Agenda

TimeItemPresenter
1
13:00 – 13:05
(5 minutes)
Introductions, actions reviewRachel Skentelbery
2
13:05 – 14:05
(60 minutes)
Statistical Design of Future Population and Social Statistics System
(EAP132)
Jennet Woolford
14:05 – 14:20
(15 minutes)
Break
3
14:20 – 15:20
(60 minutes)
Admin based population estimates for census Quality Assurance
(EAP133)
Owen Abbott
15:20 – 15:35
(15 minutes)
Break and NS joining
4
15:35 – 16:35
(60 minutes)
Census Scenarios
(EAP134 & EAP135)
Jon Wroth-Smith
5
16:35 – 16:50
(15 minutes)
Terms of Reference ReviewOwen Abbott
6
16:50 – 17:00
Any other BusinessRachel Skentelbery

Chair: Sir Bernard Silverman

Attendee List

External Panel Members
Sir Bernard Silverman (Chair)
Prof Natalie Shlomo (External Panel Member)
Prof David Martin (External Panel Member)
Dr Nik Lomax (External Panel Member)
Dr Oliver Duke-Williams (External Panel Member)

Office for National Statistics
Rachel Skentelbery (Vice-Chair, Chief Methodologist)
Sir Ian Diamond (National Statistician, Present for Item 4 only)
Cal Ghee (ONS Panel Member)
Owen Abbott (ONS Panel Member)
Sarah Henry (ONS Panel Member)
Jon Wroth-Smith (ONS Panel Member)
Jennet Woolford (Presenter)
Christopher Lydiat (Secretariat)
Gareth Powell (Secretariat)

Actions

Agenda ItemActionOwner
[6,4]A62: The panel requested an update on Census Scenarios in the autumn.Jon Wroth-Smith
[6,5]A63: ONS and the committee to amend the terms of reference document before signing offOwen Abbott

Minutes

There were no actions to update the panel on.

Background

The statistical design of the future population statistics system is a transformation of the social statistics system to provide statistics on the size and characteristics of individuals and households in a timely manner. Ways to adapt a set of linked datasets to produce an administrative based population estimate were described to the panel.

For the population ‘stock’, ways to estimate overcoverage were highlighted, including the possibility of dependent interviewing. The Integrated Population and Characteristics Survey (IPACS) was presented as the mechanism of handling undercoverage.

Methods to estimate population changes or ‘flows’ were also presented, with the intention to reconcile and integrate stocks and flows described.

Discussion & Suggestions

2.1 – The panel questioned the current status of IPACS. ONS said that there is a small scale pilot study underway. However, delays are likely due to survey teams responding to COVID-19 requirements.

2.2 – Information on the design of IPACS was also requested by the panel, with ONS stating a pilot study has not yet been completed, so refinements to IPACS design have not yet started.

2.3 – The panel asked why the census was not used as a data source for the demographic index. ONS stated their desire to use the census as a test of the index’s accuracy and discussed the desire for census data to be linked to the index to a certain extent afterwards.

2.4 – The coverage of communal establishments was discussed with the panel, with their inclusion intended but the exact scope currently being established.

2.5 – The necessity for additional research to address the impact on statistical geographies was highlighted by the panel as an additional consideration.

2.6 – Internal migration was also mentioned as a unique challenge, with the panel asking how it is intended to be handled as a flow. ONS intend to establish a coherent stocks and flows method, with internal migration addressed by the integrating of stocks and flows.

2.7 – The panel raised the possibility of learning from other national statistical institutions (NSIs) with population registers, to gain experience and insights into how to manage the UK’s lack of one.

2.8 – The work and intentions of other NSIs was mentioned, with ONS intending to continue current relationships.

2.9 – The relationship with Scotland and Northern Ireland was also raised by the panel, with ONS stating that we are working closely with them.

No actions given.

Background

The panel were taken through a range of options for handling a reduced IPACS survey in 2021, with the paper concluding use of the CCS as a viable alternative.

Estimates of variance using this approach under different scenarios were presented along with considerations of other sources of bias, with further work required in this area.

Discussion & Suggestions

3.1 – The large variance was discussed, along with its implications on using this for the census quality assurance.

3.2 – The panel asked why other methods were not selected for use. ONS explained that false positive rates were not known but estimated, and this can cause large biases in other methods.

3.3 – The approach was accepted by the panel as suitable for use as a comparator.

3.4 – Increasing the resource for clerical matching was also raised, with ONS stating the increases to resources raises risks to the census matching with uncertain potential benefits.

No actions given.

Background

Potential disruption to the census by COVID-19 was discussed, with 3 scenarios described: Disruption for 3 months, disruption until winter/spring and ongoing disruption in 2021, which would cause disruption to the field stage of the census.

Quality was expected to be impacted in all scenarios, with more severe impacts in the third scenario. Additionally, a proposed method to handle field disruption was detailed, alongside administrative data requirements.

Discussion & Suggestions

4.1 – The panel commented that there were issues that cannot be mitigated by ONS, such as if there are significantly fewer students in one year, or students not moving close to universities.

4.2 – The expected uptake in online response was questioned. It was suggested that ONS use data on current government online services to inform this prediction.

4.3 – The likelihoods of the scenarios were discussed. ONS assured the panel that the third scenario was being considered and emphasising the need to continue work around it, even though likelihood might be low.

4.4 – The potential for use of utilities data as a data source for addresses was highlighted as promising by the panel.

4.5 – The issue of gathering data from other sources was again highlighted as an issue for administrative statistics.

4.6 – The potential for using telecoms data was discussed. The panel emphasised caution in the consideration of mobile phone data, as issues in its acquisition, data accuracy and data coverage are present among other issues.

4.7 – The potential for delaying the census was also raised by the panel. ONS stated that alongside quality issues, a delay would lead to large fiscal costs and this would need to be borne in mind alongside other factors.

Actions

A62: The panel requested an update on Census Scenarios in the autumn.

5.1 – The panel were happy to approve the amendments to the terms of reference, on the condition that the number of days the panel are expected to work are well defined.

A63: ONS and the committee to amend the terms of reference document via correspondence and sign off.

6.1 – Adaptions made to assist in home-working were praised by the panel.

6.2 – The panel asked if issues involving the hashing of data from other government departments among other security barriers of data sharing should be costed to allow proper cost-benefit comparisons.

6.3 – Regarding data sharing from other government departments, the committee believed that as a principle, ‘data in the clear’ is a requirement for an administrative based census.