The purpose of this thesis is to measure the impact of social structure on voting in a Latin American context through quantitative and qualitative analyses of Peruvian industry labor’s support to the left parties at national and sub-national level in the 1980 presidential election. The left party family include on one hand the communist parties, on the other hand the social democratic Alianza Popular Revolucionaria Americana. Lipset and Rokkan’s cleavage theory, Lafferty’s typology of local community contexts and political response, and Knutsen’s theory of crossing cleavages’ effect on party choice constitute the theoretical framework.
Data employed in the quantitative analyses are aggregated at the commune level and include electoral data from the 1980 presidential election, census data from the 1981 National Census, and a geographical database. The proportions of industry workers who voted for the communists and for APRA are estimated with Gary King’s model for ecological inference and the EzI software. The estimates from this model are analyzed with spatial statistical techniques and presented in maps using Luc Anselin’s software GeoDa.
The conclusion is that an economic cleavage was not consolidated at the time of the 1980 elections. However, social structure did have effect on labor’s vote to the communist parties. In the case of APRA, there was no effect. Clusters of strong labor support to the communist parties are mainly identified in mining regions in the Andean highlands, while clusters of social democrat support dominated in the northern coastal agro-industrial regions and in the northern lower Andean highlands.