Women and Equalities Committee
Oral evidence: Preparations for the 2021 Census, HC 1909.
Wednesday 6 February 2019
Ordered by the House of Commons to be published on 6 February 2019.
Members present: Mrs Maria Miller (Chair); Tonia Antoniazzi; Sarah Champion; Angela Crawley; Vicky Ford; Eddie Hughes.
Questions 1–73
Witnesses
Iain Bell, Deputy National Statistician for Population and Public Policy, Office for National Statistics; Frankie Kay, Director of Census Transformation, Office for National Statistics.
Witnesses: Iain Bell and Frankie Kay.
Q1 Chair: Thank you so much for joining us today, and to those in the gallery and anybody who is watching online. This session is a one-off evidence session to explore what you are doing, in terms of the work you are doing in the Office for National Statistics regarding the next census. The census, obviously, is taken every 10 years and is the most robust information that we have, in terms of making policy decisions. Collecting the information is an enormous task and we thank you for all the work that you do there. We are really interested in the equality aspects of what you are doing.
The Government have published their White Paper on the census, and we are going to be talking to you about some of our thoughts around that and just probing a little bit more on equality issues before you finalise what you are going to be doing next. We are going to follow the usual format. If you could just say your name and the job you have, then we will have colleagues ask a series of questions.
Frankie Kay: I am Frankie Kay. I am the director of the census and data collection programme at the ONS.
Iain Bell: I am Iain Bell. I am deputy national statistician for population and public policy. I am the senior responsible officer for delivery of the census.
Q2 Eddie Hughes: You have recommended that a number of questions be added to and removed from the 2021 census. What, if any, demographic data will not be captured by this census, and what other data sources will be capturing that information instead?
Iain Bell: As you say, we are recommending three questions be added. Those were the ones on sexual orientation, gender identity and the veterans of the UK armed forces.
One that was recommended for removal was the “number of rooms” question, where we have alternative data available from the Valuation Office Agency. Evidence from the last census that people found it hard to fill in that question accurately. The second one recommended for removal was year last worked, where it turned out the actual user need was for whether somebody worked in the last year or not, and people, equally, found it hard to complete that form, particularly if you imagine you retired 25 or 30 years ago. A number of people found, “Actually, I do not know”, and we did not have any need for that particular bit of information.
That means, if you look across the nine protected characteristics, the only one that is not collected through the census is maternity, where we need to look at alternative data sources, through the NHS, in order to collect such information.
Q3 Eddie Hughes: Sorry, what does that aspect mean: “We need to look at alternative data sources”? Damn right you do, and what and when?
Iain Bell: As part of the Digital Economy Act, the ONS was given the ability to get data in from across Government, in order to improve the quality of our statistics. We are currently working with NHS Digital to bring in health service data, in order to fill that gap.
Q4 Eddie Hughes: You will be filling the gap prior to the 2021 census anyway. Are there any other gaps that are not captured by other means?
Iain Bell: There were no topics where we did not have an alternative source. As we went through it, there were several topic areas that we rejected earlier on in the process, but for each one there were alternatives available on these. An example of that would be volunteering, where we have the information available. We can collect the information through our surveys, because that meets the user need.
The other big one is there has long been a demand for income data to be available through the census, but many people find it too sensitive to ask that question or to answer that question through the census. We are going to utilise the administrative data from DWP and HMRC as the alternative for that. We will link it into the census data to provide income breakdowns by all the sociodemographic characteristics. At this point in time there was nothing we had in the topic reports and consultation that could not be filled by alternative means over the course of time.
Q5 Chair: Just to clarify, that data in areas such as volunteering will be linked to the census data. Therefore, it can be cross-tabulated by wizard people who can do those sorts of things. Although it will not be collected by you, it will be compatible with the data you collect.
Iain Bell: That is correct, yes.
Q6 Chair: Secondly, in terms of maternity and pregnancy data, which we found woeful, in terms of its availability, how confident are you that the NHS will actually provide you with robust data that is accurate?
Iain Bell: Until we actually get the data and work with it, we will not be able to give a full description of the quality and accuracy.
Q7 Chair: How can you make the decision now not to include pregnancy and maternity in your datasets if you do not know whether the NHS actually collects the data correctly, in line with the sort of robustness we need?
Iain Bell: The process for how we decide what goes into the census is based around user requirements. We launched a consultation in 2015, which set out to identify the areas of highest priority from users for collecting the data. Alongside that, we set out a number of criteria; they are around user criteria, operational ability and public acceptability and legal requirements. I will need to check thoroughly, and will confirm to the Committee in writing afterwards, but my understanding was that, in the 2015-16 consultation, the subject of maternity data did not come up from users as being an area of high demand.
This may surprise you, as the Committee. This is what came through the consultation process, and so, against our criteria and the way we operate, it did not come through. What I am saying to the Committee is, despite that, we recognise the gap, following the work we have done on our inequalities audit, and are looking at how we fill the gap anyway.
Q8 Chair: When you write to us, will you be able to confirm that you have done a quality assurance check on the NHS data, or that, if you have not, you will be?
Iain Bell: I will, yes.
Q9 Eddie Hughes: On the issue of the Roma ethnicity classification, how do you intend to ensure that people are aware of which tick box to use, and that the headcount is accurate?
Iain Bell: First, we know that we have already begun engagement with these groups. In the research and stakeholder engagement in order to determine the tick boxes, we have already formed links with national organisations representing the Roma community and with local organisations. We have already made links in Sheffield and elsewhere.
