Complaint by David Farrar from KiwiBlog

I very rarely lodge formal complaints with the media. In fact I am unsure if I have ever done a formal complaint, as I prefer to just critique stories on Kiwiblog.

But this story in the Dominion Post and on Stuff was so bad, I have filed the below complaint. I will advise of any response [it will just tell him in effect to Stuff off —Eds].

Anna Fifield
The Editor
The Dominion Post

Dear Anna,

I write to complain about the article titled “Poll: Whanau clear favourite” which appeared on Page 3 of the Dominion Post on 12 September 2022, and an online version at https://www.stuff.co.nz/dominion-post/news/wellington/129847685/tory-whanau-clear-leader-in-straw-poll-for-wellington-mayoralty?

The online version is slightly better than the print version, as the headline is more nuanced.

The poll reported by the Dominion Post was not a poll, but an online survey of Dominion Post and Stuff readers (and non-readers who were sent a link). It has no predictive value, and limited news value. The survey is not representative due to the following factors:

  • Respondents were not randomly selected
  • It will be dominated by people who read Stuff and are active online
  • It is trivial to vote multiple times (stealth browsers etc)
  • Candidates can e-mail links to their supporters asking them to vote
  • Respondents may be ineligible to vote by way of being aged under 18 or living outside Wellington City

The New Zealand Political Polling Code — pdf — advises media that best practice is to “Only use the term “poll” for scientific polls done in accordance with market research industry approved guidelines, and use “survey” for self-selecting surveys such as text or website surveys.”

The print story does acknowledge the survey is unscientific but not until several paragraphs in, by which time the damage is done. To take the story in order:

  • The headline of “Poll: Whanau clear favourite” is misleading as readers who only see the headline will think that a poll which is representative of residents has shown Whanau as the favourite. The lack of any qualification before the word “poll” in the headline is especially bad. An acceptable headline could have been “Online survey of readers favours Whanau”
  • The lead paragraph states “A straw poll of Wellington mayoral candidates has first-time runner Tory Whanau with a huge majority over her two main rivals” will lead readers to think Whanau is not just leading, but set to win with a massive majorityThe use of the word “straw” won’t signal much to the average reader, or clearly indicate it is not representative
  • The second paragraph states “It is intended to give an indication of which candidates are viable”. The self-selecting survey is incapable of indicating which candidates are viable. It can only indicate who is popular amongst those who choose to take part in the self-selecting survey. To suggest it has predictive value for the election is wrong.
  • The third paragraph states “It only allows one vote per computer”. This is wrong. It is trivial to get past any restriction based on cookies or IP addresses.
  • The fourth paragraph states “It is just one snapshot in time, rather than a scientific survey”. Almost all polls are snapshots in time. It is not the fact it is a snapshot in time that makes it unscientific. It is the nature of the survey.
  • The fourth paragraph states “No other known polls have yet been released.”. This elevates this unscientific survey to the level of scientific representative polls
  • The story also asked three candidates for the reaction to the survey. This again implies that the survey was somehow something of significance that should be reacted to.

I do not advocate that media should not report at all on the results of their online surveys. But I would expect any report to be relatively brief just along the lines of “The results of our online readers survey was ….”. But the Dominion Post:

  1. Wrote it up as a major story dominating the highly read Page 3
  2. Ran a headline that was deceptive
  3. Treated the survey as significant, and needing candidates to respond to it

I can’t recall any other major media outlet ever treating an survey of this nature as a story that deserves around 75% of their page 3.

In just the last 12 hours I have had multiple people comment to me on the survey, not realizing it was an unscientific survey rather than a scientific representative poll. I have no doubt that the story has led to a huge degree of misinformation.

Potentially tens of thousands of Wellingtonians have been misled by the headline and details of the news story. The brief mentions part way through of it being unscientific did not counter the overall impression that this survey was important enough to be reported on so prominently

I would ask the Dominion Post to take corrective action with regards to this story. I say this noting that any corrective action may not be able to undo the misinformation effect of the story. It may have a significant impact on how people vote.

The best outcome would be for the Dominion Post to commission an actual scientific poll to give readers a true snapshot of the race of the mayoralty and undo the damage from your story.  And for the avoidance of doubt, no I am not touting for business.

If that is not feasible, I would suggest a follow up story with similar prominence to the original story apologizing for the headline and implications in the original story that the online survey could be interpreted as indicating which candidates are viable, or leading.

For the avoidance of doubt, I am asserting that the story breached principles 1 and 6 of the Media Council Statement of Principles.

I believe principle 1 is breached as the story was inaccurate and misled readers as to the significance of the survey and the public support for the mayoral candidates.

I believe principle 6 is breached as the headline lacked any reference to the unscientific nature of the survey, by labelling it as a poll – contrary to the explicit industry best practice guideline by the polling industry.

Original article