An examination of the 2008 New Zealand general election provides indications of the existence of the influence of opinion polls on voter behaviour. This blog post shows how in that election, the National Party benefitted from the bandwagon effect, while Labour and New Zealand First received increases in support attributable to the underdog effect. Likewise, National and Labour electorate candidates, the New Zealand First, Act, and United Future Parties, and Jim Anderton benefited from strategic voting based on opinion poll results. The Green Party, however, was likely negatively impacted by third party squeeze. These results indicate that exposure to opinion poll results are likely to also influence citizens’ vote choice on election day in 2011. This blog post is the third in a series of five guest posts by University of Otago Politics student Michelle Nicol, who has recently carried out research on opinion poll influences. [Read more below]
Diversities in electoral systems and political landscapes provide citizens with altered electoral incentives, resulting in different voting behaviours and outcomes. The findings of the last blog post therefore cannot be directly applied to the New Zealand situation. Further analysis is necessary to determine whether exposure to election poll results influences citizens’ vote choice on election day in New Zealand. So this blog post uses the analysis from the previous blog post as a framework to draw conclusions as to the existence of the bandwagon effect, underdog effect and strategic voting in New Zealand. Where possible, the theoretical justifications outlined the previous post are tested against the New Zealand case using a combination of election campaign events, results from each of the 63 election polls conducted between the beginning of 2008 and the general election, media coverage of the election campaign, and the results of the 2008 NZES. The blog post concludes the premise that election polls cause a bandwagon effect, induce an underdog effect and encourage strategic voting is indicated in the context of the 2008 New Zealand general election.
The Bandwagon Effect
The bandwagon effect describes the situation in which a fashionable and popular political party or candidate receives extra votes purely as a result of that popularity. There can be no doubt that the “fashionable and popular” party of the 2008 New Zealand general election was the National Party.
The National and Labour Parties were relatively close in the polls from the time John Key became leader of the opposition in 2006 until May 2007.[1] From this point on, the National Party steadily increased its lead, beginning a domination of election poll results that has continued to this day. National’s average election poll rating for the year of 2008 up until the election was 50.05%. Their nearest rival, the Labour Party, managed an average of just 34.49% support during the same period. National went on to win the election by a margin of 11%.[2]
Establishing the popularity of the National Party, however, does not in itself prove the existence or otherwise of the bandwagon effect in the New Zealand context. NZES data is illustrative in this regard. The 2008 NZES asked subjects which party best represented their political views on election day.[3] National was the choice of 33.7% of subjects.[4] At a later point in the survey, subjects were asked which party they in fact voted for at the 2008 election: 39.4% of the same subjects responded National.[5] National received 5.7% more support than would be expected if subjects voted for the party that best represented their views. This disparity is increased if including subjects who would have voted National, but for some reason did not make it to the polling booth. In that case, the 33.7% of subjects who believed National best represented their political views can be contrasted with a 41.5% share of the vote.[6] In this scenario, National would have received 7.8% more votes than expected among NZES subjects. These figures are consistent with the numerical representations of the influence of the bandwagon effect outlined in the previous post.
The question then, is why National received more votes from NZES subjects than would have been expected based on representation of political views. Unfortunately, NZES data cannot provide a specific answer, as such questions were not asked. One possibility, however, is that because National was shown as the fashionable and popular party in election polls, individuals changed their positions to conform to this view. Citizens saw National as fashionable and jumped on the bandwagon. A momentum theory induced bandwagon effect is thus a possibility in the New Zealand context.
This possibility is further enhanced when examining the position of the National Party with reference to the secondary bandwagon effects discussed previously. Recall that the bandwagon effect has the potential to grow throughout a campaign, as high election poll ratings can lead to, “an outpouring of media coverage... and campaign contributions”.[7] This enables the recipient candidate or party to run a more effective campaign and improve their position in the next election poll, thereby enhancing the bandwagon effect.[8]
Bahador analysed newspaper and television campaign coverage during the eight weeks prior to the 2008 election. He found that 33.5% of all media coverage primarily devoted to parties was about the National Party. The Labour Party, however, received 38.2% of all party based media coverage – a full 4.7% more than National, despite going on to lose the election by a margin of 11%.[9] While some of this difference likely stems from advantages associated with incumbency, these numbers initially seem to disprove the existence of secondary bandwagon effects in New Zealand.
