Targeted election campaigning: an Australian case study

Elections around the world have, in recent years, become highly professionalised and orchestrated events. Parties devote considerable resources to advertising, research and overall administration of election campaigns. One estimate puts the value of spending by each party in Australia on direct mail alone at $15 million. As part of this process there has been a high degree of emphasis on the marginal seats. The use of electoral databases has been prominent in the attempt to identify voters who are likely to be persuaded to vote for the campaigning party. The majority of the effort in campaigning is devoted to direct mail, telephone canvassing and, recently, internet based campaigning. However, the effect of a party representative making a personal visit to a constituent's household is much greater than the effect of direct mail or telephone canvassing. The paradox is that it only requires a handful of voters in a number of marginal seats to sway an election. The problem is to find and persuade these voters. A candidate who attempts to sway an electorate by visiting houses at random is unlikely to find the most swayable voters by chance. Clearly, the idea of identifying which groups of voters will vote in a particular way, and the corollary of this question, which way elections will go, is of significant interest both internationally and domestically. From the discussion above, the problem, however, seems intractable. The problem will be addressed by looking at the Australian situation in detail