Official Report 184KB pdf
Item 3 is on the 2011 census. I welcome Duncan Macniven and Ian Máté from the General Register Office for Scotland. I invite them to make a few opening remarks.
As registrar general, I deal with matters demographic and with the census taking. Ian Máté is in charge of the data collection side of the preparations for the 2011 census. No decision has been taken yet to run the census in that year but, in order to keep that option open, we need to make advance preparations; it is not an exercise that can be cranked up in a morning.
You have answered my first question, which was about where you are in the process. Are you confident that you will be able to meet the timescales for the preparation work that needs to be done?
Yes. We have undertaken the task every 10 years since at least 1861, so we are quite used to it. We will stand on the shoulders of the 2001 census, as we have learned a lot of lessons from that. The 2001 census went fine, but one can learn lessons from any project. One of those lessons is that there are some things that we need to start earlier, such as public consultation and engagement with the Parliament—you were not here to engage with in 1995, when we were at the same stage in the run-up to the 2001 census.
Of course. I am interested in exploring some of the areas around involving people and getting better participation. You stress the fact that, these days, people are reluctant to be consulted. Can you explore some of the methods that are being used to encourage people and communities to participate?
We are keen to take as many steps of that kind as we can and take what opportunities come to us, although we can organise some things ourselves. We have organised a pukka consultation in the approved Scottish Executive fashion and have received quite a lot of responses to it—we were not disappointed with that. However, we are also taking the opportunity to make contact with people who contacted us as a result of that consultation. For example, we are setting up bilateral contacts with the community of travelling people to pursue the issues that are important to them. There is a lot of that bilateral contact.
It is important that the people who are involved in consultation feel that someone is listening to them. The work that you are doing in schools is interesting and important. What kind of reception have you had from teachers and people in schools, who might see such things as another burden or another job to do? Have they been keen to take the work on in modern studies departments, for example?
Ian Máté has been in touch with the school community, so he can answer the question better.
The material was written by teachers, because I did not feel that I could make it relevant, as I am not a teacher. So far, I have talked to the three enthusiastic schoolteachers and that is it. I have asked the adviser for modern studies and geography if I can meet before Christmas all the teachers in Glasgow who are using the material so that I can find out how they view it and whether it is useful.
Do you envisage the material being rolled out to other parts of Scotland? Clearly, it is about the census, but we are in a climate in which there are many consultations and, as has been stressed, youngsters are reluctant to participate in them. Will there be something in the curriculum to continue the work?
The material that Glasgow City Council kindly produced for us is relevant to the whole of Scotland; it does not focus on Glasgow alone. I am worried that the material for the initiative will just sit on a shelf in schools when we have paid the £16,000 or whatever for it. We might have to go to teacher conferences to try to push the use of the material. In December, I will try to find out from the teachers how much resource we need to roll out the initiative throughout Scotland. Learning and Teaching Scotland has said that it is willing to roll it out.
That is why I asked the question. Good material often lies on a shelf because there are not enough resources to back it up.
Do you work with local authorities to tap into the teacher training days or continuing professional development?
That is the kind of work that we hope to do now that we have the teachers pack, which was published last month. As Ian Máté said, we have good links with Learning and Teaching Scotland and with HMIE.
The respondents to the autumn 2004 census consultation highlighted the fact that certain hard-to-reach groups may not have been included in the 2001 census, particularly communities for whom English is not the first language. You have already pre-empted and answered my second question, which was about hard-to-reach younger people, but what about people in the first category?
The issue of people whose first language is not English has always been a concern for us and we have always had strategies for dealing with it. Those strategies have been fairly successful, but they could be improved, which is another lesson from 2001 that we will apply. As we said in our paper, we are keen to establish as many links as possible with minority ethnic groups throughout Scotland to ensure that we ask the right questions in the right way and with the right supporting material. That is relatively easy, because those groups are, or appear to be, well organised. It is less difficult than working with the English-speaking 18-to-30 age group that I mentioned, which is not homogeneous or well organised at all. In West Dunbartonshire, which is one of our test areas, we have good contacts with groups of young mothers. Young mothers are not such a difficult group for us as young men are, but that is the kind of direct contact that we try to build up. The good liaison that we have with local authorities in the areas where we propose to carry out the census test next year will be helpful.