The approach is under development, but it will be very similar to what we did in 2011 with Gypsy or Irish Travellers. We know local authorities are good for helping to tell us where the communities are, but both we and local authorities do suffer a little, in that we have association with Government and not all these groups fully trust Government. The approach we use is to employ community outreach workers and community outreach census teams, who then go in and work with the communities in order to get the completion in.
This means we end up doing things like attending the largest Gypsy and Irish Traveller fair in the country, encouraging census completion there. In the Somali community, we ended up with our outreach workers arranging completion events in cafes in the communities. Our approach is really to utilise the people who can build the trust of that community.
For Gypsy and Irish Travellers, it was their community organisations reassuring them that we had no interest in their number of cars in order to shop them for the car scrappage scheme, which was their worry at that point in time. If their federations, their organisations and their community leaders are reassuring them, that helps encourage completion. We are currently developing the approach for the Roma community, but it will look very similar, in working with these groups in order to do it.
Q10 Chair: It is fair to say that the last census, in terms of its ability to be able to estimate the numbers of Gypsies and Travellers, was felt to be a vast underestimation. Are you really right to be using similar tactics this time round? How are you going to make sure it is not just another huge underestimation of numbers in the Roma community this time?
Iain Bell: It is fair to say we know that the undercount in Roma was even more severe. Last time round, without the tick box and without any of the engagement with the Roma community, only 737 people wrote in “Roma” under that option. The best estimates we have for the population is they are well in excess of 100,000, so we know that without this form of engagement the undercount will be even greater. We are always happy to work with communities to ensure we can get this form of engagement further still.
Q11 Chair: The point I was making was that you did try engaging in this way with Gypsies and Travellers last time, in the way you have just said you will with Romas, but we still saw a massive underestimation. Is there not a new approach that you need to adopt?
Iain Bell: We will need to look and work with these community organisations and see if there are any additional ways of working. We are currently at the point where we will be working through the dress rehearsal in 2019, which will help us learn more. If new suggestions come out that work better and work more, we will be happy to take them on.
Frankie Kay: We are also looking to ensure that we have very specific procedures for the field officers, so that they are very clear about the best way to engage with those communities and elicit the responses. We are also looking to see, in particular communities where we believe there are a lot of Travellers or Gypsies, whether we can recruit specific officers from those communities to work with them, in terms of trying to improve the response rate and the engagement.
Q12 Chair: Is that different to the way you did it last time?
Iain Bell: The main difference will be that last time many local authorities felt our engagement with them on the location of sites where these communities were was not as strong as it could be. The main difference in these procedures will be strengthening that working, so we get the best information out of local authorities to reach the communities.
Frankie Kay: The most important thing is about that upfront engagement, so exactly that. It has been working with the local authorities so we are much clearer where those communities are, so we can make sure we focus our effort in those specific areas to get the response rate.
Q13 Eddie Hughes: A number of other ethnicity classifications were considered but rejected. I would be interested to talk about the process behind that. Specifically with regard to the Sikh community, in the ONS equalities data audit final report, which was published in October 2018, I believe it states, “Ethnicity is the primary source of data collection in the UK and is therefore used in monitoring equality”. With regard to what is a very significant Sikh population, having proposed not to include the Sikh tick box in the census, and based on your very considerable experience in this sector, how can you give assurance to the Sikh population, to the taxpayers in this country, that the 40,000 public bodies that use census ethnic categories to collect data are going to have sufficient data to ensure that the legal requirement is met? That is given the fact that it is 35 years since Sikh was identified as an ethnic group.
Iain Bell: On the Sikh area, what our proposals do is fully recognise the use in how this system works. Our decision-making criteria were based around the ones I outlined earlier to the Chair. Within that, one of the cores was public acceptability. In the community, there was not a single, unanimous view in favour of inclusion. It has been quite a strong issue within the community both ways.
That being said, we fully recognise the needs of the Sikh community. There are three ways in which we think our data-gathering processes can be significantly strengthened. The first is a search-as-you-type facility online. The second is utilising the Digital Economy Act again. The third is strengthening the guidance across public authorities.
Q14 Eddie Hughes: Did you say there was not a significant consensus within the Sikh community?
Iain Bell: Yes.
Q15 Eddie Hughes: The information I have seen seems to strongly contradict that. Is there some ambiguity in the data, or are you just absolutely sure?
Iain Bell: The evidence we got came from unacceptability. We undertook qualitative research, which is the standard approach for determining new ethnicity classifications and their acceptability. It showed that there was a group of older males in the Sikh community who were strongly in favour of the tick box, but that the younger second and third-generation Sikh population were very much not in favour of the tick box. They felt it was very unusual to have one religion in there, and they felt they would identify their Sikh-ness through the religion question rather than through there.
Q16 Chair: It is a generational issue.
Iain Bell: Yes. The survey of gurdwaras, which I think you may be referring to, was carried out by the APPG for Sikhs. We passed it to an independent methodological expert, because it was issued without consultation with us on methodology. Our advice was that it was leading in how the questionnaire was phrased, so that it was prompting a positive response, and that it was unclear what constituent base it was representing.
Q17 Eddie Hughes: It was not sent to you before it was issued.
Iain Bell: No.
Eddie Hughes: Okay. I thought it was.
Iain Bell: Certainly I never saw or commented on it before it was issued. Frankie, do you want to pick up the bits around the online facilities and the digital economy?
