Skip to main content

Language: English / Gàidhlig

Loading…
Chamber and committees

Equal Opportunities Committee, 04 Dec 2007

Meeting date: Tuesday, December 4, 2007


Contents


Census 2011

The Convener:

The final item on our agenda is an update on the preparations for the 2011 census. I am pleased to welcome Duncan Macniven, who is the Registrar General for Scotland, and Peter Scrimgeour, who is head of the census division, both from the General Register Office for Scotland.

We move straight to questions. In paragraph 17 of your submission you state that, in the 2006 census test, you trialled a question about negative discrimination

"on 12 grounds including accent, age, ethnicity, disability and ‘other'."

However,

"competing priorities for space mean it is not currently being considered for inclusion in the 2011 Census."

Will you explain the findings of that trial question? Has it definitely been ruled out for the 2011 census?

Duncan Macniven (Registrar General for Scotland):

We have a difficult job fitting all the good questions that one could ask in a census onto a form of a length that can be afforded and that people will not get fed up completing. We are guided by how useful a question is to the people who use the outputs that we produce at the end of the census process. We think that, even though we plan to have four pages rather than the three pages that we had in the past, there will not be space for that question, given the degree of user need.

The other thing that we need to check is that the questions actually work—that people understand them, and that they can give sensible answers. There was nothing wrong with the question from that point of view, but because of the limited user demand for it compared with other questions, we put it on the no-we-will-not-ask-it side of the line.

However, we have not taken a final decision. We will take final decisions on that question and the others in time to put proposals to the Parliament in about 11 months' time. In early 2010, it will be up to the Parliament to decide whether to agree to the regulations that will allow the census to take place or to say, "No, you should think again about that question and include it." Including it, of course, would be at the expense of excluding another question.

That is helpful. Will you give a little more detail on the findings of that trial question?

Duncan Macniven:

We do not want to get into that because the test was designed to try out the way in which we plan to take the census and to try out some of the questions. It was not conducted in a way that would produce statistically sensible and reliable evidence.

Mr Scrimgeour, do you have anything to add to that?

Peter Scrimgeour (General Register Office for Scotland):

As the Registrar General says, we did not do the test with a view to producing numbers that could be purveyed as meaningful statistics. We said that all along the line to a lot of people. However, we did look at the results. We needed to do that to assess whether the question appeared to have worked. I do not say that the statistics are reliable, but it was interesting to note that 16 per cent of people who answered the question said that they had experienced some form of negative discrimination.

A dozen grounds for discrimination were mentioned and the results were spread throughout the categories. The biggest category was discrimination on the ground of age, which 2.7 per cent of respondents said that they had experienced. That was followed by discrimination on the ground of religion, at 2.2 per cent. Age and religion came out top—that is probably as far as we would go in using the results, but they have been published as part of our evaluation of the tests that we did last year.

Your comments are helpful because they put the matter in context.

Marlyn Glen:

You mentioned user demand. I presume that that changes. By the 2011 census, we will have new duties such as the general equality duty, and I would expect local and national Government to be interested in such a question. In asking it, you would be following demand. Is that a problem?

Duncan Macniven:

Not really, because that kind of demand can be anticipated. People know what the general equality duty is and they can build that in to their responses to us on how useful certain questions are.

What methods did you use to improve and test accessibility in the 2006 census test? Have you evaluated the impact of the improvements?

Duncan Macniven:

The main thing that we tried out was in relation to the way in which enumerators interact with people in the area in which they are responsible for collecting census forms. That went well. It confirmed to us that we are right to go with what is perhaps a slightly old-fashioned way of conducting a census. The contact between the enumerator and the household is important. The enumerator can offer help to households that, for one reason or another, find it difficult to fill in the form.

We were both happy and unhappy with our liaison with community groups before the test was carried out. It was successful in some respects. For example, we got around the camp fires with Gypsy Travellers in Lochaber, and we felt that that was successful. On the other hand, I allude in our submission to the big problem of the underenumeration of young people. We were less successful in breaking through those more difficult barriers, because that community is more diffuse, and far more numerous, than the Gypsy Traveller community. There was a mixed message on that.