The Royal National Institute of the Blind and Deafblind Scotland have criticised the 2001 census for its inaccessibility. For example, Deafblind Scotland suggests that guide communicators could be provided to assist people. What measures does your organisation plan to take to make the 2011 census more accessible? Are there any plans to test the measures in 2006?
That is another matter that Ian Máté has studied.
First, in a general sense, I am trying to change the role of enumerators so that they can fill in the form on the doorstep. The enumerator presence will be slightly enhanced. Going down the technological route removes that context.
Are you quite confident that you are going down the right road for the next census to improve on the previous census for such inaccessible groups?
I am not quite confident; I am still trying to find out how to achieve that. I cannot claim to be an expert yet but, if we have the RNIB on board, if the DRC provides awareness training in the run-up to the census test and if all the methods are tested in 2006, we will be some way down that route. The main problem that may still remain is encouraging people to open doors when people are more wary of opening doors. We certainly have a problem with older people and we imputed many more older people than middle-aged people, largely because older people are difficult to contact, as they will not open doors.
I will explain the reference to imputation. We reckon that 96 per cent of the population supplied us with a census form. Rather than saying, "Och well, here's 96 per cent," we applied statistical techniques to fill in the remaining 4 per cent, so we have data that cover 100 per cent of the Scottish population. We had to impute—to assign values to—the returns that we did not receive in the proper fashion. Everything that we do is designed to reduce the amount of imputation that we must undertake.
Respondents to the autumn 2004 consultation also suggest that advertising the census could improve in 2011 to say how census information is used and how it can help communities. Such advertising might make the census more relevant and improve response rates from hard-to-reach communities. What plans does the GROS have to do that?
We are at quite an early stage to be thinking about such advertising, particularly if it means paid advertising. On formal advertising, Ian Máté has established contacts with the advertising world—we can use the Scottish Executive contract for that. We have secured the advice that will eventually lead to a good solution, but it is too early to say what that solution will be.
On unpaid publicity, do you have any promises of access to the BBC? If information is broadcast on television, the message gets across to the bulk of the community—apart from the blind people who obviously cannot see it. Is co-operation from the BBC, for example, guaranteed?
I would be very surprised if we did not get co-operation. The census has traditionally had good support from the BBC and the rest of the media. I was not in post in 2001, but my predecessor and our opposite number south of the border had meetings with the editors, as did ministers. There is good support for the census and I would not criticise it. Even at this stage, we have found that the media have a helpful attitude on the census test. The census interests people; it has a sort of magic and an aura about it.
I like the idea of the census having a sort of magic and an aura.
I very much agree that that community is one to which we must pay special attention. That is not a new issue for the 2011 census, as that group has traditionally presented difficulties to us and our predecessors. Ian Máté has been in direct contact with the community, so he is perhaps better placed than I am to answer the question.
About 200 people affirmed in 2001 that they were Gypsy Travellers or Romanies, so we obviously did not identify the community. We may have had responses from many more Gypsy Travellers or Romanies in 2001, but I doubt that we did. They are difficult to enumerate.
For our other work, it is important that we have a realistic census return. It would be interesting for us to consider the issue, to see how it works. I am also interested in the fact that you are having meetings around the country. The committee is always concerned that, when we meet representatives of any community, we may be meeting self-appointed gatekeepers and hearing only their views, rather than the views of the rest of the community. Are you aware of that problem?
We are very conscious of it with a couple of the groups that we have mentioned. We will read with interest the Gypsy Traveller report that the committee is about to publish. Although we are Edinburgh based, we are not Edinburgh fixated. Crofters are another community that we have not discussed. We need to be sure that we have caught the essence of the crofting community, so we will run a focus group on it in Stornoway.