Frankie Kay: I might just touch first on the harmonisation guidelines as well. It is very important to consider religion, ethnicity and national identity in the round, in order to make sure we really understand the characteristics of the population. Therefore, we want to make sure, when public bodies are considering how they make sure they do their equality monitoring, they think about those characteristics in the round, because there are different elements of that.
To back that up, we are looking at trying to make sure within the census we will have the search as you type. That means that when you start typing in “Sikh”—S-I K—then “Sikh” will come up as a prompted response within the ethnicity area. Within the Digital Economy Act, when we are looking to take data within the census, we will look at both the ethnicity data and the religious data, in terms of making sure we consider the Sikh community in the round. We will then link that to other sources, such as health, education and so forth, so that we can be really clear that we can understand the needs of those particular communities.
Iain Bell: Just to really emphasise the guidance on the harmonisation, the key point is probably that what is going on in public service monitoring is that our guidance is clear that we encourage people to look at nationality, religion and ethnicity in the round. However, many public bodies are only picking up the ethnicity. Therefore, simply moving a categorisation from a religion question into it does not tackle the wider problem, which is that actually we should have monitoring of all religions across all public services.
The Digital Economy Act helps that. We will be strengthening our guidance in two ways. One is really reinforcing the need to monitor religion. Secondly, we will make it very clear that, for those who do not take forward the religion question—the point you have made about the recognition in law 35 years ago—public sector bodies will have to undertake the due diligence themselves, to make sure they are fully compliant with their duties under that law.
Q18 Eddie Hughes: Can I just move on very briefly to a more specific geographical question? The Member for St Austell and Newquay—and other Cornish MPs, I am sure—will have perhaps lobbied you on behalf of their residents, given the fact that the Council of Europe has identified Cornish as an ethnicity. What consideration was given to that, in terms of tick boxes?
Frankie Kay: We are aware of the conversation with the Cornish around their national identity. When we did the research that Iain has laid out before on the topic consultation, it was very much around user need. The evidence and the research we did found that the need was quite localised within the Cornish area. For example, we received evidence from the Cornwall Council and from the Cornish MPs, but we did not receive any other requests across any other Government body, in terms of specifically around the Cornish. Again, we are absolutely aware of the need to make sure that they feel they are able to identify as they wish within the census. Therefore, we will have the search-as-you-type facility within the national identity question, so they can add Cornish within that.
Although the need is quite localised, we are also aware that there are people from Cornwall that live across England and Wales. Therefore, the engagement and the marketing we need to do needs to make sure people are aware outside of the Cornish area that they can put Cornish as a search-as-you-type option, so the marketing and communications in this particular area is not just localised within the Cornish area.
Q19 Chair: Can I just ask, for clarity, how you then capture that data if it is self-inputted? Somebody might put another word—I do not know if there is another word other than “Cornish”—and therefore you might end up with a lot of free data that you cannot then use.
Frankie Kay: We have some very clever people. Effectively, we look at the responses as they are coming in. We categorise them, and, as we see different phrases coming in, we will add to those categories, and we link the different variations for how people might spell “Cornish” if, for example, there is a typo or something like that. We do that by looking at the data and making sure we group everything to do with Cornish into a similar category. Also on this occasion, we are going to be producing bespoke tables and analytical reports for the Cornish community. Again under the Digital Economy Act, we can also link that to other data, in terms of education and health. Again, we can make sure we can meet the needs of that community.
Q20 Chair: It is not specified as a tick box but you collect the data.
Frankie Kay: Yes, absolutely.
Chair: I think that is really good feedback for colleagues. That is very helpful.
Q21 Vicky Ford: My questions are on the issue of gender identity. In the White Paper, it suggested you needed to do more testing in order to work out what the question should be. Can you explain to us what sort of testing that is, how that works, and whether or not you have now decided on the questions?
Iain Bell: It is a mixture of what we call, in our world of statistics, qualitative and quantitative research. Qualitative is very much in—
Q22 Vicky Ford: I am a statistician.
Iain Bell: My apologies. Qualitative is just small groups of people really getting in depth and checking which words and which phrases get the right data. Quantitative undertakes the larger, wider-scale testing to make sure it is accurate and representative and everybody is able to do it. We have done both on gender identity to a large scale. We were looking at the phraseology of the question through the research, getting that wording absolutely right. We now have the wording in a near final—
Q23 Vicky Ford: What sort of a group do you test that on? What sort of a number of people? How do you choose who you test that on?
Iain Bell: The qualitative testing is usually in a number of focus groups, usually each of about six, and usually they take you up to about 50 or 60 people. The wider-scale quantitative testing, if my memory serves me right, on the gender identity one is a sample of about 35,000, so it is quite a large-scale sample. That tests whether the wording works.
For gender identity, it had to work for both the community and the wider population. We learnt lessons from the Australian census. The question there, when they added it, provided some confusion for an elderly population. We actually spent a lot of time making sure it worked for both the LGBT community and the wider community.
Q24 Vicky Ford: Would you describe that population of 35,000 people that you tested as quite a large sample.
Iain Bell: Yes, I would.
Q25 Vicky Ford: That is helpful, because it is obviously quite a sensitive question, so it is important. Have you now decided on the questions?
Iain Bell: Yes, we have.
Q26 Vicky Ford: I am told that you have decided that the question on sexual identity should be mandatory, but the question on gender identity should be optional. What is the rationale there? Is that true, first of all?
Iain Bell: No. Sexual orientation and gender identity are both voluntary. The question on sex, i.e. the male/female question, remains mandatory.
