Election opinion polls are controversial: data gathering methods are not universally agreed upon, results are said to be open to manipulation and they have marginalised issue based political coverage in the media. Crucially, however, election polls are said to interfere with democratic integrity by influencing citizens’ vote choice on election day. This in turn is said to justify a prohibition on the publication of election poll results for a period of time prior to an election. These contentions are investigated in the following series of five guest blog posts by University of Otago Politics student Michelle Nicol. Based on her Honours dissertation, Nichol these blog posts ask whether election polls influenced voters at the 2008 New Zealand general election by either causing a bandwagon effect, inducing underdog effect or encouraging strategic voting. As no comprehensive examination of this topic has previously been undertaken, a framework for evaluation of the New Zealand case is constructed using overseas literature. Events from the 2008 election campaign, election poll results, media coverage and NZES data are then applied to that framework. Results indicate that New Zealanders are in fact ‘poll-driven fruitcakes’, as once lamented by David Lange. Despite this, Nicol concludes that claims of election poll interference with democratic integrity are weak. She argues that prohibitions on the publication of poll results would in reality be more damaging to democratic integrity in New Zealand than the influences of polls themselves. This first blog post provides an overview of the issue. [Read more below]
Since their development in the 1930s, surveys of electoral opinion have become a mainstay of election campaigns. In the lead up to the 2008 New Zealand general election the public were exposed to no less than the results of 63 election polls, conducted for various media outlets by TNS, Colmar Brunton, Fairfax and Digipoll.[1] This figure is growing at every election. Likewise, the proportion of coverage devoted to such election polls is increasing dramatically. In the eight week period leading up to the 2008 election, television and newspaper media devoted 10.1% of all election campaign coverage to poll related stories.[2] Coalition and MMP discussion occupied a further 17.5% of all media coverage, much of which was presumably based on inferences gathered from poll results. In fact, coverage of election polls and coalitions during the 2008 election campaign exceeded even coverage of the economy, during the onset of the global financial crisis.
Simultaneous to this increase in election poll incidence and coverage has been a worldwide decline in issue based political reporting.[3] The increased focus on election poll results and coalition possibilities is said to have “squeezed” out more relevant types of political information.[4] Factual reporting on parties’ and candidates’ policy positions, track records and capabilities is in sharp decline.[5] This holds true in regard to the 2008 New Zealand election campaign: of the top five most covered topics in the media (accounting for 50% of all election campaign coverage), only one topic, the economy, involved discussion of policy, track records and capabilities.[6] While media coverage of election poll results and coalition possibilities is increasing, reporting of political issues is becoming increasingly superficial and trivial.[7]
This change in media focus has important implications, given that the media is the primary source of election and political information for the majority of New Zealanders. A large and increasing proportion of the population is being exposed to the results of election polls during campaigns. The 42.2% of New Zealanders who watch television news nightly, and the further 27.4% watching television news five or six times a week inevitably find themselves exposed to the results of a large majority of election polls conducted.[8] The 43.8% of New Zealanders who read a newspaper five or more times a week will find themselves in a similar position.[9] Election polls are becoming difficult to avoid.
Despite their growing prevalence, however, election polls remain controversial. Not only have polls marginalised coverage of policy positions in the media, but those conducting election polls are often criticised for using imperfect data gathering and statistical methodologies. More sinister is that such allegations occasionally relate to intentional attempts to manipulate poll results. Of central interest to this dissertation, however, is the debate over whether election polls interfere with democratic integrity, specifically within the New Zealand situation. While democracy expects citizens to vote in a way which, “reflects their true feelings and attitudes”,[10] exposure to election poll results is said to influence citizens’ vote choice on election day. This in turn is said to justify a prohibition on the publication of election poll results for a period of time prior to an election.