The respondents also suggested that the recruitment of enumerators from ethnic minority communities, such as the Gypsy Traveller community, would encourage participation from those communities. What are your views on that suggestion?
We have discussed the suggestion with the community. Whenever we go through the detail of it, some people indicate that they would like to be enumerated by a trusted outsider, some indicate that they would like to be enumerated by someone in the encampment—we are happy to employ someone to do that—and some indicate that they would like to be treated like everyone else and to be enumerated by a census enumerator. Those are the kinds of issues that we must work through in the census test, which is not compulsory, to see how we will operate. I am afraid that I do not yet know how we will proceed.
It is a difficult question.
The committee notes that a question on income will be included in the 2006 test. Can you update the committee on the background to that? Why do respondents find it so controversial?
Many of our users would like to have better information on income. That is the key to allowing Government to focus on areas of particular difficulty across the policy spectrum. We would dearly like to include a question on income in the census. That is not a new thought. We included a question on income in the 1997 test that was used in the run-up to the 2001 census. We drew back from including such a question in the 2001 census, because the test results suggested that it would diminish the rate of return. As I have explained, that is important to us, so to include a question on income would have been to throw the baby out with the bath water.
If you included an income question, would it be split into different boxes that people would tick or would it ask them directly for a figure?
It would consist of tick boxes with bands of income.
It might help members of my generation, many of whom live below the poverty line, if you indicated what the current poverty line was and simply asked people to indicate whether they were above or below that line. Something simple like that would enable people to tick a box without declaring how poor they were, living on a pension in 2011.
That is an interesting idea that we can reflect on. The way our tick boxes are organised comes quite close to that, but it also gets information about the degree to which people are above the poverty line, which would also be desirable. However, if our test suggests that we cannot collect income information in the way that we would like to, we could fall back on the method that you suggest, which focuses on an important aspect of the issue.
It would be interesting simply to have a bar and ask people to indicate whether they were above or below it. The issue is a wee bit difficult, because we have a household income question, which means that the poverty level depends on the number of people in the household as well as on the income level.
The 2006 test form will have a question asking whether respondents have experienced discrimination in the past week. What is the background to that question and what will the results be used for?
The 2001 census had an ethnicity question that mixed geography and colour and was generally unwelcome. However, people are discriminated against on the ground of colour. We felt that, if the ethnic question reflected geography, we should have a question alongside it that would pick up discrimination on the ground of colour. That was the starting point. Once we started discussing the issue, we found out that a lot of people welcomed the question. For instance, Age Concern wanted a question on discrimination on the ground of age and the Disability Rights Commission wanted a question on discrimination on the ground of disability. Each organisation wanted some way of determining the level of discrimination against the people whom they represent.
How would the question be worded? Would you be asking about any form of discrimination?
We ask about all of the forms of discrimination that have been legislated against or which will have been legislated against by 2011. There are 10 tick boxes, but there ought to be another one that says "None". We ask about: accent; colour; ethnicity; language; religion, faith or belief; sexual orientation; age; disability; gender; and nationality. We think that it will be quite interesting to use the ethnicity question to determine how groups feel that they are discriminated against. That will give us an idea of the quality of life that they have in Scotland as well as what ethnic group they belong to.
That is a huge improvement.
For question 2, the basic demographic question, I think that we have always used "sex"—
In the past.
And it has worked. If it ain't broke, don't fix it.
But you use "gender" in question 16.
We deliberately use "gender" in question 16 because it conjures up a different image. It is asking not, "What sex are you?" but, "Are you being discriminated against on the ground of your gender?" It is, perhaps, a subtle difference.
Very subtle.
We would be happy to have another look at that.
Good.
I suspect that the honest answer is that question 16 has evolved—
And question 2 has not.
It has evolved to a point that question 2 has not.
Exactly.
Nonetheless, question 2 has always worked. We do not have a problem with the way in which people answer question 2, which suggests that people do not have a problem with the way in which question 2 has been asked.
