28854-20 Ross v The Sunday Times

    • Date complaint received

      22nd April 2021

    • Outcome

      Breach - sanction: action as offered by publication

    • Code provisions

      1 Accuracy

Decision of the Complaints Committee 28854-20 Ross v The Sunday Times

Summary of Complaint

1. David Ross complained to the Independent Press Standards Organisation that The Sunday Times breached Clause 1 of the Editors’ Code of Practice in an article headlined “Sturgeon's surge tilts balance on independence”, published on 5 July 2020.

2. The frontpage article, which appeared in the Scottish edition of The Sunday Times, reported on a poll undertaken by Panelbase and commissioned by the publication. It reported that support for the Scottish Nationalist Party (SNP) and Scottish Independence “jumped to the highest level recorded”. It stated that “many people who used to oppose Scottish independence, have switched to the ‘yes’ camp, turning a 55%-45% vote against leaving the UK in the 2014 referendum into a 54%-46% lead for the nationalists”. The article went on to report that the poll of 1,026 voters in Scotland, conducted between Tuesday 30 June and Friday 3 July, gave the SNP a “commanding lead” before the Scottish Parliament elections and the highest level of backing in almost a decade of Panelbase polling for the publication. It also provided a breakdown of both the constituency and regional proportional votes by political party.

3. A substantially similar version of the article also appeared online with the headline “Nicola Sturgeon’s polling surge tilts balance on independence”.

4. The complainant said that this article had distorted the findings of the survey, in breach of Clause 1. Whilst the complainant acknowledged that the poll had clearly found higher support for Scottish independence than against it, he said that it did not find a “54-46” split as reported by the publication. In fact, the poll found that 45 per cent of the 1,026 respondents were in favour of independence with 38.8 per cent against it. The complainant said that in order to claim that the poll had found that “54%” of respondents supported Scottish independence, the publication had excluded the respondents who were “undecided” on the matter as well as those deemed “unlikely to vote”, then calculated the level of support as a split between the remaining respondents. This had, in total, according to the complainant, excluded 16 per cent of respondents, with 66 respondents “undecided” on the matter and a further 98 deemed by the polling company to be “unlikely to vote” as they scored between 1-7 on a scale where 1 was ‘Certain Not To Vote’ and 10 ‘Certain to Vote’. He said the newspaper’s failure to make clear that the respondents who were “undecided” and “unlikely to vote” were excluded from the analysis rendered the article a misleading report of the poll’s findings.

5. The publication did not accept that the article had distorted the overall findings of the poll or that it contained any significant inaccuracies. It said that, as a general principle, it was legitimate for reputable polling companies, such as Panelbase, to exclude “don’t knows” and those they judge unlikely to vote from their headline figures. It noted that this was a common and accepted practice across the polling industry and media alike. As such, it did not accept that readers would be misled, and maintained that the article presented the relevant information concisely and clearly. Whilst it did accept the complainant’s point that the article should have made clear the number of “undecided” respondents, it argued that further detailing the polling methodology, including the number of those deemed “unlikely to vote”, was unnecessary, impractical, and counter-productive. Nevertheless, upon receipt of the complaint, the publication offered to amend the online article to make clear that those who were undecided or unlikely to vote were excluded and to publish a correction in its established Corrections and Clarification column to address this.

6. The complainant said the publication’s offer was inadequate, as it failed to address what approach was used by Panelbase to determine which respondents should be excluded on the grounds that they were unlikely to vote, noting that the poll excluded almost 10 per cent of respondents for this. He said this was particularly salient given the subject matter and the multitude of varied approaches and methodologies used across the polling sector.

7. The publication noted the concerns raised by the complainant, acknowledging that they had given rise to a healthy discussion in the publication’s own Editorial team on the reporting of opinion polls. However, it stressed that there was a limit to the level of detail regarding polling methodology that is practical, helpful, and necessary to include. It added that the methodologies of Panelbase were rigorous and the full details of the poll in question were publicly online to any interested reader to explore them further, as demonstrated by the complainant’s own research. Notwithstanding this, the publication offered to amend the online article and to run the following correction in their Corrections and Clarification column, both online and in the Scottish print edition, where the article appeared:

Correction: “Our story "Sturgeon's surge tilts balance on independence" (News, July 5), which reported a poll finding that voters favoured independence by 54%-46%, should have made clear that the figures excluded undecided voters and those who indicated they were among the least likely to vote. When these groups are included, the poll found that 45% favoured independence, 39% did not, and 16% were undecided or unlikely to vote”.

8. The complainant said this proposed correction was insufficient to resolve his complaint as it did not address his concerns in full and needed to detail what particular filters were applied by the polling company in this instance – for example, on what grounds it designated a respondent as “unlikely to vote” - and how many respondents fell respectively within the category of “undecided” and of “unlikely to vote” and then excluded from the overall figures. Following this, the publication offered to add further context to the wording of the proposed correction to further address the exclusion of those unlikely to vote. The publication said its established Corrections and Clarification column was the most appropriate location for a correction to appear as its location (both online and in print) was familiar to its readership and, as such, was suitably prominent.

Relevant Clause 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

9. The article reported that the poll had found a reversal of the 2014 Scottish Independence referendum, with a “54%-46% lead for the nationalists”. These figures only reflected those who had answered “Yes” or “No” to the statement “Should Scotland be an independent country?”, excluding the significant proportion of respondents who were undecided and/or were deemed to be unlikely to vote. This exclusion was not referred to anywhere in the course of the article. Failing to make clear that a significant proportion of respondents (16 per cent) had been excluded in order to achieve this headline figure constituted a failure to take care over the accuracy of the article, raising a breach of Clause 1 (i). The reporting of the poll’s findings without making this clear, particularly when this interpretation formed the basis of the claim in the article that a “reversal” of the 2014 referendum had been recorded, was significantly misleading and as such required correction under Clause 1 (ii).

10. The Committee recognised the detailed and measured approach taken by both the complainant and publication to this matter. It noted that once the publication had been made aware of the complaint and previous IPSO rulings on the reporting of opinion polls, it had offered to amend the online article and publish a correction in its established Corrections and Clarifications columns both online and in print. The wording of this correction identified the error and made the correct position clear. This was offered promptly and with sufficient prominence to meet the terms of Clause 1 (ii).


11. The complaint was upheld under Clause 1 (i).

Remedial action required

12. The correction which was offered clearly put the correct position on record, and was offered promptly and with due prominence, and should now be published online and in print.

Date complaint received: 04/11/2020

Date complaint concluded by IPSO: 31/03/2021