Predicting the Vote in Kinship-Based Municipal Elections: The Case of Arab Cities in Israel

Abstract

This paper examines the feasibility of using pre-election polls in kinship-based municipal elections in Israel, and the methods needed to increase the accuracy of predicting the results. The research focuses on the Arab society in Israel, which is an ethnic minority within a nation-state and a traditional society and that its municipal elections are characterized as kinship-based elections. The results of municipal elections in four Arab cities and towns show that pre-election polls succeeded in predicting the elections’ outcome with high accuracy. This research suggests the use of “Clan Sampling” in which the sampling is according to the sub-tribal societal structure—and further suggests applying the “Cross-Section” treatment for the undecided. Separately applying one of the two treatments gave a high level of accuracy; the accuracy further increased when applying the two methods together. By adjusting the methods used, pre-election polls were found to be accurate in predicting the vote in elections within a traditional Arab society.