Party leaders, however, play a crucial role in election campaigns in these days of professionalised politics. It is therefore important to include coverage of party leaders in any calculations of media coverage for the purpose of establishing secondary bandwagon effects. It is here that the possibility of a secondary bandwagon effect is supported. Bahador found that John Key received a total of 40.5% of all media coverage devoted to party leaders. Helen Clark, in contrast, was the subject of just 25.7% of such coverage.[10] When combined, the National Party and John Key actually received a much greater proportion of media coverage than the Labour Party and Helen Clark: 37% of party and leader coverage compared to just 31.95% respectively. Previously mentioned incumbency advantages make these figures all the more impressive. The secondary bandwagon effect of an outpouring of media coverage is therefore indicated in favour of the National Party at the 2008 general election.
A secondary bandwagon effect is also suggested in the context of campaign contributions.[11] High election poll ratings are said to cause an increase in campaign donations.[12] In 2006, Labour had a slight edge over National in terms of election poll support. Levels of large donations required to be declared with the Electoral Commission reflected this: Labour received $240,988.04, while the National Party received a somewhat lower total of $169,666.00.[13] In 2007, however, the National Party overtook Labour in terms of election poll support, retaining the lead throughout 2008. For this period the National Party received $911,101.00 in large donations. Labour, in contrast, received only $771,695.95; a full $139,444.05 less than National. Large donations reflected levels of election poll success. A secondary bandwagon effect in favour of the National Party at the 2008 election is therefore indicated.
Unfortunately an examination reference group theory and cognitive dissonance theory is unable to be undertaken. There have been no studies in these areas in New Zealand, and accordingly the necessary information is unavailable. Gathering such information is well beyond the scope of this dissertation. Despite this, it is clear that the National Party was the fashionable and popular party of the 2008 election campaign. National received more votes from NZES subjects than would be expected had subjects voted for the party that best represented their views. Furthermore, high election poll ratings seem to have led to, “an outpouring of media coverage... and campaign contributions”.[14] A bandwagon effect in favour of the National Party at the 2008 New Zealand general election is indicated.
The Underdog Effect
The underdog effect works to the advantage of candidates and parties thought to have little chance of winning an election. As outlined previously, the standard explanation for this is sympathy.[15]The New Zealand context, however, may offer a related but alternative explanation: tall poppy syndrome. Tall poppy syndrome is the habit of denigrating those who are successful and it has long been associated with New Zealand culture.[16] These feelings of disapproval over the success of others then benefit those who are less successful: the underdogs.
To examine the existence or otherwise of the underdog effect at the 2008 New Zealand general election, a combination of campaign events, election poll results, and results of the NZES are used. There were two clear underdogs of the election: the Labour Party and New Zealand First.
The Labour Party
The Labour Party was the undisputed major party underdog of the 2008 election. After three terms in power, election polls never showed Labour exceeding the 40% support mark.[17] In fact, election polls from the beginning of 2008 up until the election showed that on average only 34.5% of citizens intended to vote Labour.[18] In contrast, 50% of citizens intended to vote National.[19] Labour was widely and correctly thought to have little chance of winning the 2008 election.