Frankie Kay: There are three questions.
Iain Bell: The question on sex is the standard question that is the same as last time, which is, “What is your sex?” The answer is “female” or “male”. Then you have a question on sexual orientation, which is, “What do you consider your sexual orientation?” which is “heterosexual or straight”, “lesbian or gay male” or “bisexual or other”. Then there is a gender identity question, which will be, “Do you consider your gender to be different from that at birth?” The second two will be the voluntary questions.
Q27 Chair: How do you deal with people who are intersex?
Frankie Kay: Just to add to that, in the sex question we will also be putting a note there to say that we will also be asking another question about gender identity later on, so people are aware when they are answering the sex question that there will be a gender identity question coming as well.
In terms of the intersex, we have done some testing on that and we found that it was very difficult for people to clearly understand what that question was. The research found that was not a question we felt able to ask on the census. We are now looking at whether there are any alternative sources that can help meet that need. In the testing, most people found the question a little confusing.
Q28 Chair: Who found the question confusing?
Frankie Kay: The research that we were doing, when we were looking into that particular topic.
Q29 Chair: So the confusion caused amongst those who are not intersex has meant that people who are intersex cannot register that on the census.
Frankie Kay: There will be a write-in option if you want to identify as intersex. Sorry, I should have said that.
Chair: You are back into your free box.
Frankie Kay: Yes, apologies.
Iain Bell: Under the gender identity question, there will be a free box.
Q30 Chair: I am not talking about gender identity. I am talking about biological sex identity. Will there be a box on that?
Iain Bell: No, there will not be. There are two important things I would like to highlight to the Committee here. First, in discussions with the intersex community, they were happy with the question as it was. The second aspect is we have to design the census so that everybody wants to fill it in. As we completed the research on gender identity, we started off and we explored a three-box option under the sex question. That provoked a very strong reaction from certain groups, particularly in the feminist community, who were very against this being included. They felt this was not appropriate and it was males taking over the female sex question as well.
Our work had to come through and navigate a path that enabled the LGBT community to identify and feel included in the census, without having other groups feel as though they could not take part in the census because of it. Sorry, I can see by the looks that this is surprising to members, but that is genuinely what happened.
Q31 Chair: That feels slightly like a politicisation of the census—that you have made that decision because of the reactions of another group. There is a group, albeit a small group of people, who fall in this category.
Iain Bell: Our criteria are quite clear. Public acceptability is a clear part of the criteria. When you are not getting unanimity amongst the stakeholder groups, that is an important part of the public acceptability criteria. Therefore, it fits within our independent objectives, statistically.
Q32 Vicky Ford: When you did your testing, you tested on 35,000, and then a separate, specific testing within specific groups, like the LGBTI group. You tested that as a subsection of that 35,000, or in addition.
Iain Bell: That was separate.
Q33 Vicky Ford: Would non-binary gender identities, for example, be part of the write in?
Iain Bell: Yes.
Q34 Vicky Ford: Do you think that there is a risk that, because your first question is the sex question, that could put people off completing the rest of the survey, because that is such a specific question? That is the M/F tick box that is mandatory. Is there a risk that people will not complete the census or will not answer the questions on gender identity or sexual identity because of the way they are phrased?
Frankie Kay: The research we are doing is to try to ensure that does not happen, hence the note on the sex question, to make sure people feel they will be able to enter their gender identity later on. The research we did in the 2017 test on both the sexual orientation and the gender identity questions showed that there was only a very small percentage, with less than 2% in both of those questions, where people felt they would then stop answering the rest of the census as a result of those questions being included. That is well within the acceptability, in terms of margins of what we would expect. In fact, I think it was below 1%. We have done that sort of testing to make sure that those questions do not have an adverse effect on the broader response rates.
Q35 Vicky Ford: I am sure this is a question I should know the answer to already. At what age do people start filling in the survey? How do you manage, for example, teenagers and young people, who are often raising these issues?
Frankie Kay: The way the census is filled in is the head of the household will fill it on behalf of those that live within the household. Both for sexual orientation and for gender identity, we are saying that we would not expect those questions to be completed for anybody under the age of 16. There are two other aspects to this. One is that we will be making both the sexual orientation and the gender identity optional; therefore, nobody need answer those questions if they do not wish to.
In addition, if you are a member in that household and you would like to answer your own questionnaire, you can privately request your own questionnaire, either on paper or online. Therefore, if you want to make a private response, you will be able to do so, and then we will make sure we take that answer in consideration over the one that has been prepared by the household. It is very important to make sure the confidentiality is maintained.
Q36 Vicky Ford: Yes, but you are not going to be capturing data on under-16-year-olds in this area.
Frankie Kay: We would not be, no.
Iain Bell: Our research showed the public acceptability really dropped off for asking it for under-16s. Particularly because it is a household and people can fill it in on behalf of others, that public acceptability drop-off was quite marked at about the age of 15 or 16.
Q37 Vicky Ford: Moving on from gender, I have just one question more generally on the data audit. The ONS audit of protected characteristics found that the majority of data sets are missing data on a number of different protected characteristics. This is what we have been told. What advice would you give the Government about how to roll out the 2021 census questions across public data?
Iain Bell: We are already actively taking forward some of that work. The NHS is particularly far behind on its categorisations of ethnicity. I think it is on 2001. We are already in discussions with the NHS about how we bring its characteristics up to date, and how it will implement sexual orientation and gender identity alongside it. Following on from the Committee’s hearing on race disparity last year, jointly with the Race Disparity Audit we are taking forward a workshop with all Government Departments on how we can take forward the upping of ethnicity classification and data collection across them.