I see.
We shall look at it again.
We would be grateful if you could.
My first question follows on neatly from that. The Equality Network's consultation revealed that some respondents thought that a question on gender identity should be included. Do you think that that is feasible? Do you have any reservations? In view of the fact that the male/female question has always been effective, might that be compromised by introducing a question on gender identity?
To give a personal point of view, the answer is yes, but we obviously have to test my personal points of view. If we have something else in there, people may give a frivolous answer, so we always have to be a bit wary of that. For instance, we had 14,000 Jedis in 2001. If they were given the opportunity to write things in, people took that opportunity.
There were also Terry's old geezers and gals, among which I hope none of you numbers yourself. The problem of frivolous responses is always at the back of our mind.
So you do not really see a way of introducing that question?
It is too early to say.
Are you testing it?
No. We were testing a sexual orientation question, but that is as far as we went. That is obviously different from gender and sex, but we are not testing such a question.
I notice that the sexual orientation question was included in the consultation in the summer but that it is not in the draft 2006 test. Are you testing it in some other way? Was there a reason for missing it out of the 2006 test?
My personal opinion was that such a question would affect coverage. In the 2006 census test, we are already taking a risk with coverage to test the income question, which, from the point of view of our data users, is much more important. We decided not to include the sexual orientation question in the 2006 census test, but Richard Morrison, who is sitting behind me, has done a 4,000-household survey with a fairly bland set of census questions. He added the sexual orientation question to 50 per cent of them to test it. We are keying in the data now; we have had a 28 per cent response rate, and there is no difference between the sexual orientation returns and the non-sexual orientation returns. We are now looking at the accuracy of the returns, and we also asked people for their opinions on all the questions to see what they thought about that question.
Your survey indicates that the inclusion of such a question would probably not have a detrimental effect. Is it now too late to include it in the 2006 test, or do you not want to have too many test questions in the test?
We are running out of space on the 2006 census test questionnaire. We are still waiting for information from the racial equality scheme implementation group. We have left a space for the group, but it appears that it is going to ask us to run two questions.
Do I recall correctly that you will have another census test in 2008? There might be an opportunity to fly that one again, as it were, at that point.
That is right. We would have preferred to have flown it in 2008 if we had felt that that was the way that the wind was blowing.
So if you thought that a sexual orientation question was coming in for 2011, you would want to do the test then.
Yes.
Another thing that the Equality Network consultation threw up was the suggestion that an online form could encourage people to respond, because they might feel that that was more private. What consideration have you given to that?
We have been thinking hard about online returns, and we have been making some international comparisons. On one level, and from one point of view, the technology gives us a really good, heaven-sent opportunity that should offer a way of capturing men aged between 18 and 30 in particular. If using an online form achieved that, we would be very keen for that to happen. Based on our international contacts, we are a little sceptical about it at the moment, particularly because the cost of adding on that option would be significant. We are currently not including online returns in our 2006 test. However, our opposite numbers south of the border, who perhaps have deeper pockets than we do, are planning to include them. We will look very carefully at what that does for the return rate there. If it produces a decent increase in the return rate, we will be keen to adopt it.
The point is well made. The last time we discussed the matter, we very much took that on board in considering whether to ask for the question that, in the end, you included. We were conscious of the law of diminishing returns: the more we put in, the less we get back—roughly speaking.
I absolutely agree, although that is another personal view. I personally do not see very much benefit in web-based completion. Canada, New Zealand and Australia seem to be spending about $10 million on web-based completion, the completion rates for which are about 10 per cent. I do not think that people who would not complete a paper-based census would complete a web-based census. Web-based completion does not cover underenumeration in that sense. Those who do not fill in the census form do not think to themselves, "Oh, goodie, now we can comply by filling it in on the web." They take a different view.
This is quite a political question, because it is a matter of resources. I am dredging up my personal memories now. Was there more in the way of personal contact when the forms were collected during the 1991 census than happened during the 2001 census, or is that a misperception on my part?