Establishing the existence of an underdog, however, says nothing about the existence or otherwise of the underdog effect in the New Zealand context. NZES data is again illustrative in this regard. The 2008 NZES asked subjects which party best represented their political views on election day. Labour was the party chosen by 29.5% of subjects.[20] At a later point in the survey subjects were asked which party they in fact voted for at the 2008 election; 31.9% of the same subjects responded Labour.[21] Labour received 2.4% more votes than would be expected if subjects voted for the party that best represented their views. This disparity is increased if including subjects who would have voted for Labour, but for some reason did not make it to the polling booth. In that case, the 29.5% of subjects who believed Labour best represented their political views can be contrasted with a 36.1% share of the vote.[22] In this scenario, Labour would have received 6.6% more support than expected. These figures, but particularly the first, are consistent with the numerical representations of the influence of the underdog effect outlined in the last post.
The question is then why Labour received more votes from NZES subjects than would have been expected based on representation of political views. Again, NZES data does not include a measure of why this is the case. One possibility, however, is that feelings of sympathy led to Labour, the predicted loser in election polls, gaining votes as a result of the publication of those polls. That is, NZES subjects succumbed to the underdog effect. The underdog effect is therefore a possibility in the New Zealand context.
New Zealand First
This possibility is further supported by an examination of New Zealand First’s 2008 election campaign. Like the Labour Party, New Zealand First was an underdog of the 2008 election, albeit of the minor party ranks. New Zealand First began the year amid speculation that the Party was the recipient of a $100,000 donation from Owen Glenn. Winston Peters, the Party’s leader, denied this emphatically with his infamous “NO” sign. Later in 2008 Peters was shown to have lied, and questions arose over further undeclared donations. [23] These events culminated in New Zealand First and Peters being investigated by the Serious Fraud Office, the Privileges Committee, and the Police. Not only was Winston Peters seen as unlikely to win the Tauranga electorate seat, the Party was thought to be in real danger of falling below the 5% threshold mandated by MMP.
In terms of whether this underdog status translated into an underdog effect, NZES data is again illustrative. Like Labour, fewer subjects saw New Zealand First as representative of their political views than voted for the Party. In raw terms, 3.5% of subjects saw New Zealand first as representative of their views,[24] while 3.7% of subjects voted for the Party.[25] This figure increases to 4% when controlling for those subjects who would have voted for New Zealand First, but for some reason did not make it to the polling booth. In this scenario, New Zealand First would have received 0.5% more votes than would be expected if individuals voted for the party that best represented their views. This is a significant percentage for a minor party, and again consistent with the numerical influence of the underdog effect outlined previously.
As with the Labour Party example, NZES data does not include a measure of why subjects voted for a party other than that which best represented their political views. What is different about the New Zealand First example, however, is the clearly observable nature of their election year difficulties. After an examination of newspaper coverage of the New Zealand First 2008 election campaign, 16 key events were placed on a timeline. This timeline was then contrasted with a timeline of election poll results, giving a clear picture of how voters reacted to each key campaign event. Results are presented in Graph 1. The key events were:
- First election poll of 2008 (20 Jan);[26]
- Beginning of speculation over Glenn donation (20 Feb);[27]
- Presentation of the infamous “NO” sign (28 Feb);[28]
- New Zealand Herald’s publication of Glenn email confirming donation (12 Jul); [29]
- Winston Peters admits Glenn donation (18 Jul);[30]
- Questions raised over undeclared Vela family and Jones donations (22 Jul);[31]
- Privileges Committee asked to investigate Glenn donation (5 Aug);[32]
- Glenn gives evidence to Privileges Committee to effect that Winston Peters lied about sourcing donation (27 Aug);[33]
- Serious Fraud Office launches investigation into Vela family and Jones donations (28 Aug);[34]
10. Rodney Hide lays complaint over undeclared donations with Police (Sep);[35]
11. Electoral Commission begin investigating undeclared donations (8 Sep);[36]
12. Winston Peters censured (23 Sep);[37]
13. New Zealand First launches official campaign with attacks on National, Act and the media (5 Oct);[38]
14. Winston Peters and New Zealand First are cleared by the Serious Fraud Office (11 Oct);[39]
15. Winston Peters and New Zealand First are cleared by the Electoral Commission (23 Oct);[40] and
16. New Zealand First cleared by Police (4 Nov).[41]