Exposure to the results of election polls is said to influence citizens’ vote choice in three possible ways: first, by causing a bandwagon effect, second, by inducing an underdog effect, and third, by encouraging strategic voting. The bandwagon effect is a simple case of popularity breeding popularity. The term bandwagon is synonymous with an activity or cause that has suddenly become fashionable or popular. Individuals change their views, or act in a particular way that makes them consistent with the majority. Specifically within the umbra of election polls, the bandwagon effect describes the situation in which, “the predicted winner in an election poll gains additional votes as a result of the publication of that poll”.[11] Voters are exposed to polls showing a particular candidate or party is fashionable and popular, and taking that information into account, are then said to “jump on the bandwagon” and vote for that candidate or party.
Conversely, the underdog effect is said to work in favour of a candidate or party thought to have little chance of winning the election. It covers the situation where the, “predicted loser in an election poll gains additional votes as a result of the publication of that poll”,[12] the standard explanation for which is a feeling of sympathy.[13] A candidate or party may increase their vote share simply because election polls showed them losing.
The third alleged influence of election polls is that they encourage strategic voting. Instead of being swept along by a bandwagon or underdog effect, some voters are said to employ election polls to, “make calculated, maximising decisions about how to vote”.[14] The strategic voter uses polls to form expectations about candidate or party success, and votes according to those expectations to bring about the most desirable of possible outcomes. Strategic voting involves individuals casting a vote for a party or candidate other than their most preferred, often to bring about the defeat of their least preferred candidate,[15] or with coalition possibilities in mind.[16]
Each of these election poll influences may result in votes that are not reflective of true feelings or attitudes. Such “false” votes go toward electing governments, and as such are said to put democratic integrity at risk. This in turn is said to justify a prohibition on the publication of election poll results for a period of time prior to an election. Crucially, however, there is no actual consensus upon whether election polls cause a bandwagon effect, induce an underdog effect, or encourage strategic voting. Some academics believe these election poll influences are reality, yet others dismiss the possibility of such influences altogether.
While this disagreement has generated huge volumes of literature, research has primarily been undertaken in the United States (US) and the United Kingdom (UK). No comprehensive examination of the New Zealand experience with election polls, their possible influences, and corresponding impacts on democratic integrity has been undertaken. Neither are the conclusions (albeit mixed) from the US and UK able to be applied directly to the New Zealand situation; the differences between electoral systems and corresponding electoral incentives are simply too great.
There are indications in recent political history, however, that suggest such election poll influences are worthy of examination in New Zealand. Take for example United Future leader Peter Dunne’s experience with the “worm” at a 2002 televised Leaders’ Debate. The worm was a real time election poll, controlled by a cross-section of floating voters as a response to feelings of approval or disapproval as to what a leader said.[17] Dunne received a highly positive worm reaction,[18] subsequently followed by a jump in election poll ratings for the Party.[19] The Party began to receive widespread media attention, and United Future gained eight seats in parliament, rather than the single seat expected prior to the debate.[20] Voters were exposed to the worm, which showed United Future as being fashionable and popular. Taking that information into account, citizens jumped on the bandwagon and voted for United Future.
Similarly, indications exist that strategic voting as a result of election polls occurs in New Zealand. The most well known example is Jeanette Fitzsimmons’ Coromandel electorate success in 1999. An election poll released in October 1999 showed the left-wing Green candidate to be polling just 3% below National’s incumbent MP.[21] The Labour candidate for Coromandel had polled at just 15%, well below the other candidates. This resulted in Labour leader Helen Clark effectively calling for Labour supporters in the Coromandel to give their electorate vote to Fitzsimmons.[22] The incumbent National MP was thus denied re-election.[23] Individuals within the Coromandel cast their electorate vote for a candidate other than their most preferred in order to bring about the defeat of a least preferred candidate.
These indications of election poll influence, the upward trend in poll incidence and coverage, and the resulting potential for interference with democratic integrity necessitate investigation into the New Zealand case. This dissertation therefore uses a case study of the 2008 New Zealand general election in an attempt to establish whether election polls cause a bandwagon effect, induce an underdog effect, or encourage strategic voting in the New Zealand situation. This in turn enables evaluation to be undertaken on whether election polls in fact interfere with democratic integrity.
