The dog that didn't bark: Would increased electoral turnout make a difference?

The question of whether increased electoral turnout would make a difference has long been a matter of interest to psephologists and laymen alike. Several approaches to this question have been employed, one of which has been an analysis of responses to survey questions about how non-voters would have voted had they actually voted. The problem with this approach is not only that it involves responses to a hypothetical question, but also that the responses are given after the election outcome is known and hence are subject to the impact of various post-election effects – “bandwagon effects” in particular. In the present article an alternative approach based on an analysis of party sympathy scores is pursued. Party sympathy scores are calculated among voters and non-voters alike. On the basis of this information, an algorithm is then used to estimate what percentage of potential voters were “lost” by the failure of different parties to mobilize non-voters with sympathy scores comparable to their voters, and thereby assess the impact of non-voting on the election outcome. Analyses are carried out using data from six national parliamentary elections held in Norway in the period from 1981 to 2001. Results indicate that the impact of higher levels of turnout is marginal at best.