As shown by Graph 1, New Zealand First began the year polling at 3.5%, and remained around the 3% mark until shortly after Peters’ presentation of the “NO” sign. Support then dropped to a campaign low of just 1.1%. During the Party’s relatively quiet period, from April to June, support hovered around 3.5%. When the Peters story unravelled revealing donation lies, however, and ensuing investigations begun, election poll support for New Zealand First spiked to campaign highs of 6.5% (twice), and 5%. Then, between being cleared by the SFO, Electoral Commission and Police, support for New Zealand First fell back to around 3%.[42]
Not only does NZES data reveal that New Zealand First received more votes than would be expected if subjects voted for the party that best represented their views, analysis shows that election poll support for New Zealand First was at its lowest when Winston Peters was a non-story and when he was cleared of any wrongdoing, and at its highest when Winston Peters was being vilified in the media. It is unlikely this peak in election poll support is due to a public perception that lying, fraud or other illegal election behaviour is desirable. More likely is that the qualitative information received by individuals through the media operated to induce an underdog effect, much like Fleitas found in the second blog post on this topic. Individuals sympathised with Winston Peters and New Zealand First. This is a key element of the underdog effect.
NZES data and campaign events contrasted with election poll results indicate that the Labour Party and New Zealand First benefitted from an underdog effect at the 2008 New Zealand general election.
Strategic Voting
Strategic voting requires individuals to form expectations about candidate and party success, and then vote according to those expectations to bring about the most desirable of possible outcomes. To examine the existence or otherwise of strategic voting at the 2008 New Zealand general election, a combination of campaign events, election poll results, and results of the NZES are applied to each of the three theories of strategic voting outlined in Chapter One: coalition insurance strategy, third party squeeze, and the wasted vote phenomenon.
As in Chapter One, however, it is necessary to first establish that polls are used by individuals to form expectations about election outcomes. While there is no New Zealand study directly on point, Bowler touched on the issue in the context of the 2002 NZES.[43] The 2002 study included a pre-election survey of 3590 subjects, a set number of which were interviewed each day for 36 days prior to the election and asked to assess each party’s chance of being in government post-election.[44] Responses were consistent with shifts in election poll results; that is, as support for a party increased in election polls, so too did expectations the party would be in government.[45] Subjects formed expectations of election outcomes through election polls. The scene for strategic voting in New Zealand is thus set.
Coalition Insurance Strategy
Given the similarities between New Zealand and Germany’s political systems, it is would be sensible to assume that Gschwend’s coalition insurance strategy is applicable. Recall the theory suggests:
- Minor party supporters will recognise that their minor party candidate has little chance of winning the electorate vote. They then give this vote to the candidate from their most preferred (major party) coalition partner.
- Major party supporters may find their desired (minor party) coalition partner is in danger of falling below the 5% threshold. They then give the minor party their party vote.
Two further elements with particular relevance to the New Zealand situation may be added to this theory. First, coalition insurance strategy could apply to major party supporters who wish to ensure their party maintains or strengthens a particular ideological platform. Such voters give their party vote to a party with a slightly stronger ideological foundation with a view to a coalition being formed. Second, coalition insurance strategy could also be used to explain the situation in which major party supporters in a particular electorate (where their desired coalition partner has a realistic chance of winning a candidate seat) find their desired coalition partner is in danger of falling below an electoral threshold for representation. They then give the minor party representative their candidate vote. The first of these is consistent with Gschwend’s pattern of minor parties benefitting from party votes at the expense of major parties, while the second provides a minor exception to the general rule that minor party candidates are deserted at the expense of major party candidates. Therefore:
- Major party supporters wish to ensure their party maintains a particular ideology. They then give their party vote to a (usually minor) party with a slightly stronger ideological foundation.
- Major party supporters in a particular electorate (where their desired coalition partner has a realistic chance of winning the electorate seat) find their desired (minor party) coalition partner is in danger of falling below the 5% threshold. They then give the minor party their candidate vote.