The final part of the White Paper, which I will go back to, is the importance of the Digital Economy Act here. At the moment, if you consider it, every public service is collecting these data every time. There are many repeat collections of these data across public services, which must come at considerable burden to British citizens and expense to the taxpayer. Part of the key proposals is, as we look at bringing these data in, we can look to identify where we have good data across the whole population on that protected characteristic, and then utilise the linking of the data to provide equality monitoring back across the other protected characteristics and help fill in that gap.
The three-pronged attack to this is, first, getting our harmonisation guidance up there; secondly, working and engaging with all the public services to get them engaged, as we are doing with the NHS and will be with other Departments soon; and, thirdly, really exploiting the power of the Digital Economy Act to fill the gaps without increasing the burden on both citizen and state.
Q38 Sarah Champion: The Government are going to need to pass legislation to implement your recommendations. Do you foresee any problems with that? Is there anything you are kept up at night worrying about in case it comes up?
Iain Bell: There will be legislation brought forward on sexual orientation and gender identity in primary legislation. There will then be the census order later this year, and then regulations early next year, which enable the census. At this point in time, I do not have any particular concerns around the legislation per se. We are working closely with Cabinet Office in terms of preparing the primary legislation and working through in discussion with communities, who may still be wanting a different position, in order to make sure things go through.
Being totally transparent, the honourable Gentleman here mentions the Sikh tick box. We are aware of a strong campaign on that side particularly, and on Cornish for that. We are working through with the Minister for the Constitution on those.
Q39 Sarah Champion: You would have seen my eyebrows shoot up on the sexual orientation—that you did not think there would be any issues around that. For those of us who work in that field and the work we do on this Committee, it is a very contentious issue.
Iain Bell: We went through the process of preparing a White Paper with Government. That goes through a standard write-around procedure and agrees the policy for the legislation being primary. The census elicits a large amount of parliamentary interest. Despite a lot of publicity around our proposals around both sexual orientation and gender identity, it has not elicited much parliamentary interest at all on that side yet.
Sarah Champion: Give it time.
Iain Bell: Yes, exactly. I am fully cognisant of how sensitive an issue it is, but at this point in time the registration there on that side has not come through. I am aware that Scotland has hit more issues on this than we have with their proposals. They have been largely relating to the gender identity question. In particular, they put the proposal of having the “other” box in the binary sex question. That elicited the same response in Parliament as what I mentioned earlier about the male/female and that lobby group.
That is the other part around, particularly, sexual orientation. Certainly our experience from the Scottish Parliament side is the sexual orientation question has not elicited as much controversy. The legislation is to make it a voluntary question, so nobody need complete it if they do not want to.
Q40 Chair: Iain, can I just ask, practically, if Parliament were to decide to amend the primary legislation on the Floor of the House—I hope it does not, for your sake—what would you then do? Would you have to go through your qualitative and quantitative assurance process? How would you square that circle, because Parliament is asking you to do something that you then may find does not actually work in practice?
Iain Bell: Ultimately, the census is brought together and enabled through the 1920 Census Act, which means that the decision of Parliament is the decision of Parliament. In terms of the legislation on sexual orientation and gender identity, the Bill will be to make the question voluntary.
Q41 Chair: Is that primary or secondary?
Iain Bell: Primary. I would need to check whether Parliament could amend the question during that process, given the Bill would be to make the question voluntary. It would not be about the question wording.
Q42 Chair: It would be really useful if you could write to us on that.
Iain Bell: I just want to make absolutely sure I get my legal and parliamentary process right.
Chair: Yes, we know how complicated it is. It would be very useful to know, just in case that situation arises. We would have your expert opinion to hand.
Q43 Sarah Champion: You have spoken about how you are going to get groups to engage and fill in the questionnaire. Looking specifically at people with protected characteristics, how are you making sure the form is accessible and that they actually go forward and complete it.
Frankie Kay: There are obviously a range of different things that we are doing. In particular, we are making sure we are working with groups such as the RNIB and the British Deaf Association—those kinds of interest groups—to make sure we are providing the best access to the census. We are still going to have the online census. We want it to be digital first. However, we are aware that not everybody would wish to respond online, so we are going to still provide people an offline response. We are going to be looking, in particular, if there are certain areas and populations where we feel it best to provide paper as the first way of completing the census, albeit with still an opportunity to fill online if they wish to do so.
We are also looking at our assisted digital offering. We are working and looking, in terms of how we can provide that service. At the moment, we are about to go out for a tender on that basis, to make sure that those that would like to be able to fill the census in online have the ability to do that. We will also have translation services available, in terms of translating our guidance into a number of different languages. We will also have our contact centre, so that if people would like to complete the census they will able to phone our contact centre and get assistance that way.
We will also be having a wide range of marketing communications to outreach to those communities, very much engaging with our local communities, working with local authorities, interest groups and charities, to make sure we have as wide a reach as we possibly can. We employ very specific community liaison officers to try to make sure that is as effective as possible. We also look to make sure our field officers are as representative as they can be of the communities in which they are working. We look for people to speak other languages. For example, in Wales we look to make sure we have Welsh speakers in our field offices as well. We try to make sure we have a full range of different ways of making sure people can complete the census.