The next blog post examines and reconciles overseas literature on each of the three possible election poll influences. The main theoretical justifications for each of the bandwagon effect, underdog effect and strategic voting are evaluated, allowing the chapter to conclude that each of the three election poll influences are in fact reality. A profile is drawn of the types of individuals most likely to be affected by each influence, and where possible numerical indications of the strength of each influence are given. In doing this, the next blog post achieves its aim of developing a framework for analysis of the New Zealand situation.
The third blog post then uses the framework set out in the previous chapter to draw conclusions as to the existence of the bandwagon effect, underdog effect and strategic voting at the 2008 New Zealand general election. Where possible, the theoretical justifications for each influence are tested against the New Zealand situation using a combination of election campaign events, media coverage, results from each of the 63 election polls conducted between the beginning of 2008 and the general election, party donation declarations and the results of the 2008 New Zealand Election Study (NZES).[24] The post concludes that all three election poll influences are indicated at the 2008 New Zealand General Election.
Taking these findings into account, the fourth blog post then examines the specific ways in which election poll influences are said to interfere with democratic integrity. That post then looks at the appropriateness of the often proposed solution to these issues: restrictions on the publication of election poll results for a period of time prior to an election.
[1] Note: This dissertation refers to “election polls” rather than the more generic “opinion polls” to make clear it is concerned only with polls of pre-election popularity, and not with the results of wider issue-based opinion polls.
[2] B Bahador, “New Zealand Media Coverage of the 2008 Election Study” (paper prepared for publication in, Key to Victory: the New Zealand general election of 2008, ed. S Levine and N Roberts (Wellington, NZ: Victoria University Press, 2010)), 17.
[3] S Ansolabehere and S Iyengar, “Of Horseshoes and Horse Races: Experimental Studies of the Impact of Poll Results of Electoral Behaviour,” Political Communication 11 (1994): 413-430, 427-8.
[4] Ibid.
[5] Ibid.
[6] Bahador, “New Zealand Media Coverage of the 2008 Election Study,” 17.
[7] S Cross and J Henderson, “Public Images and Private Lives: The Media and Politics in New Zealand,” Parliamentary Affairs 57, no. 2 (2004): 142-156, 143.
[8]NZES, “Results from the 2008 NZES: New Zealand Politics and You,” New Zealand Election Study, http://www.nzes.org/exec/show/freq_2008a (accessed 22 September, 2011).
[9] Ibid.
[10] J Matthes, “Mass media and public opinion: Manipulating or enlightening?,” (paper presented to the 9th Dialogue on Science, Engleberg, October 13-15, 2010), 1.
[11] R Henshel and W Johnston, “The Emergence of Bandwagon Effects: A Theory,” The Sociological Quarterly 28, no. 4 (1987): 493-511, 494.
[12] Ibid.
[13]Ibid., 498.
[14] P Nesbitt-Larking, Politics, Society and the Media (Peterborough, Ontario: Broadview Press, 2007), 299.
[15] S Bowler, J Karp, and T Donovan, “Strategic coalition voting: Evidence from New Zealand,” Electoral Studies 29 (2010): 350-357, 350.
[16] B Cain, “Strategic Voting in Britain,” American Journal of Political Science 22, no. 3 (1978): 639-655.
[17] Cross and Henderson, “Public Images and Private Lives,” 150.
[18] Television New Zealand, “Worm likes Peter Dunne,” Television New Zealand, July 16, 2002,
http://tvnz.co.nz/view/news_politics_story_skin/116854 (accessed 17 May, 2011).
[19] Cross and Henderson, “Public Images and Private Lives,” 149.
[20] Ibid., 150.
[21] J Vowles, “Did the Polls Influence the Vote? A Case Study of the 1999 New Zealand General Election,” Political Science 54 (2002): 67-77, 72.
[22] Scoop, “The Greens & the Coromandel: Labour Begin To Move,” Scoop Independent News, 26 October, 1999.
[23] Vowles, “Did the Polls Influence the Vote?,” 72.
[24] Note: The NZES is a post-election study of voters funded by the University of Auckland and the Electoral Commission.