It is not really a misperception. In 1991, the enumerators dealt on average with 250 to 350 households; in 2001, they dealt with 350 to 400 households. The methodology might have been the same, but we had fewer enumerators. Because we used post-back forms, the enumerators did not have to return to so many households. Therefore, we increased the number of households that each enumerator could deal with.
That is helpful, and it is something that we might want to bear in mind regarding the resources that are allocated to the census.
Question 6 of the 2006 census test is a very important health question:
I want to compare the language used in question 7 with that used in question 18. Question 7 asks:
Question 18 was in the 2001 census, and we have adapted it slightly to have one more tick box and to break down the numbers of hours a week of caring. The other question is entirely new.
They do not match up, and from a statistical point of view one would want the questions to match up.
I worked with that question after it was used by enumerators in Dublin. It is very difficult to ask certain questions if one has an enumerator-based collection. One would have to go into a house in which one of the residents suffered from mental ill health to ask, "Are you mentally ill?" That creates a wee problem. The wording in question 7 is, perhaps, softer.
I do not want to talk about the Scottish Executive's see me campaign; however, we are trying to get rid of the stigma that is attached to mental as well as physical ill health. I am concerned about that. I see that you are trying to be helpful, but, in trying to be helpful, you could cause confusion.
So you think that, rather than use the wording
Yes.
I see your point. We will be happy to have a look at that.
That is smashing—thank you.
I should say that the form that you have in front in you is not the settled form for 2006. There is an excellent opportunity for such comparisons to be made. It is interesting that your fresh eye has noticed two points that had not come over to us strongly.
I return to the ethnicity classification. The committee understands that the Executive has just concluded a consultation on proposals to review the ethnicity questions for the 2011 census. Will you give us the background to the review and update us on when the results will become available? You mentioned that you were expecting two questions on ethnicity to be asked.
The consultation period has just finished and we have had a report on the responses to the consultation that blandly states what the responses are and whether they are from an organisation or an individual. From those responses, we will develop our recommendations, which I think will come out at the end of October. That is the current position.
Is the GROS satisfied that, in carrying out the consultation on the census ethnicity classifications, the Executive ensured the widest possible participation of community groups?
The problem of getting through the gatekeepers has already been mentioned. I am not sure that we have managed to do so completely, although many of the respondents say that they consulted people in their groups.
The 2006 test will be the first examination of the modified ethnicity questions. What plans has the GROS put in place to ensure that the questions will be tested effectively?
The plans are those to which Ian Máté has just alluded. We have been careful to pick an area that is the richest in different groups from the ethnic minority community, because we attach such importance to that community.
We still have the problem that we do not know who we are going to use as interviewers for those delicate questions. It would be quite nice to use members of the ethnic communities but we always have problems recruiting them.
We have come across that problem in many areas.
We are very conscious of the Scots language and what the group has said about it. Perhaps Ian Máté is the best person to tell you what we are doing about it.
We have developed a matrix question that meets all the needs that have been expressed, but we are not sure whether it will work—that is what we are going to test. It tries to get at literacy, whether people speak English or another language, whether people speak Scots or Gaelic, whether someone's first language is an ethnic minority language and whether someone uses British Sign Language or other methods of communication. It is the Jack-of-all-trades language question.
I must admit that I had some concerns about the question on Scots. What kind of Scots are you asking about? I mentioned it in the north-east and people immediately asked whether it was about Doric or Lallans. What is Scots? I would be interested to know whether my colleagues would tick the box or not.
Is the tongue that people speak in Shetland Scots or not? That is one of the issues that arose during consultation. It is hard to define Scots, whereas Gaelic is not hard to define. Question 10—the one that Ian Máté spoke about—is basically the same as the question that we asked about Gaelic in 2001, on which we are publishing a report next Monday. The question worked pretty well, but extending it might run into problems. We will see. It would be good to have a question on the Scots language in the census, but there are practical difficulties—we are exploring how great they are and how they can be overcome.
