This research studies the determinants of electoral choice in emerging democracies, where the number of parties is large, there is a low degree of partisanship, and volatile electorates. In such environments, how do citizens learn about issues and candidates? How do social background and economic circumstance combine with partisan ties, ideological predispositions, and media-delivered information to shape mass political behavior? What are the effects of conversations with neighbors and political activists? Is there accountability in electoral processes with these microfoundations? This project tests a comprehensive model of political attitude formation and change in Brazil, a country ideal for the study of mass behavior under multiple parties and weak partisanship. With the nation about to embark on its fourth presidential campaign since redemocratizing, the number of parties is high but stable. Although a rough left-right continuum differentiates the major parties and many voters, only a small minority of citizens expresses any partisan identification. The media are politically and technologically sophisticated, and civil society is increasingly mobilized. At the same time, more Brazilians point to informal conversations with friends, family and colleagues as their main source of political information than to any other source. The result is that Brazil, like many countries with weak, multiple parties, has very volatile and unpredictable election campaigns. Finally, Brazil is also ideal for testing the institutional sources of mass political behavior because it varies internally in terms of partisan alignments, individual party strength and electoral rules. To capture these various influences on mass political attitudes, this project proposes a context- sensitive panel study during the elections of 2002 in two Brazilian cities: Juiz de Fora (Minas Gerais) and Caxias do Sui (Rio Grande do SuI). Neighborhoods will be sampling units: 100 adults, enough to establish aggregate neighborhood opinion, will be interviewed in each of 20 neighborhoods in each city. The first wave will take place before the campaign begins, in March. The second wave will be just before the first round of the election, and the third wave will be after the second round, in the middle of October. About 800 "discussion partners" of the main respondents will also be interviewed. Concurrently, print media sources, TV news, and party-controlled TV programs will be content analyzed. Finally, members of the associations Iand political organizations that comprise each neighborhood's "political context" will be interviewed. This project will produce the first theoretically and empirically grounded model of vote choice in a milieu of numerous weak parties; that is, the first model including all the major factors potentially influencing electoral behavior. Because the Brazilian environment is matched by many emerging democracies in Eastern Europe and Latin America, our results will illuminate questions of democratic accountability, media and campaign influence, and the effects of political and social networks in many other settings as well.
