Members present

• Jonathan Portes (Chair)
• Harvey Daniell (Bank of England, Delegate for Huw Pill)
• Tom Pybus (HM Treasury, Delegate for Daniel Gallagher)
• Nye Cominetti (The Resolution Foundation)
• Steve Ellerd-Elliott (Department for Work and Pensions)
• Anthony Wilson (Department for Work and Pensions)
• Kirsty Yule (Scottish Government, Delegate for Richard Murray)
• Stephanie Howarth (Welsh Government)
• Alexandra Fitzpatrick (Welsh Government)
• David Bell (University of Stirling)
• Alan Manning (London School of Economics)
• Xiaowei Xu (Institute for Fiscal Studies)
• Philip Wales (Northern Ireland Statistics and Research Agency)
• Tim Butcher (Low Pay Commission)

Labour Market Technical Group members

• Alice Crundwell (Bank of England)
• Michal Stelmach (Bank of England)
• Douglas Rendle (Bank of England)
• Paul Cheston (HM Treasury)
• Elizabeth Hunt (HM Treasury)
• Anisha Rasan (HM Treasury)
• Sunita Bali (Office for Budget Responsibility)
• George Koutsoumanis (Office for Budget Responsibility)
• Andrew Ward (Department for Work and Pensions)

External Methodology Experts

• James Brown (External Methodology Expert)
• Ray Chambers (External Methodology Expert)

ONS presenters

• David Freeman
• Tim Gibbs
• Leone Wardman
• James Harris

Secretariat

• Niamh Davies

ONS colleagues

• Owen Abbott
• Sarah Ash
• Tom Evans
• Mel Gore
• Liz McKeown
• Sarah Henry
• Mike Keoghan
• Sumit Dey-Chowdhury

1. Introduction

1. The Chair opened the meeting and thanked panel members for their attendance.
2. The Chair welcomed the increased number of attendees for this session including members of the separate LFS/ TLFS Technical Stakeholder Group, ONS colleagues with responsibility for survey design, collection and methodology, and academic technical advisors James Brown and Ray Chambers.

2. Overview and Update – David Freeman

1. David Freeman (DF) provided an overview of recent activities since the last panel meeting, including the reweighting of the Labour Force Survey (LFS) using updated population projections, the higher LFS achieved sample size, and the publications released in early December.
2. DF then introduced the bias work, emphasising the importance of understanding and addressing bias in the LFS. DF highlighted the initial work carried out by Tim Gibbs and his team and invited feedback on further analysis and potential solutions.

3. Introduction to analysis – Tim Gibbs

1. Tim Gibbs (TG) introduced the analysis to the panel. He explained it is focused on non-response bias only and it includes all LFS waves with respondents aged 16+ and GB only. Key variables analysed included age, sex, region, tenure, index of multiple deprivation, disability, country of birth and country of birth by ethnicity (LFS).
2. The presentation included a discussion of the challenges in the data and the implications of weighting for bias. There was consensus that the characteristics of non-responders need to be understood more, particularly young people and non-white populations.
3. Feedback from panel members indicated that there is an appetite for more detailed weighting schemas, with the suggestions of using admin data to support this. The ONS confirmed this is currently being explored.

4. Future work– Leone Wardman

1. Leone Wardman (LW) presented the panel with the ongoing and proposed future work that her team at the ONS are undertaking to investigate bias. These include exploring admin data, using census non-response studies data and looking into adaptive designs for sampling.
2. The panel discussed the future work with strong support for the use of admin data sources such as PAYE for triangulation purposes. The panel gave suggestions for further future work including exploring whether the household composition of achieved interviews was contributing to potential bias. LW confirmed this has been considered as planned exploratory work. Panel members also briefly explored the possibility of using mandation in the future, to address these issues in the longer term.
3. Academic technical advisor Ray Chambers noted that the lessons being learnt on bias for the Labour Force Survey (LFS) must be transferred to the Transformed Labour Force Survey (TLFS), this is essential to future proof the work.

5  Questions – James Harris

1.  James Harris (JH) presented the panel with several questions following the bias analysis they had seen. JH asked panel members to reflect on what conclusions they have drawn from the analysis, what further analysis would be beneficial, and any methodological actions that could be taken to address bias.
2. JH welcomed further written discussion and feedback from panel members after the session, with recognition that they had been presented with a lot of data.
3. The discussion from the panel built on suggestions made earlier in the meeting to conduct further analysis into youth responders versus non-responders, the utilisation of administrative data sets including from the Department for Work and Pensions and consider expanded reweighting work including migration status.
4. ONS will reflect on the feedback received both during and after the meeting, including on relative priorities for further analysis, in determining next steps.

6. Conclusion

1. The Chair thanked participants for their valuable contributions and welcomed any further written feedback, comments and discussion offline after the session.
2. The next meeting is scheduled for the 29^th January 2025, focusing on the TLFS design.

Please note: these minutes were produced with assistance of AI