The Meeting was held at ORB on 20th February 2018

Present

Sir John Curtice, President
Nick Moon, Secretary/Treasurer
Simon Atkinson, Management Committee
Johnny Heald, ORB
Ivor Knox, Panelbase
Anthony Wells, YouGov
Gideon Skinner, Ipsos MORI
Laurence Stellings, Populus
Jansev Jemal, ComRes
Adam Drummond, Opinium
Luke Taylor, Kantar Public
Damian Lyons Lowe, Survation

1. Apologies for absence

BMG, Lucid Talk, and ICM

2. Minutes of the last AGM

The minutes were approved

3.Officers’ reports

• Financial report

  We currently have £9,985 in the bank, and £3,750 of outstanding membership fees (from 2017 and 2018), making a due total of £13,735. There are expenses expected, but probably no more than £500.

  NM raised the possibility of another fees holiday but it was agreed that the BPC needed to build a reserves fund in case the need for another enquiry, and fees would be levied as usual in 2019. With the existing members NM reported that annual subscription income is £2,750.

  It was agreed NM would send a note of thanks to MRS thanking them for their financial support for the Inquiry.

• Disclosure issues dealt with by the officers

  Nothing that fell within our remit. ComRes audience recruitment raised a new serious issue about whether this fell within the Rules. The Officers decided it didn’t, and a clarification was published on the BPC website making this clear

• Membership applications dealt with by the officers

  DCYFOR have become members

4. Appointment of the officers

• President
  NM nominated JC, SA seconded, elected nem con
• Secretary
  JH nominated NM, DLL seconded, elected nem con
• Management Committee member
  AW nominated SA, LS seconded, elected nem con

5. President’s report

The report raised the issue of whether we need to think about rules to cover Multilevel Regression and Post-stratification (MRP). It was agreed we keep a watching eye on this for now. All other main points from the report are covered in later agenda items.

6. House of Lords Committee

JC and SA attended on behalf of BPC, NM and various BPC members also gave evidence.

BPC might have to make some kind of holding statement quickly after the final statement is released. We have already said we’re opposed to both statutory regulation and banning polls. We have slightly moved into quality domain by having enquiries into elections and agreeing certain procedures as a result, so may be under pressure to venture further.

JC encouraged members who were not CEOs of their organisation to ensure that their CEO was fully aware of the potential difficulties following publication.

Cttee due to report by end of March.

Agreed that JC should be authorised to give a strong defence of the value of self-regulation – citing all the openness and attempts to improve that happened after the election. We are much more transparent than most other countries.

It was agreed that the BPC would hold at least one mid-Parliament meeting, ideally within the Palace of Westminster, as well as the pre and post-election sessions to explain latest thinking.

7. Reporting uncertainty in the polls

Gideon introduced paper produced by Luke Taylor and him. Aim was to try to find a simple approach of explaining likely levels of uncertainty in polls. They decided best way, to avoid having to explain statistical theory to lay people, was to use history of reliability of polls in previous elections. Looked at all final polls done by BPC members (including 2001 before BPC existed).

Suggested level of uncertainty of +/-4pp, (achieved by >90% of polls), but also mention that most polls are within +/-2pp.

The Sturgis bootstrapping approach, achieved by taking repeated random sub-samples from polls, was felt to be harder to understand. Also bootstrapping ignores bias rather than variance by not comparing with actual result
Members would be free to calculate and publish their own.

Agreed to principle of what the paper says.

We need a PR statement to accompany the change in rules.

8. Proposed rule change following item 7

Insert at the end of the bullet points at Rule 2.4.

In the case of a poll of voting intentions for an election, include in the published documentation the following statement: ‘All polls are subject to a wide range of potential sources of error. On the basis of the historical record of the polls at recent general elections, there is a 9 in 10 chance that the true value of a party’s support lies within 4 points of the estimates provided by this poll, and a 2 in 3 chance that they lie within 2 points.’

This amendment to take effect from 1 April 2018.

GS asked all members for comments, and to check their own companies’ figures.

JC proposed NM seconded nem con.

9. Any Other Business

SA suggested we look again at FAQs and journalists’ guide to opinion polls on the website in the light of the Lords Cttee

JC suggested we investigate cost of updating the design of the website.