If the theory and additions are applicable in New Zealand, large party electorate candidates will receive more votes than the level of candidate preference would suggest; small parties will receive more party votes than level of party preference would suggest; and small party candidates in particular electorates will receive more votes than levels of candidate preference would suggest. 2008 NZES data tends to support each of these trends.[46] It must be noted at this point that figures in this section cannot take into account hypothetical votes of non-voters due to the NZES not recording this data for electorate votes.
The first element of Gschwend’s coalition insurance strategy is indicated in the New Zealand context. Candidates from the Labour and National parties received more votes than should be expected based upon the measure of most liked candidate according to NZES data. Labour candidates received 32.7% of electorate votes while only 26.2% of subjects liked Labour candidates the most. Likewise, National candidates received 39.7% of electorate votes, while only 34.5% of subjects liked National candidates the most.[47] Large party candidates received 11.7% more votes than levels of preference suggested.
Both Gschwend’s second element and the added third element involve minor parties receiving more party votes than levels of party preference would suggest. The picture presented by NZES data largely confirms these expectations; New Zealand First, Act and United Future all received more votes than expected; collectively 1.9% more. The Maori and Green parties, however, received fewer votes than the number of subjects who claimed to like their party the most.[48]
The case of the Maori Party is easily and simply explained; the Party is electorate based.[49] Despite a push for party votes at the 2008 campaign launch, the Party has continually emphasised the importance of winning all seven Maori electorate seats. The case of the Green Party is also easily explained (see below), and may actually lend credence to the theory of strategic voting: the Green anomaly could well be a case of third party squeeze.
The fourth element added to Gschwend’s theory also proves accurate in the New Zealand context. Along with Labour and National, United Future and Progressive Party candidates received more votes from NZES subjects than expected. United Future candidates received 1% of votes, while only 0.4% of subjects liked the United Future candidate the most. Likewise, Progressive candidates received 1.1% of votes, while only 0.8% of subjects liked the Progressive candidate the most.[50] This 0.9% departure from Gschwend’s original element one trend can be explained by both Parties’ absolute reliance on winning an electorate seat to get into parliament.
The Progressive Party has been in coalition with Labour since first entering parliament in 2002. The Party began with two MPs, but has since become reliant on Party leader Jim Anderton winning the Wigram electorate. This gives Labour Party supporters in Wigram an incentive to vote strategically by giving their candidate vote to the Progressives. The strategic incentive rises greatly when considering the next highest polling candidate in Wigram at the 2008 election was a National Party member. While Jim Anderton won the electorate with 44.5% of votes, if Labour supporters had voted for the Labour candidate (trailing far behind on just 15.15%) it is likely the National candidate who polled at 30.66% would have won the electorate.[51] Similar incentives exist for voters in Ohariu; an electorate United Future is dependent on Peter Dunne winning.
Gschwend’s coalition insurance strategy seems to apply in New Zealand. Movement in votes is consistent with predictions. Furthermore, these movements account for a total of 14.5% of votes – very close to Gschwend and Meffert’s estimate of 15% of voters acting strategically. While that this analysis does not prove definitively that these shifts in votes occurred as a result of the coalition insurance strategy, the consistency with which moves are made indicates that the coalition insurance strategy very likely exists in New Zealand.
Third Party Squeeze
Third party squeeze, “refers to the flow of third party support to one of the other two parties when the race is close between those two but not the third”.[52] This could explain why (as above) 5.9% of NZES subjects responded that they “liked” the Green Party the most, yet only 5.5% of the same subjects actually voted Green.[53] Third party squeeze may also explain why the Greens traditionally receive such poor results on election day.
