11445-22 Arnott v express.co.uk

Decision of the Complaints Committee – 11445-22 Arnott v express.co.uk

Summary of Complaint

1. Richard Arnott complained to the Independent Press Standards Organisation that express.co.uk breached Clause 1 (Accuracy) of the Editors’ Code of Practice in an article headlined “Nicola Sturgeon suffers devastating polling blow as SNP support collapses”, published on 31 August 2022.

2. The article, which appeared online only, reported on a recent poll of 2,000 adults in the UK; the poll asked respondents about their voting intentions in the event of a general election. The article focused on the Scottish National Party (SNP) and stated that the poll results “will come as a huge blow to Ms Sturgeon, who has threatened to use the next Westminster election as a proxy vote on Scottish independence. […] The SNP leader suggested she would have a mandate for independence if support for pro-independence parties topped 50 percent at the ballot box.” The sub-headline of the article stated: “Support for the SNP has collapsed, new polling suggests, with just 27 percent of Scots suggesting they will vote for the party at the next general election.” The article further reported: “[market researcher] spoke to 2,000 UK adults on August 28. However, only 180 of those spoken to were in Scotland, meaning the results should be treated with caution.”

3. The complainant said that the headline and sub-headline were inaccurate and misleading in breach of Clause 1. The complainant said the publication had based their headlines on a survey conducted across the whole of the UK, with only 180 of those polled being in Scotland. He said that the small number of Scottish people polled meant that the conclusions drawn by the publication were inaccurate, and also implied that 27% of Scottish people would vote for independence. The complainant said that a  sentence in the body of the article explaining that the figures should be read with caution did not remedy the misleading headline. He said that it could not suggest a drop in support for the SNP based on the size of the data sample used, and that the publication was unable to provide other examples of articles where it had used similarly small sample sizes to make such headline claims.

4. The publication said it did not accept a breach of Clause 1. It stated that the poll was open to respondents from the whole of the UK, of which 180 were from Scotland. Out of the 180 Scottish respondents polled, 27% said they would vote for the SNP. The publication said that the complainant had acknowledged that the article made clear that only 180 people from Scotland voted in the poll. It also said that due to the small sample size, the article had stated “the results should be treated with caution.”

5. The publication said that straw polls are often used in national newspapers and that the headline, specifically referenced a “devastating polling blow” which accurately reflected how damaging the results of a general election would be if the poll were to be replicated in a national vote. It said that care had been taken in the sub-headline which stated that the poll "suggest[ed]" a drop in support for the SNP in order to not be misleading and suggest that the result was unambiguously explicit.

6. Furthermore, the publication provided the raw poll data set which the article was based on to show it had taken care not to publish significantly inaccurate information: it contained the methodology, a breakdown of the statistics according to characteristics such as age, religion, and region. The data was also weighted according to factors such as the respondent’s likelihood to vote, age, gender, region, education, and 2019 General Election vote. In response to the question “If there were to be a General Election in the United Kingdom tomorrow, for which party would you vote?”, in a sample of 2,000 people, 156 people in Scotland were asked who they would vote for, and this subset was calculated to be a weighted amount equal to 180 people.

7. The data which the publication had highlighted on the raw data set had calculated the 27% figure by incorporating the weighting of “Likelihood to Vote” and including those who did not know which party they would vote for. This said that a weighted 42 people who lived in Scotland said they would vote SNP out of a weighted 153 people.

8. The publication said that the polling company was a member of the British Polling Council (BPC) which was committed to promoting transparency in polling and upholding standards within the sector and that this was widely seen as a mark of quality within the industry. It said the polls conducted by the company had previously been reported on by a range of media outlets including the BBC, ITV, Sky and several national newspapers. It said it was satisfied that the pollster and its methodology were suitably reliable and that it was entitled to rely on its data. Further to this, it said news outlets often relied on the breakdown of data in such polls, which it had done in this example geographically.

9. While the publication did not accept a breach of Clause 1, it offered to remove the word “devastating” from the headline as a gesture of goodwill.

10. The complainant did not accept the publication’s proposed course of action would resolve their complaint. Having reviewed the raw data the publication provided, the complainant said the survey was conducted UK-wide and not just in Scotland so the overall results were unreliable. He said the pollsters had only asked 128 people in Scotland for their views, which would not give an accurate result. He said that an ideal sample size giving a +/-3% margin of error based on 4,245,200 registered voters in Scotland would be 1842, not 128.

Relevant Code Provisions

Clause 1 (Accuracy)

i) The Press must take care not to publish inaccurate, misleading or distorted information or images, including headlines not supported by the text.

ii) A significant inaccuracy, misleading statement or distortion must be corrected, promptly and with due prominence, and — where appropriate — an apology published. In cases involving IPSO, due prominence should be as required by the regulator.

iii) A fair opportunity to reply to significant inaccuracies should be given, when reasonably called for.

iv) The Press, while free to editorialise and campaign, must distinguish clearly between comment, conjecture and fact.

Findings of the Committee

11. The complainant believed that the headline and sub-headline were inaccurate because – although it was part of a larger poll – the “27%” claim was based on a sub-group of only 128 respondents in Scotland. His concern was that the number of respondents was too small to draw robust conclusions and therefore could not support the headline claim that the SNP had suffered a “devastating polling blow”.

12. The Committee noted that its role was to establish whether the headline and sub-headline were significantly inaccurate or misleading and whether, in line with Clause 1(i), the headline was supported by the article. In addition, it needed to determine whether the publication had taken care on reporting the statistics accurately within the article, rather than making a judgement on the methodology used in the poll.

Firstly, the Committee noted that the headline referenced the basis for its characterisation of “SNP support collapse[ing]” where it mentioned “polling blow” and the sub-headline further clarified this by stating “Support for the SNP has collapsed, new polling suggests, with just 27 percent of Scots suggesting they will vote for the party at the next general election”. The article later set out the details of the poll itself: the name of the polling company and its sample size,  “[market researcher] spoke to 2,000 UK adults on August 28. However, only 180 of those spoken to were in Scotland, meaning the results should be treated with caution.” Therefore, the headline was sufficiently supported by both the sub-headline and article, and there was no breach of Clause 1.

13.  The Committee considered the statistics which the article had referenced and the raw data the publication had provided to show it had taken care over the accuracy of the article. The raw data showed that 156 people who resided in Scotland had been asked the relevant questions and that the “180 of those spoken to” referred to in the article had been a weighted figure. The Committee acknowledged that both figures were small relative to the size of Scotland’s population and that the article had not made clear 180 people was a weighted figure. However, where the text of the article had noted its limitations by stating the sample size, and where the 27% figure was an accurate calculation of the percentage of people who lived in Scotland who said they would vote SNP from the weighted figures, the Committee did not consider this reporting to be significantly inaccurate or misleading. The publication was entitled to use the 27% figure as a basis for the headline’s claim, and there was no breach of Clause 1.

14. The Committee noted that the complainant also appeared to be concerned that the article had gone further than simply describing Westminster voting intentions and implied that 27% of Scottish people would vote for independence. However, the article did not suggest that 27% of Scottish people would vote for independence; it simply stated “27 percent of Scots suggest[ed] they will vote for [the SNP] at the next general election.” There was no breach of Clause 1 on this point.

Conclusion(s)

15. The complaint was not upheld.

Remedial Action Required

16. N/A

Date complaint received: 31/08/2022

Date complaint concluded by IPSO: 31/01/2023