BSA Decisions Ngā Whakatau a te Mana Whanonga Kaipāho

All BSA's decisions on complaints 1990-present

Dunbar and Television New Zealand Ltd - 2005-108

  • Joanne Morris (Chair)
  • Diane Musgrave
  • Tapu Misa
  • Paul France
  • Geoffrey Dunbar
One News
TV One

Complaint under section 8(1)(a) of the Broadcasting Act 1989
One News – Colmar Brunton poll surveyed voters’ party vote preferences – did not make correct assumption about likely Māori Party result – use of poll data in “virtual Parliament” format allegedly misleading and inaccurate

Standard 5 (accuracy) – poll relied on reasonable assumptions – not upheld

This headnote does not form part of the decision.


[1] TV One broadcast political items on One News at 6pm on 28 August and 4 September 2005. The items reported the outcome of two political polls conducted for Television New Zealand Ltd, by research company Colmar Brunton.

[2] Both items reported how the outcome of the polls would translate to the make-up of a new Parliament, using a “virtual Parliament” to illustrate how many seats each party might win in the forthcoming election.


[3] Mr Dunbar complained to TVNZ, the broadcaster, that the items failed to consider the potential impact of the Māori Party on the make-up of the next Parliament.

[4] He argued that the reports of the polls misrepresented the situation by anticipating the Māori Party gaining only one or two representatives in Parliament. Because the reports did not allow for the Māori electorate seats or the preferences of voters on the Māori Roll, Mr Dunbar contended, TVNZ’s analysis persisted in conveying a distorted picture of possible election results.

[5] He further stated that TVNZ’s poll results consistently understated the Māori Party’s likely seats in the next Parliament. By not at least issuing a disclaimer to the effect that Māori electorates were not polled, he stated, TVNZ offered a seriously distorted view of voter preferences. In his e-mail of 4 September 2005 to TVNZ, he contended that both items were electoral programmes and accordingly should be dealt with within 48 hours.


[6] TVNZ assessed the complaint under Standard 5 of the Free-to-Air Television Code of Broadcasting Practice, which provides:

Standard 5 Accuracy

News, current affairs and other factual programmes must be truthful and accurate on points of fact, and be impartial and objective at all times.

Broadcaster's Response to the Complainant

[7] Prior to its decision on his formal complaint, TVNZ advised Mr Dunbar that it considered the two items did not fall within the definition of “electoral programme” contained within s69 of the Broadcasting Act 1989, and accordingly declined to process the complaint on that basis.

[8] In its response to the complaint, TVNZ asserted that each poll was a “snapshot”, rather than a prediction, of the make-up of the next Parliament, based entirely on evidence available from the Colmar Brunton surveys. These snapshot polls assumed that three particular candidates would win electorate seats and thus their party would not need to cross the 5% party-vote threshold to gain parliamentary representation – Jim Anderton for the Progressive Party, Peter Dunne for United Future, and Tariana Turia for the Māori Party.

[9] TVNZ stressed that the poll was a measure of the party vote, and was intended to show how seats would be allocated, assuming that there would be 120 seats in the new Parliament. Accordingly, the poll contained no data measuring the likelihood or otherwise of the Māori Party gaining any “overhang” seats, which would result in a Parliament with more than 120 seats.

[10] The broadcaster observed that the two polls allocated the Māori Party one or two per cent of the party vote. This meant that, because Tariana Turia was certain to retain her electorate seat, if an election had been held on either of the two occasions the polls had been conducted, the Māori Party would have been allocated one or two seats through the party vote, TVNZ said.

[11] It noted that while TVNZ’s political editor could safely predict the presence of the Māori Party (as well as the Progressive Party and United Future) in Parliament, TVNZ maintained, the poll did not contain data which would allow speculation on what might arise as a result of voting in Māori electorate seats.

[12] TVNZ stated that in both polls, the number of seats shown on the virtual floor plan of Parliament added up to 120. It considered that to explore the possibility of a minor party increasing their representation through overhang seats was outside of the scope of the polls and thus the news items.

Referral to the Authority

[13] Dissatisfied with the response from the broadcaster, Mr Dunbar referred his complaint to the Authority under s8(1)(a) of the Broadcasting Act 1989. He alleged that the items were inaccurate and resulted in unfairness.

[14] The complainant alleged that the item contained two key inaccuracies:

  • TVNZ gave the impression that the Māori Party party vote was relevant to the seats that it would potentially gain in Parliament, when to a large extent it was not
  • TVNZ created the misleading impression that the Māori Party would be only minimally successful in electorate seats, ignoring evidence from other sources suggesting it would win between three and seven electorate seats in the House.

[15] The complainant observed that Colmar Brunton sampled the voting public as a whole, and that the use of the result was up to the client – in this case, TVNZ. He asserted that the Colmar Brunton poll made no effort to include a meaningful sample of the voting preferences of voters on the Māori roll who intended to vote in Māori electorates.

[16] Mr Dunbar compared the Colmar Brunton poll results to another research company’s (ACNeilsen) results. He noted that these polls included Māori seat results as well as an estimate of nationwide party-vote preferences. In support of this argument, he cited results for an ACNeilsen/Sunday Star-Times poll which gave the Māori Party only 0.4% of the party vote. However, he stated, the same poll awarded the Māori Party four seats in Parliament, based on projected wins in electorate seats.

[17] Mr Dunbar considered that a simple disclaimer in each of the items, to the effect that the Colmar Brunton poll was inapplicable to Māori Roll seats would have “successfully blunted charges of inaccuracy and unfairness”.

Broadcaster’s Response to the Authority

[18] In its response to the referral, TVNZ emphasised that the information Mr Dunbar complained was missing in the items was not available from the data produced by the poll. It considered that, unlike the constituency seats of Tariana Turia, Jim Anderton and Peter Dunne, the election of additional Māori Party MPs could not be assumed, and there was no data in the poll to indicate such an event occurring.

