Testing the Applicability of Perry's Model of "Probable Voters" for Predicting the Election Results for the Zagreb City Assembly in 2001

Tests the applicability of Perry's model of "probable voters" in Croatian electoral circumstances, as well as its possible influence on the predictive validity of pre-electoral research of public opinion in Croatia. Results of comparative field research & telephone opinion polls carried out on occasion of the 2001 Zagreb City assembly elections are used in the analysis. To define the relative significance of a specific variable from Perry's model, as a possible predictor of electoral participation, the first level of analysis established a factor structure & the reliability of included variables for determining the probability of electoral turnout. On the second level, the authors tried to establish to what extent a certain variable or group of variables, selected as significant in the previous analysis, could influence the validity of research on voting intentions at elections. Based on the latter it is concluded that in Croatian circumstances the application of certain elements of Perry's model has a more or less satisfactory but still limited impact on the validity of pre-electoral research & that, for Croatian purposes, it should be revised by introducing some additional variables of electoral behavior, first & foremost those that are more situationally connected to Croatian political & electoral circumstances. 8 Tables, 16 References. Adapted from the source document.