Q44 Sarah Champion: Could I drill down on a specific group, the deaf community? I do a lot of work with my deaf community. I have to say, before I worked with them, I assumed if you wrote it down they would be able to complete it, but of course English is their second language and British Sign Language is their first. For that community, how will you get them to complete it? Will you actually have people that they could go to?
Frankie Kay: Unless you are aware, I think we would have to specifically get back to you on that one.
Iain Bell: The bits I know of are that certainly online there will be sign language interpretative help available to the community on how to do it. We will then work with the British Deaf Association and other groups about how best to outreach into those communities. These are communities that we develop the engagement with once we get our field force up and working, which we are at the stage of doing.
Q45 Sarah Champion: What have you done in the past?
Iain Bell: In the past, it was very much what we have talked about before. In 2011, it was working with the British Deaf Association and making sure there were sign language interpreters available to help out, and community outreach. Often people come together in community groups for fellow deaf people. Once we work with the British Deaf Association and others, what we tend to find is they know where the groups come together. We can then arrange completion events in those groups, and that helps them to overcome those barriers.
Q46 Sarah Champion: What is your budget for all this outreach stuff?
Iain Bell: I do not have the breakdown for the entire outreach to hand. I can provide that to the Committee later.
Q47 Sarah Champion: You were talking about the online component and the offline, paper component. I am going to read this, because it is complicated. You are hoping that the census is predominantly completed online, but there is still a component of the form on the unique access codes that will be sent out by post. However, accessing post can be difficult for some communities, particularly the ones you are trying to reach out to. What consideration have you given to other options that do not require that postal element?
Frankie Kay: You are quite correct in those two options. We are looking at trying to predominantly get about a 75% response rate online. Our field officers will have paper forms they can provide to people on the doorstep. At the sort of community events Iain has described, they will also have availability of paper forms, so that if people would prefer to fill those in they will be able to do so in that way.
Iain Bell: Online outreach will be important there as well. Many groups who can be hard to reach and mobile communities will often have online forums where they come together. An awareness through there, so they know they can phone up the contact centre, get their own codes and fill out the form that way, is also part of this activity.
Sarah Champion: What you are doing is fascinating.
Q48 Tonia Antoniazzi: The census White Paper mentions groups that are temporarily housed or those with no fixed abode. It says that further analysis is needed to provide for these groups. What further analysis has been completed and what strategies are you proposing?
Frankie Kay: We have been looking at this further. One of the most important things we have realised, or we are aware of, is to make sure that we engage with these communities as early as possible. We are looking at engaging with those communities. Again, we are aware that you need very specific enumeration procedures for those types of communities. For example, we will look to go to homeless shelters, where obviously those communities quite often congregate. We will look to work with charities such as Shelter so we can understand where those communities are and what the best way to engage those will be. For example, we are aware that some of these people like very minimal involvement with those types of services. If we are looking at a particular homeless shelter, we will do multiple visits to that shelter, so, if people engage on different days, we can make sure we have as much opportunity as we can to ensure those communities can respond to the census.
Q49 Tonia Antoniazzi: We have spoken about some of those communities, like the travelling community, but what about people in domestic abuse refuges? That is quite sensitive. Is it the same strategy?
Frankie Kay: We have specific field officers that will work with communal establishments. A particular field force looks at those types of establishments. We also want to make sure we work with local authorities, because we are aware that some of these, as you say, can be quite sensitive, or they can be quite small. It is really important that we work with the local authorities to understand where those establishments are and that we can then approach those in the most appropriate way.
Q50 Angela Crawley: Coming back to the Gypsy, Roma and Traveller inquiry, we have heard from that particular community that there is an air of suspicion, and that some of the communities have been recorded in official statistics and do not really trust the ONS. What I am interested to know is how you are planning to overcome that issue of trust.
Iain Bell: As I mentioned earlier, it is really about working with the local community to build that trust and working with the outreach groups in order to do it. First off, we use local authorities to help reach and identify where the communities might be. The census advisers who are out there become critical to this. It is them and their role to build trust with that community group that is critical to it. They have to go in and understand exactly what it is they do not trust about us. Sometimes it is the perception that something will happen and that we will give the data to other Government for something to happen.
For the Gypsy, Roma and Traveller community last time round, it was at the time of the car scrappage. They were worried the question on the number of cars was going to be utilised and sent over to another bit of Government. They will not believe us saying it will not happen, but if we build the trust in with the leaders of that group and they are saying, “No, that does not happen. This just makes sure our community gets the funding we need”, actually people build trust.
Then you have to work through the next layer of misinterpretation. We had two rival camps down the road in the last census, both leaders fighting over it, because they thought the funding meant the other camp was going to get the money. The adviser had to explain to them fully that, no, this was about more funding to the whole community. This is why we cannot quite do the “This is it. There is a one-size-fits-all. This is how we do it” type approach. We are going to have to really empower and enable our field representatives out there to build that trust and build that community.
Repeatedly, across many groups, we have found that once the leadership are engaged, they really invest in this to get their communities represented. I know it is totally different communities, but, for example, last time round leaders of the Chinese community funded a census completion bus that went round the major cities with large Chinese populations. The Kashmiri community were looking to put up billboards to encourage completion. We will be looking to work with the Gypsy, Roma and Traveller community in order to get what tactics and what trust-building will work in order to do this.
Q51 Angela Crawley: You touched earlier on the fact that part of your work, as well as your own field representatives, is working with local authorities. As the White Paper states, there is also the issue that local authorities will be key partners in the delivery of the census. What if the local authorities are also mistrusted? What more do you think you can do to work on that?
