data used came from the Goianiense Electoral Study (ESEG), conducted in 2015 with 1200 Goianienses. In order to achieve the goal, the binary logistic regression method was used to estimate voting coefficients in relation to the ideological measurement models. The ideological measurement models, in brief, do not present explanatory power taking into account the low value of the estimated coefficients of determination. However, when using models selection criteria such as Akaike and Bayesian, even with non-significant coefficients, the classic self-positioning scale was the best model that was able to explain ideology and vote.