It is important that you find a way to do it. We know that many Scots speakers do not recognise that they speak Scots and that many of them cannot read or write it. The issue is quite difficult, because of the different regional aspects.
Why are the language questions closed and not open? Would it be better to ask "What is your native language?" or "What language do you speak in your normal day-to-day activities?"
It is far easier to construct precise outputs from closed questions. Precise outputs are important. However, we use more open questions or we have a box marked "Another—please specify". For example, we had such a box for religion in the 2001 census and we got the Terry's old geezers and gals and the Jedi knights, to which Ian Máté alluded, as well as proper religions. However, the complexity of analysing those returns meant that we did not get the information out in the first-round results.
The languages question in the 2006 test will ask people to indicate which languages they can understand, speak, read and write. Evidence from the Commission for Racial Equality suggests that people may find the format of the question confusing. The CRE is concerned that respondents may enter all the languages in which they can communicate, including BSL and French, for example. Further to your previous answer, are you confident that people will understand the question or will its format confuse them?
We are not confident on that and we will have to test it. That said, the format worked for Gaelic in 2001, which is not a poor starting point.
If the question does not work, we have a problem. We now have a Gaelic question and a Scots question, which we tried to develop fairly rigorously in 1997 through cognitive research and that sort of thing. We have the literacy question to address the issue of literacy levels. Moreover, the issue of service needs requires a question on whether people speak English. We have a horror of ending up with about five questions in order to meet all the needs that the language question is trying to address. Although that may eventually have to become the solution, I hope that that is not the case.
Are you trying to concentrate on the question of people's first language?
No.
That would be a different question. The issue is complicated.
It is. As we have said, we feel that we have not thought our way through the issue as yet. We have time to do so and the test will help.
We need to know from the communities what languages people talk among themselves and whether they also talk English. The question is not just about one language.
If I may, I will return to question 9, which is on marital or civil partnership status. I like the change that has been made to the question. The Justice 1 Committee is looking at the Family Law (Scotland) Bill, which provides for legal status to be given to cohabitation. As that change may come into law by 2011, are you thinking of adding a question on cohabitation?
No, not to question 9. The issue is picked up in the questions on household structure on the first page.
In 2011, as we did in 2001, we will run a relationship matrix, which includes the option of cohabitation. The question is not included in the current paper for economic reasons. We do not need to test it again.
But the issue will be picked up at the household level.
The Scottish Council of Jewish Communities thinks that the question
It would be possible to amend the question. The wording in the 2006 form is the same as that used in 2001, which was the first time that we asked a religion question in a census—we had to change the legislation for that. I feel that the question worked quite well in 2001 and that the arguments for having a continuous series to allow us to examine changes over 10 years might outweigh the point that the Jewish community has made. However, we will consider the suggestion.
Last week, we had a meeting with a liaison group of religious leaders, at which the Buddhists made the same point: Buddhism is not really a religious body to which people belong. The other religious leaders agreed. However, we have not yet changed the wording of the question, mainly because I am worried about comparability with the 2001 census. Many of the users of the results from that question were academic, so I want to ask academic users to consider what the effect of such a change would be. If we ask "What is your current religion?" we might get a much larger number of people affirming a particular religion than happened with the present wording, which might mean that we lose comparability. However, in the census test, we will move towards the wording that you have suggested.
I imagine that the point is that, if we ask which religion people belong to, some people might consider that, unless they regularly attend a church, synagogue or mosque, they do not really belong to a religion and so would have difficulty answering the question.
We must decide whether the purpose of the question is to find regular churchgoers or people who just generally affirm a religion.
I think that, with the kirk, the number of people who chose the option Church of Scotland in 2001 exceeded considerably the membership figures that were held by the headquarters of the church at 121 George Street. Therefore, it seems that people have not applied the strict definition that the Buddhists and Jewish people who have been mentioned may fear.
I am slightly sceptical about the wording of question 20 in the 2006 form, which is:
That is interesting—I will read the question with a fresh eye. However, I believe that the question is a traditional one that has been in censuses since time immemorial and it seems to work, despite its lack of English logic.
