Consideration and Choice: Analyzing Party Choice in the Swedish European Election 2014

Abstract

For a long time the question of to what extent party choice in the European Parliament (EP) elections is primarily dependent on voters’ orientations towards the European Union (EU) or just a mere reflection of orientations towards issues and actors in national politics has been debated. By combining insights from individual‐level models of party choice in second‐order elections with theories of sequential decision making this article investigates if, how and at what stages in the decision process attitudes to European integration matters for party choice. In line with previous work on first and second decision rule criteria in EP elections, this article develops and tests hypotheses about how voters’ orientations work at different stages of the voter decision process. The findings, based on Swedish data from a probability‐based three‐wave Internet campaign panel, indicate that many voters are in fact considering more than one party to vote for in the beginning of the election campaign. As expected, left‐right orientations function as a main decision rule with respect to which parties voters even consider voting for, while proximity on the European integration dimension mainly matters as a second decision rule in the final stage of the decision process. Using a sequential model with consideration and choice stages, the article reveals a much larger complimentary effect of EU proximity on party choice than has generally been found in previous research. This serves as a distinct contribution to the emerging research field of individual party choice in second‐order elections.

How (often) do voters change their consideration sets?

“…Studies that applied this model to voting behaviour typically ignored the awareness set, as well as the distinction between consideration set and choice set. Consequently, two phases in the vote choice process are distinguished: a consideration stage, in which voters select one or more viable alternatives from the available options (e.g., all competing political parties or candidates), and a choice stage, in which they select one party (or candidate) from within the consideration set (see e.g., Karlsen & Aardal, 2016;Oskarson et al, 2016;Schmitt-Beck & Partheymüller, 2016). Some voters may regard only one party or candidate as a serious option and then the consideration set is identical to the final choice.…”

“…Moreover, such countries often have a two-party system, for which consideration set models have less to offer. Sweden was selected also because there is presumably no other country for which more extensive and suitable data about electoral consideration sets are available, in large part due to the pioneering work by Oscarsson and his colleagues (Oscarsson et al, 1997;Oskarson et al, 2016). The availability of panel data about consideration sets is also another reason to include the Netherlands.…”

“…Because this kind of phased decision making is most likely when the number of choice options is large, consideration set models are particularly relevant for multi-candidate and multi-party elections. Indeed, several studies show that applying these models in such contexts leads to valuable insights (e.g., Wilson, 2008;Karlsen & Aardal, 2016;Oskarson et al, 2016;Schmitt-Beck & Partheymüller, 2016;Vassik et al, 2016;Moral & Zhirnov, 2018). Given the novelty of this approach in electoral research, obviously many pivotal questions about voters' consideration sets are still unanswered.…”

Individual narcissism, need for closure and the two sides of negative partisanship

“…For the analyses of negative party ID strength, the data was transformed into long format, meaning that the unit for analysis is no longer the respondent, but the respondent's negative identification strength with a single party (see also van der Eijk, van der Brug, Kroh, & Franklin, 2006). This stacked data format allows to analyse the general patterns between personality and negative party identification over all parties and is often used for analyses of multi-party systems (Nasr, 2020;Oskarson, Oscarsson, & Boije, 2016). As now voter-party dyads are the unit of interest, the stacked data format increases the sample size by the factor six, it is thus important to account for clustering on the level of the individuals (Beck & Katz, 1995).…”

Tactical Voting in Norway: Context, Motives and Occurrence

“…; De Vries & Rosema ; Oscarsson ; Oskarson et al. ). It follows from this definition of a tactical vote that tactical assessments are more likely to become decisive when the political advantage of the preferred party decreases.…”