Made available in DSpace on 2015-04-14T14:46:16Z (GMT). No. of bitstreams: 1 458327.pdf: 1372960 bytes, checksum: 0d2573de878d931ff4c34cce58aaeaab (MD5) Previous issue date: 2014-02-28 / This thesis is about the relationship between the mechanisms of direct democracy (MDDs), i.e. plebiscites, referendums and citizen-initiated votes, and citizens' perceptions of the democratic regime. Hypotheses are tested on the influence of the MDDs on the presidential approval, on trust in democratic institutions (Congress and political parties) and on regime performance, measured through external efficacy, namely the belief that voting can change the future. Data are from Latinobar?metro and encompass surveys carried out annually in 16 countries, from 1996 to 2010. Multilevel models are used to analyze the data. If institutions as MDDs can increase the legitimacy of these more concrete dimensions of democracy, this is a quicker solution than economic, social and cultural transformations. This study relies on rational choice theory assumptions which state that citizens can make informed and rational judgements as to whether or not their regime meets their standards of how it should work, while still considering the various influences of diverse contextual and institutional factors. The explanatory variable the MDDs was tested in five different forms: short-term effects (occurrence and weighted occurrence), long-term effects, i.e., the cumulative number of MDDs, type of mechanism and issues quantity and relevance. Multilevel models, the alternative to time series analysis when there is not a sufficient number of observations over time, have not looked like good substitutes when information regarding the cases and the dependent variables behavior were considered. However, despite the methodological limitations imposed by data and the rarity of the events, it is possible to say that, holding constant citizens' perceptions of the country current economic situation, interpersonal trust, education, age, and in some cases, presidential approval and trust in institutions, in Latin America, the use of national mechanisms of direct democracy about issues of broad interest do not damage citizens' support to the more concrete regime dimensions (or representative democracy specifically, maybe the biggest concern on MDDs), on the contrary, MDDs improve, in different degrees, according type, nature of the issues and the historic of the country's MDD experience, the political support. Related to the explanations taken into account, although in the presidential approval model and in the models of trust in institutions the government performance is much more important, in the model of the vote, the cultural and institutional explanation gain more ground and all the three have basically the same relevance. Among them, it is possible to highlight the interpersonal trust, education, mandatory plebiscites and MDDs involving issues, many or few, of great general interest. / Esta tese versa sobre a rela??o entre os mecanismos de democracia direta (MDDs), isto ?, plebiscitos, referendos, iniciativas populares, e a percep??o dos cidad?os sobre o regime democr?tico. Testam-se as hip?teses sobre a influ?ncia desses MDDs na aprova??o presidencial, na confian?a que os cidad?os depositam em algumas institui??es da democracia (Congresso e partidos), e na sua avalia??o sobre a performance do regime, medida atrav?s da cren?a na efic?cia do voto como instrumento de mudan?a. Os dados s?o provenientes do Latinobar?metro e englobam pesquisas de opini?o realizadas anualmente em 16 pa?ses de 1996 a 2010. Modelos multin?veis s?o utilizados para a sua an?lise. Se institui??es como os MDDs podem aumentar a legitimidade dessas dimens?es mais concretas da democracia, tratase de uma solu??o mais r?pida do que transforma??es estruturais econ?micas, sociais e culturais. Esse estudo apoia-se nos pressupostos da teoria da escolha racional de que os cidad?os s?o capazes de fazer julgamentos racionais, tais como se o regime funciona do modo no qual eles acreditam que deveria funcionar, enquanto considera a influ?ncia de fatores contextuais e institucionais diversos. A vari?vel explicativa os MDDs foi testada de cinco formas diferentes: efeito de curto prazo (ocorr?ncia e ocorr?ncia ponderada), efeito de longo prazo, isto ?, n?mero acumulado de MDDs, tipos de mecanismos e natureza dos assuntos envolvidos na vota??o. Os modelos multin?veis, alternativa ? an?lise seriada quando n?o h? n?mero de observa??es suficientes ao longo do tempo, n?o se mostraram bons substitutos quando informa??es a respeito dos casos e at? relativas ao comportamento das vari?veis dependentes foram analisadas. No entanto, apesar das limita??es metodol?gicas impostas pelos dados e pela raridade dos eventos ? poss?vel dizer que, mantendo constantes quest?es como a percep??o da situa??o corrente da economia, a confian?a interpessoal, a escolaridade, a idade e, em alguns casos, a aprova??o presidencial e a confian?a nas institui??es, na Am?rica Latina, a utiliza??o de mecanismos de democracia direta nacionais sobre assuntos de ampla relev?ncia n?o danifica o apoio dos cidad?o ?s dimens?es mais concretas do regime (ou ? democracia representativa especificamente, talvez a maior preocupa??o quando se trata de MDDs), pelo contr?rio, MDDs melhoram, em diferentes graus, de acordo com seu tipo, a natureza de suas quest?es e com o hist?rico de experi?ncia do pa?s com MDDs, o apoio pol?tico. Em rela??o ?s explica??es levadas em conta, embora no modelo de aprova??o do chefe do Executivo e nos modelos das institui??es a performance do governo seja muito mais importante, no modelo do voto, a explica??o cultural e a explica??o institucional ganham espa?o e se tornam quase equivalentes, com destaque para a confian?a interpessoal, a escolaridade alta, os plebiscitos mandat?rios e os MDDs que envolvem assuntos de grande relev?ncia, sejam poucos ou muitos.