Culture, Tourism, Europe and External Affairs Committee

Meeting date: Thursday, October 29, 2020

Official Report 635KB pdf

---

Subordinate Legislation

---

Census (Scotland) Amendment Order 2020 [Draft]

The next item on our agenda is evidence on an affirmative instrument. I welcome Fiona Hyslop, the Cabinet Secretary for Economy, Fair Work and Culture; Jamie MacQueen, a lawyer in the Scottish Government; and Pete Whitehouse, the director of statistical services at National Records of Scotland.

I remind everyone to allow broadcasting colleagues a few seconds to operate the microphones before asking or answering a question. Also, I would be grateful if questions and answers were kept as succinct as possible.

I invite the cabinet secretary to make a brief opening statement, after which we will move to questions. If any member wishes to ask a question, they should please type R in the chat box.

Good morning. It was back in February that I last appeared before the committee in respect of Scotland’s census. At that appearance, you recommended approval of the draft census order, and National Records of Scotland was making good progress and was on track to deliver a successful census in 2021.

I am here today to seek your approval of the amendment census order that will change the date of the census to March 2022. Let me make it clear that the decision to move the date of the census was not taken lightly. We are all aware of major events around the world that have been cancelled or significantly disrupted due to the impact of the coronavirus pandemic. Like many of those events, planning for Scotland’s census has been seriously impacted.

You will have heard me say before how important census data is. It is crucial that a high response rate is achieved, which, in turn, produces high-quality outputs for use by data users in the short, medium and long terms. It is for that reason that the decision was made to move the census to 2022. We must ensure that the census produces the high-quality data that is required by data users, and moving the census date provides the best chance of achieving that.

No other survey provides the range of information that the census provides. We get only one chance every decade to ask the people of Scotland to complete a census, which makes it all the more important that a full census is taken and that it reaches all communities across Scotland.

NRS officials have told you about the different options they considered to preserve the 2021 census date. None of those options provided the necessary confidence in relation to securing high response rates and achieving a successful census.

I appreciate that some data users will be disappointed at having to wait an additional year for data or will be concerned about Scotland’s census being out of sync with the rest of the United Kingdom. However, I can offer you and all data users and stakeholders the reassurance that NRS will continue to work with the other UK census offices and data users to ensure that Scotland’s census delivers the high-quality analyses and outputs that are required. The change of census date provides the best opportunity to do that.

I am grateful to the committee for your careful consideration of the census legislation thus far and for your support in ensuring that the census is a success. I formally invite the committee to recommend approval of the Census (Scotland) Amendment Order to the Scottish Parliament.

Thank you, cabinet secretary. As you know, representatives of National Records of Scotland came to the committee to talk about the delay, and, as you have just said, we all agree that it is important that we get accurate data from the census.

You said that NRS was on track to deliver the census before the pandemic struck. However, we know that, ahead of the pandemic, the delivery confidence assessment had moved to an amber grading. When the NRS representatives were before us, I raised the issue of the Office for Statistics Regulation report from October 2019 that looked at how all the census authorities across the UK were progressing, and I pointed out that, of course, it is only Scotland that has delayed the census.

That report suggested that NRS had experienced difficulties. For example, in paragraph 4.5 it said that NRS had told the regulator that it was facing challenges that included

“procurement issues, concerns over effective decision making ... and contingency planning arrangements.”

It went on to say that it was putting a new governance structure in place to deal with those things, but it said that there remained a delivery risk for census outputs in Scotland.

Did that report give you cause for concern, and do you believe that, notwithstanding the pandemic, NRS will overcome those issues and get into a position in which it can deliver the census in 2022?

Yes. That report was published back in 2019, and I have been keen to ensure that there is continuous programme board oversight and review of the progress. That is why we have the ratings—it was important that we had them previously, and it is important that we have them now.

We must separate out two issues. Clearly, the decision that was taken this year involved the completion rates and, therefore, the success rate of the census in 2021. In relation to the planning for the Scottish census, we were embarking on issues around procurement, particularly with regard to recruiting and training operators, and we were thinking about the provision that would be required in the summer and autumn in that regard.

NRS was moving from amber towards a more positive place—I will bring in Pete Whitehouse on that, because the governance and the changes in the oversight, particularly with regard to the operational side of the census, are obviously the responsibility of NRS. Nevertheless, as a Government minister, I spent a great deal of time on the matter and have had regular updates on progress, and I have pressed back on some of the issues that were highlighted in the report that you mention, which was from the previous year.

I appreciate what you are saying, convener, but we believed that NRS would be able to deliver, and we pressed back to test whether it really had to delay the census. However, like other countries, we have taken the view that, increasingly, the wisest and most sensible thing to do is to ensure that the participation rate is as high as possible. Given that we are spending £100 million of taxpayers’ money, a 60 per cent return rate would not represent good value and would not provide the data that we need.

Pete Whitehouse can talk about the amber rating that was in place in March.

