GAVIN MILLAR AGAINST THE NORTHERN ADVOCATE
Case Number: 3244
Council Meeting: March 2022
Publication: Northern Advocate
Balance, Lack Of
Comment and Fact
1.Gavin Millar complains that Dr Shane Reti in a regular column in the Northern Advocate was the author of a misleading headline and graphs about the relationship between hospital admissions for self-harm Pre-Covid and during Covid published in hard copy and on-line. The graphs were wrong and the headline therefore inaccurate. Following the errors being pointed out, the Advocate published articles of “Clarification”, and removed the article from its on-line publication. The clarifying articles were not an adequate correction, and indeed Dr Reti was given an opportunity to make further comments about mental health and covid restrictions to the same effect as before. The complaint is upheld.
2. Dr Shane Reti publishes a regular column in the Northern Advocate, an NZME publication. His on-line column published on 7 March 2022 was headed “No Covid victory parade for mental health”. The print version was headed “Mental toll of Covid era” and shows a picture of a man with his head in his hands. Both the on-line and print versions contained graphs which show admissions for intentional self-harm from February 2019 to February 2020 as increasing at a much more gradual rate than the rate of increase between March 2020 and June 2021. The much steeper line is shown for the latter Covid period.
3. It is to be noted that when he published the graphs, Dr Reti stated that “the following graphs require further data analysis but seem to point to a story of generally increasing intentional self-harm”. After complaints, a “clarifying” article was published online by the Northern Advocate on the same day, and in the print edition the next day. The Shane Reti column was removed online.
4. Mr Gavin Millar complains that the graphs were deliberately misleading. NZME should have checked them. He says there were intentional faults in the graphs, from using two charts, changing the time periods, and using a linear regression to cover up data points that disagree with the thesis of the column. Mr Millar says that the response by NZME was not sufficient as it was argued that the point of the column was still valid, even though the theory that self-harm has increased due to the Lockdown was “ridiculous”.
5. In response it is said that the article was an opinion piece, and that the Advocate believed that the data used by Dr Reti was accurate. As soon as concerns were raised by readers, the graphics and reference to them were removed from the column. Shortly after that, they agreed with Dr Reti to remove the whole article from the website. They also published a clarification on their website that day, and in print on page 3 the next day.
6. It denies any intention to mislead. It published an opinion piece from the Health Minister Andrew Little on 14 March 2022, where he set out views on the data used by Dr Reti.
7. The column was marked “From Parliament”, and Mr Millar argues it is not an opinion piece because it is not labelled as such. However, an opinion piece, if that is what it plainly is to a reader, does not have to be labelled “Opinion”. Here the heading “From Parliament” was a clear indication that what was to follow was from a politician. A politician would be expected to be putting forward a political point of view. As a high-profile opposition politician speaking on a controversial topic, nothing less would be expected. This was plainly his opinion.
8. That it is an opinion piece is not the end of the complaint. Opinion pieces must be based on a foundation of fact, and should not contain clear errors of fact. There is no doubt that the graphs presented by Dr Reti as justifying his headline and comment in his column were misleading in that they did not fairly show the trend in cases through the period. For the covid period graph just two data points were used from the source material, rather than all the data points. The mistake appears to be that the data points are based on contrasting figures that are months apart, with no recognition of the monthly figures in between. Even that last data point shown in the misleading second graph, shows an inaccurately high end-point for June 2021.
9. If all the data points are used, they do not indicate the same increase of mental health admissions as shown in the charts. Therefore, the actual figures do not support the thesis that during covid mental health in New Zealand has markedly deteriorated. Dr Reti himself accepts there are issues with his data. The Northern Advocate does not argue that the graphs were not erroneous.
10. Therefore, the publication, while clearly an opinion piece, was inaccurate. Principle 5 of the Media Council principles require opinion pieces to not contain a clearly wrong factual basis.
11. While publishers of opinion pieces are expected to examine what they propose to publish and query and remove any obviously incorrect facts, they are not expected to investigate every asserted fact or graph or figure.