The Green Party polled moderately well throughout 2008. From the beginning of the year until the launch of their “Vote for me” advertising campaign, average poll support was at 6.46%, after which time the Green’s average poll rating increased to 7.23%. The Party saw a further increase in support with the launch of their official campaign on 5 October. The average election poll rating from that date until the election was 7.86%. Despite this relatively solid level of support and a lack of campaign disasters, the Greens received just 6.72% of votes on election day.[54]
The key question is why many of those expressing support for the Greens fail to deliver on election day. Two interrelated aspects of the 2008 election may explain this. First, Green supporters were likely aware of the Green’s much publicised ruling out of dealing with the National Party post-election.[55] This placed the Green Party firmly within Labour’s camp, effectively making Green Party influence in parliament dependent upon Labour success. Second, Labour polled consistently below National for the duration of 2008 – but not so far below that all commentators had given up predicting Labour success. Headlines such as, “Undecided voters can turn election”,[56] “Nats lead but only just, election could go down to the wire”,[57] and “Clark believer in Auckland voters”,[58] continued to raise the slight possibility of a Labour victory until the day prior to the election. These headlines may have given Green supporters the strategic incentive to turn their backs on the Green Party and vote Labour. What some (perhaps incorrectly) perceived as a close race between National and Labour may have reduced Green Party success; third party squeeze operated to the Green’s detriment.
Third party squeeze is also observable in New Zealand at a level higher than individual party politics. Since the implementation of MMP, there is a trend towards minor parties receiving a reduced vote share when the race between National and Labour is tight, and an increased vote share when the race between the two major parties is one-sided. For example, of the five elections held under MMP, the 2005 election was indisputably the closest. Labour received a mere 2% more votes than National on election day.[59] Minor parties collectively received the lowest share of votes since MMP began – just 18.5%. Conversely, the most one-sided race since the advent of MMP was in 2002: Labour trounced National by a margin of 20.4%. Minor parties, however, did incredibly well; receiving their highest ever collective share of votes at 32.9%.[60] This pattern also holds true for elections at neither end of the spectrum: 2008, 1999, and 1996. When competition between Labour and National is neither one-sided nor very close, minor parties receive neither very high nor very low collective shares of the vote. It seems as though third party squeeze plays a role in shaping New Zealand’s political landscape.
Wasted Vote Phenomenon
Despite the seemingly sensible nature of the wasted vote phenomenon, it is questionable whether it occurs in New Zealand. Two elements of the 2008 election are particularly illustrative in this regard. First, 58,105 individuals, or 2.48% of all voters, cast a party vote for a party that did not once register on any of the major elections polls.[61] It is unlikely that any of these people expected their vote to gain their preferred party representation in parliament.
The second element tending to disprove the existence of the wasted vote phenomenon in New Zealand is the lack of movement on election day away from parties that had been polling below the 5% threshold. In the week leading up to the 2008 election, New Zealand First, the Maori Party, Act, United Future and the Progressives were all polling below the crucial 5% threshold. Despite this, all five parties increased on their final week’s polling average on election day.
Of course there is another ‘safety net’ to representation in New Zealand. Electorate seats were won by Maori Party, Act, United Future and Progressive candidates. The widely expected nature of these wins could be the reason there was no move away from these parties despite low party vote expectations. New Zealand First, however, did not win an electorate seat and was widely seen in the lead up to the election as being unable to. Despite this, the Party still increased on its final week’s election poll average of 3.44% support to finish the election with 4.07% of votes: an increase of 0.57%. A further 95,356 votes (on top of the 58,105 mentioned above) were effectively wasted at the 2008 election.[62]
This is consistent with Bowler, Karp and Donovan’s findings in their study of the 2002 New Zealand general election. Throughout that campaign, minor parties were polling around the 5% threshold, “so there was sufficient ambiguity to raise some doubt that they may not in fact secure representation”.[63] Despite this, and contrary to expectations, it was found that citizens are not discouraged from casting a vote even when sceptical about the likelihood of their preferred party securing representation in parliament.[64] New Zealanders (or at least 153,461 of them in 2008) are not afraid of “wasting” their vote.