Iain Bell: They are. We both suffer and are tainted. Local authorities will help us very much with the key partners, enabling and setting up locations, such as libraries, as we mentioned, and helping to identify where the establishments are that may be particularly of interest. There is then the outreach, the marketing and the working with the community leaders that matter. Our job is to get the community leaders fully on board so that they trust us. If they trust us, the outreach activity enables the completion that way. That is the approach.
Q52 Angela Crawley: On that point about outreach and going out into communities, the White Paper states that there will be a number of completion events and people can come along. One of the suggestions was that you should attend public libraries. I wondered if there are any other venues you have considered or scoped out as a potential opportunity for people to come and complete or come to those completion events.
Iain Bell: Certainly there is the library. I know people may have been a bit surprised, but we tested the library for assisted digital in the 2017 test on the Isle of Wight. It was a remarkably positive and well-received event. We are currently tendering our assisted digital offer, as Frankie said earlier. They will help us find where the best locations are for the assisted digital offer, so at this point in time I cannot sit here and say, “They are these locations”, because the partner will help us identify those.
The final bit of the jigsaw is that gives you some fixed venues for things like assisted digital. For the completion events, we go where the community is. Once we establish, with the leaders of the community, where they meet—we talked earlier about the British Deaf Association and other communities—we will turn up at whichever meetings or events will enable completion, in partnership with the local community. Last time round, we were in knitting clubs.
Q53 Angela Crawley: That is fine. You are looking at venues. You are planning to recruit 17,000 field staff to support this completion work. How are you planning to train these staff specifically to be sensitive to the needs of people with various and multiple protected characteristics?
Frankie Kay: The recruitment, the pay and the training of the field officers is something that we are working with a supplier on. At the moment, we are developing the training material in order to make sure, exactly as you say, that people are aware of the need to make sure of that across all of the different equalities. We will be doing that by learning lessons from the 2011 census, but also from our current field force. The field force we use to collect data for our social surveys is a fantastic way of gaining the best practice, in terms of making sure we pass that on to the organisation that will be providing the training for that. It is a combination of making sure we learn from best practice and taking the expertise from the supplier.
In terms of that field force, we are also trying to make sure it is as representative as the communities as well. That is an important part of the recruitment, so that therefore those people feel they can connect with the field officers, in terms of the engagement we have with them.
Angela Crawley: Thank you very much. It is really interesting to hear about the work you do.
Q54 Chair: Before we go on to budgets, can I just talk about this notion of community acceptance, which is a really integral part of the way you develop your questionnaire? How do you manage the problems around groups who are not accepted by the community as readily as those who are? How do you remove the risk that you hide groups of people because of that need? I understand why you need to do that, to have community acceptance of the questions you ask. I am thinking back to your response to the questions on people who might be non-binary, not because they are making a choice but because that is literally the way they are in life. This is not about people who are transgender making decisions. It is about people who are non-binary when they are born.
Iain Bell: The primary job of the census is to count the whole population. For the census to work, everybody must be able to feel included within the census. For the process we undertake, particularly going back to the intersex, gender identity question, the key first stage is to talk with the communities themselves about how they would like to be—
Q55 Chair: Sorry, can I clarify? I meant intersex, not non-binary.
Iain Bell: Yes, intersex. The discussions are very much making sure that we find a solution that is acceptable both to the community, because I need every community to take part for this to be a successful consensus, and to the wider population.
Just to clarify the point earlier about the feminist lobby versus the recognition of transgender, this was not about us saying we were taking a political trade-off here. This was us saying that we had to find a way of allowing people to express their gender identity and, equally, a way of people expressing their sex that was acceptable to both. As Frankie put it very nicely, the sex question then signals that the gender identity option comes later and that allows it. The core component, in terms of decision-making on intersex, was actually that when we talked with the leaders of those groups, they were happy with how they were able to identify in the current form at that point in time.
Q56 Vicky Ford: That first question says male/female. What happens if you tick neither box? What happens if you tick both boxes?
Iain Bell: We take the data as it comes in.
Q57 Vicky Ford: Does it make the whole thing void?
Frankie Kay: No.
Iain Bell: No. We accept the data as it comes in. Our team will then work through how we clean the data in order to provide representative results. At this point in time, there are many aspects of the census where my exact knowledge of how the data coding would work in those scenarios is not quite with us. I will come back to you with details. It may be under development at this point in time, because we are currently focused on how we collect the data. How we process, edit and impute it is still to be decided.
Q58 Chair: Particularly in the world in which we live now, whilst you have noted that Parliament has not shown undue interest in your work to this point, these things mean that changes within nanoseconds. I am particularly interested in the fact you say you have been talking to representatives of different groups. You say that people who represent those who might be intersex have said they are happy about this. It might be useful just to get a little bit more information from you on that.
Vicky Ford: How did you consult them? What percentage, et cetera?
Iain Bell: I am happy to do that.
Q59 Chair: Going on to the very important issue of budgets, what I have been struck by in the way you have been talking this morning is the importance that you put on to building community relations to build trust. As we all know as politicians, to build that trust takes a lot of man, woman and person hours. It is just extraordinarily expensive to do that. You have got an enormous amount of money from the Government—£900 million—but that is for the entire census programme until 2026. We are struck by the fact that there is very little public information about how that money is going to be used. What assessment have you made on how much funding will be needed to deliver the sort of incredibly inclusive approach you have talked about? Do you think the current level of funding is sufficient?