12. However, if a fact that was not obviously wrong is shown to be wrong after publication, immediate steps must be taken to correct it. If that is done and with fair prominence, the correction may be able to defuse the complaint and what would have been an upheld complaint can become a complaint that is not upheld.
13. We should add that we are satisfied that the Northern Advocate did not know that the graphs were not based on satisfactory data. There is nothing to indicate that they had the source data. The issue is whether when they got information indicating the error in the charts, they took adequate steps to ameliorate the error.
14. In his final response Mr Millar complained that Dr Reti in his twitter feed continues to publish this particular graph in the Herald
styling. This is correct, and the graphs are still there, despite Dr Reti’s acknowledged reservations about them. However,
although he retains the Northern Advocate reference, the Northern Advocate will have had no control over a private twitter
account. We do not take into account the continued presence of the graphs in the private twitter account in assessing this
complaint. Such a private posting is outside our jurisdiction, which is limited to publications by media organisations.
Were the actions of the Northern Advocate sufficient to defuse the error?
15. At 6.02 pm on the day of publication of the erroneous column and graphs, and after complaints, the Northern Advocate published an article headed “Dr Shane Reti’s column and graphics removed for more analysis”. It stated that the graphics produced by Dr Reti “which the Advocate published in good faith” were “subsequently called into question”. It stated that the Advocate and Dr Reti agreed to remove the article from the newspaper’s website until he clarified data. Dr Reti wanted to “double check” it. It quoted him saying “After doing so I consider the focus of the article is correct – although the graphs I produced could have been better directed”. It then quoted this from Dr Reti, where he adds to his theme that following covid mental health problems increased:
a. “Data provided by the Ministry of Health confirms the number of hospitalisations for self-harm for 10 – 14-year-olds increased by more than 30 per cent follow[ing] Covid arriving”
16. The erroneous column is not now available on the internet, and we assume it has not been since later on the day of publication. The hard copy edition of course could not be withdrawn.
17. The clarification was published in the hard copy newspaper the next day, under the heading “Reti article and graphs: Clarification”. The clarification states that the graphics used have been called into question and that the article has been removed from the Advocate website. It quotes the statements of Dr Reti set out in paragraph  above.
18. There have been articles by Marc Daalder in Newsroom (11 March 2022) and Stuff (14.46 7 March 2022), referring to the charts, and the fact that there are different interpretations of the data, and that experts from the Herald have said that the data was misused. The Stuff article was headed, “Shane Reti article on mental health deleted after data in it attacked as misleading”. It quoted a number of reputable persons including the NZ Herald data journalist Keith Ng criticizing the graphs. A quick scan of the actual data quickly reveals the errors.
19. The Northern Advocate had withdrawn the article, and stated that more analysis was required. It gave Dr Reti the last word, with
his statement that the graphs could have been better directed, but allowing him to make a similar point using a different statistic.
There is force in Mr Millar’s observation in his final response that this is hardly a retraction, but rather provides a “further
opportunity for Dr Reti to make incorrect statements”.
20. The implication of the Northern Advocate “clarification” is that Dr Reti is not convinced he is wrong, and he is allowed to make a further re-enforcing point. Having received the complaint, the Northern Advocate should have checked the data (which as stated is not complex), and published a clear correction, and not allowed the response to be fudged. While the article was promptly removed, and the data was stated to be “in question”, a correction if it is to assuage error should be plainly a correction. A correction is the rectification of an error or inaccuracy. It is not a “clarification” allowing the error to stand doubted but unrectified.
21. The Council concludes that the Northern Advocate did not do enough to correct the error. The article was factually inaccurate, and readers were not presented with a correction of the error.
22. The complaint is upheld.
Media Council members considering the complaint were Hon. Raynor Asher (chair) Hank Schouten, Tim Watkins, Jonathan MacKenzie, Jo Cribb, Marie Shroff, Liz Brown, Katrina Bennett, Reine Vaai and Richard Pamatatu.