Publication Type:Conference Paper
Source:5th ECPR General Conference, Potsdam, Germany (2009)
Keywords:electoral analysis, voting behaviour
Consideration Set Models (CSM) are becoming increasingly popular in electoral research. Since the predictive power of most standard models of voting behaviour is slowly diminishing, CSM are needed to meet the challenges of a highly individualized voting behaviour. In an era of dealignment and increasing proportions of party switchers and late deciders, the process of choosing a party is no longer a routine manifestation of group loyalty, but rather a real individual choice between considered alternatives.
This paper addresses key measurement issues in developing consideration set models: how to effectively identify citizen’s consideration sets in multiparty systems. Some instruments tend to generate too large sets (include parties that are actually never considered), while others generate too small sets (do not include a party that was actually voted for). In this paper I evaluate a large number of different survey questions and techniques applied to isolate how many and what choice options are actually considered by voters close to Election Day. One tentative result is that the most recent instruments designed to meet the requirements of CSM tends to generate smaller sets than expected.