Gschwend’s coalition insurance strategy, additions to that theory, and third party squeeze have been used to indicate the existence of strategic voting in New Zealand. The wasted vote phenomenon, in contrast, seems to hold little weight in our country. Whether this is because our inclusion of proportionality gives citizens a hope that does not exist in pure plurality systems, or because we have “personal devotees” of particular parties,[65] who are simply not concerned with election polls or other indicators of success is outside the scope of this dissertation.
Conclusion
The 2008 New Zealand general election provided indications of the existence of each of the three election poll influences. The National Party seems to have benefitted from the bandwagon and secondary bandwagon effects, while Labour and New Zealand First showed increases in support attributable to the underdog effect. Likewise, National and Labour electorate candidates, the New Zealand First, Act, and United Future Parties, and Jim Anderton – the Progressive candidate in Wigram – benefited from strategic voting. The Green Party, however, was likely negatively impacted by third party squeeze. It has therefore been indicated that exposure to election poll results influenced citizens’ vote choice on election day in 2008. This conclusion necessitates the evaluation of whether election polls in fact interfere with democratic integrity, and whether a prohibition on the publication of election poll results for a period of time prior to an election is required.
[1] C Robinson, “2008: Images of political leadership in the campaign,” in Key to Victory: the New Zealand general election of 2008, ed. S Levine and N Roberts (Wellington, NZ: Victoria University Press, 2010), 141.
[2] Note: While National won the election by a margin of 11% over Labour, election polls suggested the margin would be more like 16%. This discrepancy is explained by strategic voting. Strategic voters likely saw National was in a comfortable position, and moved to support an ideologically compatible minor party. See the discussion on Strategic Voting and particularly the coalition insurance strategy below.
Chief Electoral Office, “Official Count Results – Overall Status 2005,” Elections New Zealand, http://www.electionresults.govt.nz/electionresults_2005/partystatus.html (accessed 21 August, 2011).
[3] NZES, “Results from the 2008 NZES: Issues, Problems, Parties and Leaders,” New Zealand Election Study, http://www.nzes.org/exec/show/freq_2008c (accessed 22 September, 2011).
[4] Ibid.
[5] NZES, “Results from the 2008 NZES: Party Preferences and Voting,” New Zealand Election Study, http://www.nzes.org/exec/show/freq_2008e (accessed 22 September, 2011).
[6] Ibid.
[7] Ansolabehere and Iyengar, “Of Horseshoes and Horse Races,” 415. See also: Henshel and Johnston, “The Emergence of Bandwagon Effects.”
[8] Mehrabian, “Effects of Poll Reports on Voter Preferences.”
[9] Bahador, “New Zealand Media Coverage of the 2008 Election Study,” 3.
[10] Ibid., 9.
[11] Note: Only “large donations” are declared by political parties. The law in force at the time set “large donations” as anything greater than $10,000.
[12] Ansolabehere and Iyengar, “Of Horseshoes and Horse Races,” 415. See also: Henshel and Johnston, “The Emergence of Bandwagon Effects.”
[13] Elections New Zealand, “Party donation returns,” Elections New Zealand, http://www.elections.org.nz/rules/parties/party-donations/donations-summary.html (accessed 25 September, 2011).
[14] Ansolabehere and Iyengar, “Of Horseshoes and Horse Races,” 415. See also: Henshel and Johnston, “The Emergence of Bandwagon Effects.”
[15] Henshel and Johnston, “The Emergence of Bandwagon Effects,” 498.
[16] J Kirkwood, “Tall Poppy Syndrome: Implications for entrepreneurship in New Zealand,” Journal of Management & Organisation” 13 (2007): 366-382, 366.
[17] See Appendix I.
[18] See Appendix I.
[19] See Appendix I.
[20] NZES, “Results from the 2008 NZES: Issues, Problems, Parties and Leaders.”
[21] NZES, “Results from the 2008 NZES: Party Preferences and Voting.”
[22] Ibid.
[23] NZPA, “Timeline of Events in NZFirst Donations Row,” NZPA, 29 August, 2008.