Iain Bell: I will start at the final bit. Yes, I do. Our assessment is the current level is sufficient for all the outreach activities, the development of the online tools, et cetera.
If I may, I will clarify a bit more about the £906 million. This census programme has three elements to it, unlike the last census. The last census developed an online system that was then not used again because of intellectual property issues. Part of the programme is to make sure we roll out the online facility across our business and social surveys, which make them cheaper and more efficient, so that comes in under the £906 million. The second bit is the role of administrative data and looking beyond this census.
If you take out those elements, you get to a figure of just under £800 million for the census. If you look at that in the round, compared to the last census, that came in at £478 million. There are a number of factors that drive the increase in this. First off, we actually just have a bigger population, so we need more staff in order to do more outreach. The first element of that is the growth in the population. The second is inflation and the living wage adjustments. The third element is at this stage of a business case you always have a degree of optimism bias for a programme of this size, because, as you say, things can get a bit more acceptable.
I heard your point earlier that we need to learn from 2011. However, in a total sense of delivering a high-quality census, 2011 was the best respected census for a long time. What I am basically saying, in a long-winded way, is the funding has grown in line with the changes you would expect. Therefore, I am comfortable on that assessment that I have sufficient funding to do all the outreach, all the activities and all the trust-building that I have talked about already.
Q60 Chair: 2011 cost £478 million. The next one, which will be released in—
Iain Bell: The first results will be by April 2022.
Q61 Chair: 2022 will cost £900 million.
Iain Bell: Sorry, no; that is including moving our business and social surveys online and exploring the use of administrative data for whether that can replace the census in the longer term.
Q62 Chair: Those two figures are not directly comparable.
Iain Bell: No. The figure I gave of just under £800 million is closer to directly comparable.
Q63 Chair: It is still a significant increase.
Iain Bell: Yes, it is, but largely driven by inflation, living wage and population growth.
Q64 Chair: £478 million plays £800 million is more or less the comparable figures.
Iain Bell: It is around about there, yes.
Q65 Eddie Hughes: That feels slightly counterintuitive, given the idea that, as we move into a more digital age, you are not printing forms because people can access it online. Creating algorithms that process data means you must be able to do that on a mass scale. I just imagine that, for all reasons, it should be less expensive, not more.
Iain Bell: The difference between online and paper is not that much, in terms of the cost of completion. The main cost is the minute we need to chase up because there is non-completion. Our entire discussion today has focused on how we reach out to these communities who may not respond first time or second time, and how we do that outreach. Basically, our costs modelling is that getting a form online versus paper, if they respond straightaway, is only £1 cheaper.
Q66 Eddie Hughes: You say only, but we are talking about scale here, with millions of people.
Iain Bell: Yes, I know we are, but that is within the figures and that reduced the estimate by about £30 million. Whether they complete online or they complete paper, the big expense is that the minute we need to chase up and send our field staff out, it is an additional £5 to gather every response. More outreach activity and more getting people to respond first time is what will drive down the cost of the census. Online helps. Do not get me wrong; it does help lower the cost, but actually the big driver is getting everybody to complete it. With a larger population, you have factors working against you.
Of course, the other thing to be quite clear about is those figures of £478 million and £800 million are both in cash terms in those years. Much of the change in that is inflation and living wage adjustments.
Q67 Chair: Is it not true to say that in the past we have been more accepting of inaccurate data?
Iain Bell: I would say we always deliver the highest quality data we can.
Q68 Chair: We have been more accepting of the fact that we have not necessarily accurately reflected minority groups. That is why you need the extra money: to be more reflective of absolutely every aspect of society, because of the need to cater for services for those groups.
Iain Bell: There is a worldwide trend of censuses, which is more inclusion and more data around gender identity and sexual orientation, in order for these groups to feel included and take part, to deliver the highest quality census information. Every time we strive to include and to reach out to as many communities as possible. We will continue to put every effort into reaching every community.
Q69 Chair: When do you need Parliament to pass the primary legislation in order to stick to your timetable of undertaking the census? Was it March 2021? Sorry, I cannot remember the exact date. When do you need the primary legislation?
Iain Bell: The primary legislation is a matter for the Cabinet Office, and it is one to follow up with the Minister for Constitution on the exact timing. The core date for me is the end of the legislative sequence. In order to recruit the field, I need the regulations passed. That enables me to recruit all the people into it. If they were passed by April 2020, we are in good shape for delivering the census on time.
Q70 Chair: April 2020 is your cut off.
Iain Bell: For regulations.
Q71 Chair: Otherwise you would potentially compromise your ability to be able to recruit in the sort of way you have been talking about.
Iain Bell: That needs the census order, which is secondary legislation, on both Floors, so that needs to come through around December of this year.
Q72 Chair: Okay, so that is not long.
Iain Bell: It is not long.
Q73 Sarah Champion: What was the population in 2011 and what is it now?
Frankie Kay: We should know that.
Iain Bell: It is just over 66 million just now, UK-wide. I would need to get back to you. I cannot remember 2011 off the top of my head. My apologies, I should know that.
Chair: You have known the answer to almost everything else, so thank you very much. We are very grateful to you for all the work you do and the real care you have taken to look at new ways of reaching out to marginalised communities. That is really important, though you have raised, in a number of our minds, the challenges you have. It has been really useful to hear that from you today. What we will do now as a Committee is consider what we have heard and then perhaps how we can best support the work you do as you go forward. We will get back to you on that, whether it is through a report or whether it is through further correspondence with Ministers. Thank you so much.