Based on regional socio-economic indicators and consolidated results from 1994 and 1998 elections for governor in the state of Paraná, this article presents data that demonstrate high levels of voting "volatility" that would not be perceived through data analysis at the aggregate state level. I analyze the electoral results from the gubernatorial dispute in Paraná through geographic divisions, using municipalities as my unit of analysis. Votes that candidates received in the diverse municipalities of the state are organized according to size of the latter and according to their levels of socio-economic development, as measured through the IDH-M (1991 Municipal Index of Human Development). Municipalities are classified as micro, small, medium and large. In order to establish a parameter of socio-economic development, I use the IDH, classifying municipalities into the following groupings: very low IDH, low IDH, medium IDH and high IDH. My aim is the social and geographic identification of the electorate that supported the major candidates for governor of Paraná for the above-mentioned consecutive elections. I show that there was an important level of electoral volatility in both of the elections which Jaime Lerner won, an aspect that is not perceived when analysis is based on aggregate election results. The deeper issue at stake here is related to the discussion of voting behavior in mass democracies, in which the voter is identified as a subject of the electoral process who has constantly changing demands and desires.
voting behavior; mass democracy; voting volatility; Paraná State Government; Jaime Lerner