[24] NZES, “Results from the 2008 NZES: Issues, Problems, Parties and Leaders.”
[25] NZES, “Results from the 2008 NZES: Party Preferences and Voting.”
[26] See Appendix I.
[27] NZPA, “Glenn donated $100,000 towards legal costs, Peters says,” NZPA, 18 July, 2008.
[28] D Fisher, “Winston’s woes,” NZ Listener, 2 August ,2008.
[29] NZPA, “Timeline of Events in NZFirst Donations Row.”
[30] Ibid.
[31] Ibid.
[32] Ibid.
[33] Newstalk ZB, “Businessman Owen Glenn is again challenging Winston Peters and Brian Henry’s account of how he ended up donating$100,000,” Newstalk ZB, 19 September, 2008.
[34] NZPA, “Timeline of Events in NZFirst Donations Row.”
[35] The Press “Peters lashes out after NZ First cleared in donations probe,” The Press, 5 November, 2008.
[36] NZPA, “NZ First into the breach over breaches this week,” NZPA, 7 September, 2008.
[37] Dominion Post, “Now, had it gone the other way...,” Dominion Post, 26 September, 2008.
[38] NZPA, “Peters launches campaign with attack on National and media,” NZPA, 5 October, 2008.
[39] NZPA, “Electoral Commission Clear NZ First,” NZPA, 24 October, 2008.
[40] Ibid.
[41] The Press, “Peters lashes out after NZ First cleared in donations probe.”
[42] Note: Correlations appear imperfect on Graph 1 because polls are conducted over long periods, and campaign events do not match these periods perfectly.
[43] S Bowler, T Donovan and J Karp, “Strategic Voting Over Coalition Governments: The Case of New Zealand” (paper presented to the Workshop on Voters, Coalitions, and Democratic Accountability, Exeter, October 5-6, 2007).
[44] Ibid., 7.
[45] Ibid., 11.
[46] NZES, “Results from the 2008 NZES: Party Preferences and Voting.”
[47] Ibid.
[48] NZES, “Results from the 2008 NZES: Issues, Problems, Parties and Leaders.”
[49] The Press, “Maori Seats Puzzle,” The Press, 25 October, 2008.
[50] NZES, “Results from the 2008 NZES: Party Preferences and Voting.”
[51] New Zealand Parliament, “Electorate profiles: Wigram,” New Zealand Parliament, http://www.parliament.nz/en-NZ/MPP/Electorates/EPData/9/3/8/DBHOH_Lib_EP_Wigram_Data_3-Wigram-Electoral-Profile.htm (accessed 1 September, 2011).
[52] Cain, “Strategic Voting in Britain,” 639.
[53] NZES, “Results from the 2008 NZES: Issues, Problems, Parties and Leaders.”
[54] Chief Electoral Office, “Official Count Results – Overall Status 2008,” Elections New Zealand,
http://www.electionresults.govt.nz/electionresults_2008/partystatus.html (accessed 1 August, 2011).
[55] Dominion Post, “Greens rule out National,” Dominion Post, 21 October, 2008.
[56] The Press, “Undecided voters can turn election,” The Press, 7 November, 2008.
[57] NZPA, “Nats lead but only just, election could go down to the wire,” NZPA, 6 November, 2008.
[58] NZPA, “Clark believer in Auckland voters,” NZPA, 7 November, 2008.
[59] Chief Electoral Office, “Official Count Results – Overall Status 2005.”
[60] Chief Electoral Office, “Official Count Results – Overall Status 2002,” Elections New Zealand,
http://www.electionresults.govt.nz/electionresults_2002/partystatus.html (accessed 1 August, 2011).
[61] Chief Electoral Office, “Official Count Results – Overall Status 2008.”
[62] Ibid.
[63] Bowler, Karp, and Donovan, “Strategic coalition voting,” 355.
[64] Ibid.
[65] Pearse, “Making the most of a two vote ballot,” 2.