As Paul Lowe said when NRS officials appeared before the committee a short while ago, the ranking of this complex programme as amber was not unusual. He spoke about how the Office for National Statistics was also in that space and said that that was the normal trajectory of a complex programme of this nature. The feeling in the programme board and within our organisation at that point was that we were going to be able to deliver a successful census in March 2021. However, that scenario changed when the pandemic hit.

Our conclusions are that we have made great progress. Some of the conversations that we have had with the Office for Statistics Regulation have taken the form of peer conversations with colleagues to talk through issues and get feedback on areas that we need to think about and on which we need to work. It is a very helpful and supportive improvement mechanism rather than something that should be seen as the marking of homework or something of that nature.

Our position was that we were ready to go; amber was the position that we were in, which, as Paul set out, meant that we were going to be able to deliver. As is absolutely the case with all censuses across the globe, the year running into the census is hugely complicated and difficult—we all know that—and it is usual for a census to be in an amber situation at that point. As I said, we continue to work with the OSR and it continues to give us good support and advice, as do other colleagues across the UK, to make sure that we overcome what will be the usual problems that we face. We are well set to do that, and we are confident that we will be able to deliver.

I have an observation: with hindsight, it is good that we are reducing the contact between enumerators and members of the public over the next year. That is increasingly looking like a good move.

My question is quite different and is directed to Pete Whitehouse. Given that there will now be an 11-year gap between censuses, which are a key statistical underpinning for much public and a good deal of private planning, how will we support data users in that extra year when the data is much more obsolete than it used to be? Will you do any additional sampling work to keep the data that was produced some 10 years ago relevant and reliable for users?

There a number of different types of data. We will continue to engage with all our users, so that they are fully aware of the work that we are taking forward and that we understand their need for data and how we can help them to, as you imply, bridge the gap between when they were expecting to get census data and when they will get that data for Scotland.

In areas such as population estimates, we will continue to use the methodology that we have and will roll that forward for a further year, so that will continue to provide the data. As Paul Lowe said in his letter back to the committee after our last attendance, we are working with the Office for National Statistics and other parts of the UK to ensure that we are consulting users of population data, migration data and other sources so that they are fully aware of how that data will come forward and how we will help them to use and have access to population data. That is one sort of data for which there is methodology in place that works at the moment, which we will roll forward for another year. Nevertheless, our position is that, as the cabinet secretary set out, providing the high-quality data of the census is paramount for us, which is why we made the recommendation to move it back a year.

On the other sets of data, the other surveys that are in place do not do what the census can do. The real value of the census is its ability to go to such a low level on different population groups and different geographies, but there are other sources of survey data out there, such as household surveys, health surveys and others, that will continue to provide valuable information to people as they wait for the census data to come on track. There will, however, be certain pieces of data that only the census can provide, and our recommendation was based on our delivering data of the quality that is required by users. The one-year delay will allow us to do that.

A couple of other members want to ask questions.

09:15  

Good morning. I have two questions; the first is addressed to Mr Whitehouse and the second is for the cabinet secretary.

Mr Whitehouse, further to its evidence session with NRS in September, the committee wrote to Paul Lowe, the registrar general, on 21 September, raising a number of points, and it received a reply. My copy is not dated, so I am not sure when that reply was written, but it states:

“As part of the re-planning exercise we are updating the full lifecycle resourcing plan and supplier costs to the Programme to inform the overall financial impact assessment and subsequent costings.”

Can you advise us whether that process has been completed? If it has not, when will it be completed?

We are still working through our replanning work. As Paul Lowe said in the set-out, we will write to the committee when we have concluded it. At the moment, we think that that will happen in December.

Thank you for that.

Cabinet secretary, in the convener’s letter to the registrar general, she asked that the committee receive

“regular ... updates on the delivery of the 2022 Census including details of Delivery Confidence Assessments provided to the Census Programme Board and of any concerns raised regarding deliverability.”

Can the cabinet secretary confirm that the committee will receive—[Inaudible.]—updates? That issue was not really addressed by the registrar general in his reply to the convener.

The sound was breaking up a bit, but I think I got that. As has always been the case, as we move into the final year, we will want to ensure that there are more regular updates for the committee on the progress of the census, and I am committed to doing that.

In relation to the census programme board reports, I will look at the level of the provision and the detail of that, but I have always been certain that I want to engage the committee on a regular basis through updates on operation and costs, deliverability and any risk aspects. It is my responsibility, if there are any risks to the census, to bring that to the attention of the committee. That is my commitment to you.

I thank the cabinet secretary for that. We look forward to those updates, because I imagine that the cabinet secretary has to rely on the assurance of the registrar general and NRS that the next deadline of March 2022 will be met, and it would be a pity if the cabinet secretary’s trust was not met in that regard. The committee is very happy to do what we can to ensure that the next deadline is met.

I can reassure the committee on that. The point of the broader oversight provided by the programme board is to give that range of test and challenge. Also, as the committee has heard, within the census community across the UK and, I dare say, internationally, there is understanding of the pressures on the census, not least because we are not the only country that is experiencing those issues. That test and challenge is really valued, and it is important to me.

