1. 09:30 – 09:35 (5 Mins) – Introduction – Sir Bernard Silverman
  2. 09:35 – 09:45 (10 Mins) – Actions Update – Rob Turner
  3. 09:50 – 10:45 (55 Mins) – EAP163 – CCS Census Matching QA Strategy – Rachel Shipsey
  4. Break
  5. 11:05 – 11:50 (45 Mins) – Presentation – QA in Census – Catherine Bean
  6. 11:55 – 12:15 (20 Mins) – Census Update – Jon Wroth-Smith
  7. 12:15 – 12:30 (15 Mins) – Any other business – Sir Bernard Silverman

Panel Members

  • Sir Bernard Silverman (Chair)
  • Prof Ana Basiri
  • Dr Oliver Duke-Williams
  • Dr Nik Lomax
  • Prof David Martin
  • Prof Natalie Shlomo

Office for National Statistics

  • Rachel Skentelbery (ONS Lead)
  • Owen Abbott (ONS Deputy Lead)
  • Cal Ghee (ONS Attendee)
  • Gareth Powell (ONS Attendee)
  • Jon Wroth-Smith (ONS Attendee & Presenter)
  • Catherine Bean (Presenter)
  • David Edwards (Presenter)
  • Pete Large (Presenter)
  • Jo Neagus (Presenter)
  • Rachel Shipsey (Presenter)
  • Rob Turner (Secretariat)

1 –Introduction

  1. The chair welcomed Professor Ana Basiri to the Panel.

2 – Actions Update

  1. The panel closed Action 73 following the conclusion of field operations.
  2. The panel closed Actions 47 and 57 and reformulated these as Action 84.

Action 84 – ONS to bring paper to panel detailing how they are assured that software code is produced in line with best practice and how they will ensure processes for retention and archiving software and methodologies are in place

3 – EAP163 – CCS Census Matching QA Strategy

  1. ONS presented on the quality assurance strategy for matching of Census Coverage Survey to Census.
  2. The panel suggested that the thorough linkage process outlined in the paper might be useful in assessing over-coverage. ONS clarified that this already takes place, by using individuals matched in different locations as part of the estimation of over-coverage.
  3. The panel raised the possibility of using crowdsourcing in future clerical matching work, as this may be a way to be able to undertake large scale exercises such as linking administrative data sources, although they acknowledged this would not be possible for Census work.
  4. The panel suggested using the automatic matching scores as a way of assessing the quality of clerical matches, either as an instant prompt to matchers when a sufficiently unlikely match is assigned or to be looked it after matching has completed. The panel highlighted that a non-zero number of difficult matches would also be expected to be made, and so this could also be used in assessing quality of clerical checking. ONS stated that they do not have plans to do this, in part because they wish the clerical matchers to be independent of the matching algorithms but would look into this as part of Census evaluation work.
  5. The panel questioned whether there is a distinction for thresholds used when matching online and paper questionnaires to CCS records. ONS clarified that there is not.

4 –Presentation – QA in Census

  1. ONS presented on the validation of outputs of statistical methodologies, with the panel commending the overall approach.

5 – Census Update

  1. ONS provided an update on Census operations.

6 – Any other business

  1. There was no other business.