Sometimes voters choose to vote for another party than their most preferred one. One reason may be considerations on government outcome. Particularly in PR systems, where government outcome tend to be a coalition, voters may change strategies from one election to the other. Voting for another party than one’s most preferred one with the intention to affect government outcome is often referred to as strategic voting (see Blais et al. 2005). Strategic voting for small parties – sometimes referred to as threshold insurance voting (Cox, 1997) – is the focus of this paper. Previous studies have only briefly elaborated the relationship between contextual features and strategic voting for small parties (McCuen and Morton, 2010; Meffert and Gschwend, 2011). In this study two hypotheses are elaborated. The first expectation is that small parties get more strategic votes the closer they are to the threshold, since there should be a lower level when a vote for the small party could be seen as wasted. The second expectation is that voters’ tendency to vote strategically for small parties increase if parties are outspoken about their coalition intentions, i.e. the coalitions they will form if winning the election. This is tested using a survey experimental design within the Citizen-panel at University of Gothenburg, in Sweden. In Sweden there is a 4 percent electoral threshold to the parliament, and some previous studies have shown that some voters vote strategically for parties at risk of falling below this threshold (Oscarsson and Holmberg, 2011; Fredén, 2012). In the experiment, a large number of participants is randomly assigned to fictive opinion polls with different information on support levels and different coalition signals. The results show that the level of threshold insurance voting is highest when the small party is very close to the threshold and coalition signals clear.