I will be providing regular updates to the committee, and, obviously, my duty and responsibility is to make sure that the Cabinet more widely understands the issues. That is why, particularly in relation to the decision about the order that is front of you, I have made sure that my Cabinet colleagues are fully aware of and understand the issues. I engage with them actively on that.

Thank you.

Good morning, cabinet secretary. I find the first paragraph of the letter that has just been received from Mr Lowe quite an interesting curiosity. It says:

“An important indicator is the overall response rate which in 2011 was 94 per cent”.

It goes on to say that

“about 340,000 were thought to be missed and subsequently adjusted for during data processing.”

It seems strange to me that we are conducting a census to get 100 per cent accuracy, yet there is an estimate of people who have clearly not responded to the census.

The same paragraph goes on to say that

“Coverage assessment and adjustment (CAA) methodology was able to derive”

census population estimates using the census coverage survey. It then says that the response rate was 89 per cent in Glasgow and 96 per cent in the Borders, with the response rate in other local authority areas falling somewhere in between. How critical is the census if that information can be delivered more or less without it—so it seems—and how confident are you that it is accurate? It is suggested that the census represents 100 per cent of the population, but that is without getting a fully accurate census.

The census is a statistical tool. Its purpose and intention is to provide an account of all the people in Scotland on the census date. Like all censuses across the globe, it attempts to get as much information as it can from across the population. However, it is never going to represent 100 per cent of the population. The census does not work like that in places like America, Australia, New Zealand or Scotland.

I know that it cannot represent 100 per cent of the population, but we estimate that 340,000 people were missed out of the last census. That seems counterintuitive. That, as well as the fact that you can get variances between 89 per cent and 96 per cent, is what I am trying to get clarification on.

If it would help, I can put a paper together to explain how we do a range of other things.

We run the census and try to maximise engagement with it, so we have strong communication and engagement strategies running up to and throughout the census. As was mentioned in Paul Lowe’s letter, we also conduct a census coverage survey, which is a statistical tool that is used across the other UK censuses to understand whether, for example, certain households in certain areas did not respond. That gives us an opportunity to produce estimates. We present our figures, and we do that with confidence intervals and information that tells the user how precise the data is.

We know that response rates vary in different areas, and we do extra work in some of the areas where response rates have traditionally been lower, so that we are engaged with the community and provide support through whatever means is needed. That could be by helping people to access it, by providing translation services or by creating awareness of the census. A whole package of information and work allows us to produce the best-quality estimates.

In response to the member’s earlier point about the role of the census and its importance, the census allows us to go into fine-grain detail and gets information about different groups in our population, different work patterns and all sorts of things that surveys that might go to 6,000 households or 10,000 people cannot allow us to get. The census is absolutely critical in allowing that work to be done and in providing a message and story about our country. It also gives people an indication of the accuracy of that data, which will have been the case in every census that we have done.

It is the role of the Office for Statistics Regulation, in its assessment of our national statistics accreditation, to be clear about the quality of the estimates that we produce. The process in Scotland is no different from the process in England, Wales, Northern Ireland, America, Canada or Australia. It is the technique that is used to produce high-quality estimates for countries.

I am more than happy to provide a paper setting that out, if it would help.

That would be helpful.

I will remark on the visit that I made to the operations of the test census, when it was targeting the areas that we were previously concerned about. That is one of the reasons why particular areas of Glasgow were targeted during the test. Engagement and publicity are important, particularly in areas where, in the past, there has been a lower level of completion. Within that test area, there were many areas—for example, in relation to the Roma community in Glasgow—where we had to make sure that people were aware that the census was coming and that they were encouraged to complete it. There was also recognition of the disparity between urban and rural areas. Therefore, the test sessions were conducted not only to check that the systems work but to improve penetration by ensuring that the communication exercise maximises completion of the census. For example, we know that poverty has a big impact in lots of walks of life, and the census is no different. How do we ensure that we get the most accurate information? That is another layer with regard to making sure that we maximise penetration and get the completion rate up.

Okay. Thank you.

Thank you. I welcome the assurance that the committee will receive regular progress updates on the delivery of the 2022 census, because we are all agreed on its importance.

The next item of business is consideration of motion S5M-22767, in the name of the cabinet secretary, on the approval of the Census (Scotland) Amendment Order 2020. As no member has indicated that they want to make a contribution, I invite the cabinet secretary to speak to and move the motion.

I thank the committee for the time, attention and focus that it has given to the census. I appreciate the desire and need to keep abreast of developments and to scrutinise the progress that is being made. I reiterate my commitment to the committee on that.

Motion moved,

That the Culture, Tourism, Europe and External Affairs Committee recommends that the Census (Scotland) Amendment Order 2020 [draft] be approved.—[Fiona Hyslop]

Motion agreed to.

The committee will shortly report to the Parliament on the instrument. Are members content to delegate to the deputy convener and me responsibility for signing off the final report?

Members indicated agreement.

I thank the cabinet secretary and her officials for their attendance this morning.

09:29 Meeting suspended.  

09:31 On resuming—