[19] TVNZ further asserted that the Colmar Brunton poll essentially measured the party vote. It considered that as far as the Māori Party was concerned, this meant one or two per cent of the electorate. It argued that the poll did not measure electorate contests, and thus there was no basis on which the additional seats that might be gained by the Māori Party could be calculated.

[20] TVNZ maintained that, contrary to the complainant’s view, neither item was an “election programme” as defined by the Broadcasting Act 1989.

Complainant’s Final Comment

[21] In his final comment, Mr Dunbar reiterated his concern that TVNZ had presented an inaccurate and ultimately unfair analysis of polls surveying voter’s preferences in the 2005 election.

[22] He noted that TVNZ’s polling company had made no effort to discern whether voters were enrolled on the general or Māori roll. He considered that this implied that electorate results could be measured by trends in the nationally based party vote, which, he maintained, was inaccurate.

[23] Mr Dunbar considered that allocating one seat to the Māori Party in one poll and two in the other compounded the inaccuracy. In doing this, he stated, the poll awarded the other five or six Māori electorate seats to Labour. He observed that no polls of Māori voters had ever projected such a result.

[24] The complainant stated that the “minimal achievement” attributed by the poll to the Māori Party was potentially unfair in that it might have contributed to some voters switching their party vote preference to a party they perceived as more successful.

[25] Mr Dunbar maintained that, given the nature and number of the Māori seats, the characterisation within a poll of a 120 seat Parliament inevitably resulted in inaccuracies and, in this case, unfairness. He considered that this could have been avoided by warning viewers of the limitations of the poll’s data. Further, he wrote, inaccuracies could have been avoided by depicting a Parliament of 114 – including Ms Turia’s “safe seat”, but excluding the six seats which had not been surveyed by Colmar Brunton.

[26] Moreover, he noted, in the interests of full and accurate reporting, TVNZ could have accompanied the qualified Colmar Brunton poll result analyses with poll results specific to Māori electorate seats.

Authority's Determination

[27] The members of the Authority have viewed a tape of the broadcasts complained about and have read the correspondence listed in the Appendix. The Authority determines the complaint without a formal hearing.

Procedural matters

[28] The Authority has on a number of occasions interpreted its governing legislation, the Broadcasting Act 1989, as barring complainants from raising new standards when referring a complaint to the Authority. They may do so, the Authority has ruled, only if the standard is raised in the initial complaint. On this occasion, the Authority does not accept that Standard 6 (fairness) was raised either implicitly or explicitly in the initial complaint. Accordingly, it confines its deliberations to Standard 5 (accuracy).

[29] Further, the Authority considers that the item was simply a presentation of poll data and did not advocate, oppose or publicise any political standpoint. Consequently, the Authority agrees with the broadcaster that the item did not fall under the definition of “electoral programme” contained in s.69 of the Broadcasting Act 1989.

Standard 5 (accuracy)

[30] As a preliminary point, the Authority notes that no issues of accuracy arise from the simple reporting of the poll results. In his complaint Mr Dunbar does not suggest that the results of that poll inaccurately represented the then current state of the nationwide party-vote.

[31] The only issue arises in respect of the projection of this party vote into the virtual Parliament. For this exercise, the party vote was directly translated into seats in Parliament, while also assuming that Jim Anderton, Peter Dunne and Tariana Turia would win their electorate seats. It was also assumed that the Parliament would comprise 120 seats.

[32] The essence of Mr Dunbar’s complaint was that based on polls reported in other media at the time, the Māori Party was likely to win between three and seven of the Māori electorate seats. It would thus receive a share of seats in Parliament disproportionate to its share of the party vote. The virtual Parliament, he said, should have taken these seats into account.

[33] The Authority is of the view that to present a consistent picture of developments in public opinion about the election, it is reasonable for broadcasters to adopt certain assumptions at the outset of polling and to keep these assumptions consistent.

[34] In the present case, TVNZ adopted three assumptions in creating its virtual Parliament, and these assumptions were made clear to viewers in both programmes:

  • Parliament would be 120 seats, and there would be no “overhang” (when a party receives more electorate seats than its share of the party vote would otherwise entitle it to)
  • The share of seats was allocated on the share of the party vote (so long as the party reached the 5% threshold)
  • The only exceptions to this were the three safe electorate seats of Jim Anderton, Peter Dunne and Tariana Turia.

[35] The Authority recognises that polling is a speculative exercise and that any prediction of Parliament’s likely make-up will require some assumptions. The Authority considers that as long as those assumptions are made clear, even if they are contentious (as was the assumption that only Tariana Turia would win a seat for the Māori Party) the accuracy standard will not be breached. So long as the assumptions are understood, viewers will themselves be able to recognise the limitations of the prediction being made.

[36] In the present case the Authority considers that the basis on which seats in the virtual Parliament were allocated was made clear, and TVNZ’s assumptions were sufficiently explained to the viewer. Accordingly, the Authority considers that the accuracy standard was not breached.


For the above reasons the Authority declines to up hold the complaint.

Signed for and on behalf of the Authority


Joanne Morris
28 November 2005


The following correspondence was received and considered by the Authority when it determined this complaint:

  1. Mr Dunbar’s formal complaint to TVNZ – 28 August 2005
  2. Mr Dunbar’s second formal complaint to TVNZ – 4 September 2005
  3. TVNZ’s response to the formal complaint – 12 September 2005
  4. Mr Dunbar’s referral to the Authority – 13 September 2005
  5. TVNZ’s response to the referral – 20 September 2005
  6. Mr Dunbar’s final comment – 1 October 2005