We will consider the wording again, but the question reflects the output. We normally ask a series of complicated questions and I did not have room for more questions, so I thought that I would try a simple solution.
I have no intention of changing something that has been used for ages.
The question reflects the output of the census rather than the questions that we have asked in the past.
It sounds as if there is scope to pick up Mr McGrigor's point, so I thank him.
I can be just as niggly. Question 5 has a diamond bullet-point in it that says:
In 2001, we carried out for the first time a follow-up door-to-door survey that followed closely on the heels of the census. We got responses to that and compared them with the census responses. Some folk completed both; some completed neither—we got no reply from them; and some completed one but not the other. By comparing the data that we got, using statistical techniques, it was possible to estimate the responses of the 4 per cent of people who did not return the census and confirm the responses of the 96 per cent who did.
Is there an argument for applying those survey techniques to the census, so that you get fuller information the first time around instead of having to impute it later?
Yes, that is exactly what we are trying to do. We would prefer to get 100 per cent the first time around and for the follow-up survey just to confirm our results.
You said that you got a 96 per cent return. Did you have a 100 per cent issue? Is there a mechanism for picking up households that someone might come across but that you do not know about? Can enumerators add households?
Yes. In 2011, we are going to strengthen two things. The first is the address list that we start with. Last time, we started with a good address list, but developments in addressing—we keep in touch with those developments, although the issue is the responsibility of local authorities—will allow us to move forward a lot and to avoid situations in which, damn it, a house has been demolished or a new set of houses has sprung up without our knowledge. We see that quite a lot round here. We will have a more up-to-date and complete address list, which will be better tested because it will be used daily by local authorities.
The issue of confidentiality is mentioned in the test census form. However, in order to get accurate information, people must be confident about the confidentiality with which their answers will be treated, so perhaps that message should be splattered everywhere.
That is an interesting subject to address, without taking up too much space. We emphasised the confidentiality of the census at the time that it was taken. Perhaps this is a naive observation, but I feel that people are confident in the census as a brand and in the registrar general as a confidential custodian of information who is independent, or semi-independent, of the Government.
It is hard to reach people who are not sure of that confidentiality.
I am not sure whether an awful lot of people say, "The census form says it's confidential, but I don't believe it and won't complete it," but that may be part of the penumbra.
Is that partly why people do not want to put down their income? They are sort of confident but have concerns.
You may be right. That is certainly part of a penumbra of factors that affect folk's attitude to the census.
I notice that the census test for 2006 involves the household form. Thousands of people in the country are not in households, such as members of Her Majesty's forces and those poor souls who are incarcerated in prison. Are people who are in prison given the visitors page? Who fills in such forms? Does a prison governor do that? Is a way found to obtain the answers that you require from the thousands of prisoners or are they ignored?
That is a good point, which also applies to people who are in care homes or in hospital. You are absolutely right: we have given the committee the household form. We have a separate arrangement for enumerating what we call communal establishments, such as those that we have spoken about. You picked prisons as an example. Ian Máté can talk about that in more detail and describe how the system operates there, if that would help.
I am sorry; I do not know how the system operates in prisons, because I was not involved in such work in 2001. I will know later, but I suspect that prisoners fill in a form themselves. The problem that we have had is that prisoners do not take seriously questions such as those about whether they have been there for six months, which makes them resident. When they are asked questions about their expectations or what their last job was—when they were last employed—they take that as a joke.
Will you update the committee on the next steps to develop the 2011 census?
As you will have gathered, our eye is on the 2006 test. We will run that and probably publish the results in a new version of the consultation paper that we issued a year or so ago. In the light of that, we will make adjustments for the rehearsal that Nora Radcliffe talked about, after which we will repeat the process in the lead-up to the census.
The committee will continue to monitor with great interest the development of the 2011 census. Thank you for appearing.
Meeting suspended until 11:38 and thereafter continued in private until 11:50.
Previous
Reporter