I want to press you a bit more on the issue of Gypsy Travellers. You talked about getting round the camp fire. Are other methods to improve contact with that group being considered?

Duncan Macniven:

I think not. We have found that personal liaison with groups of the size of the Gypsy Traveller community beats any other method. However, we are open to suggestions if you have found a secret weapon. We felt that personal liaison worked well, because although the number of Gypsy Travellers from whom we got responses was not large, it was larger than the number that the local authority expected us to get. Cross-checking with other methods of measurement is something that we try to apply to the census results and we will be accentuating the use of it in 2011.

I understand that the Gypsy Travellers requested a literacy question. Was that included? Have you drawn any conclusions from the response rate?

Duncan Macniven:

In the run up to the census test, we asked the Gypsy Traveller community whether there were any questions that it would find particularly interesting and it mentioned that one. We did not include it in the end, although we included an analogous question about language—which is rather different from literacy—that asked how well people speak or understand Gaelic or English. We concluded that the literacy question needed to be answered in more depth than was possible through a self-completion questionnaire of the kind that we used in the census test, so we did not include it.

Another of the community's suggestions, which we did pursue and would be happy to pursue further, was that we carry out an investigation into the health of the Gypsy Traveller community. Health is one of the biggest areas in which census information is used, so we would be keen to carry out such work. It seemed that although such a question would appeal to some members of the community, there was a risk that it would go down like a lead balloon with others. We drew back from including it in the test, but we would be happy to discuss it with the community again in the run up to 2011.

We look forward to hearing how that develops.

You said that although you do not plan to include a sexual orientation question, you are taking part in work to identify alternative survey sources of that information. Can you provide more detail on those sources?

Duncan Macniven:

Yes, we can certainly expect the question to be included in the Scottish health survey in 2008. There might be other opportunities to include it in surveys, which we are pursuing with colleagues in the Scottish Government and the Office for National Statistics who are responsible for running such surveys. I cannot guarantee that it will be included in other surveys, but it will be included in the Scottish health survey.

Sandra White:

In your submission you say that ethnicity is one of the most difficult topics and you note the criticism from the race equality advisory forum of the inconsistent use of colour and geography in the census—I noticed that too. Perhaps you will explain the colour and geography terms. Some of the main issues referred to in paragraph 8 of your submission remain to be resolved. Will it be possible to make a final decision that satisfies all the participants in the 2011 census?

Duncan Macniven:

That would be nice, and we are trying very hard to do that. Views within the community, particularly within the African and Caribbean communities, are so polarised that, however hard we try, we are unlikely to be able to come up with a question that everyone likes. We hope to achieve a question that none of the communities hate and that everyone will be willing to answer, which is the key.

There is an element of cutting off noses to spite faces in boycotting the census, because to do so deprives the Government and other census users of the very information on which the policies to address that community's needs are based. So when we are in touch with the community, as we are frequently, I argue strongly in favour of people completing the census, but I cannot guarantee that we will come up with a question that satisfies everyone.

In 2001, there was an interesting attempt to find a short question that people across the population would find easy to answer, but it mixed colour and geography by asking

"What is your ethnic group?"

and then offering options such as "(a) White" or "(c) Asian". That was illogical. Although I doubt that we will be entirely able to escape from that mixture, perhaps we can express the question in a way that is more sensitive to the views of the different communities involved.

Apart from rephrasing the question or finding a question that everyone will answer, do you think that you would be more successful if you used members of ethnic minority communities in the 2011 census?

Duncan Macniven:

Yes. There are two levels to that. We gained a great deal from liaising with all segments of the minority ethnic population. Our consultation in spring this year was successful in drawing out views. Separately from that general consultation, we had specific contact with minority ethnic communities, which has been hugely valuable to us and has allowed us to explain the point that I have just emphasised about the importance of completing the census.

However, if I picked you up correctly, your question was more about the enumerators on the doorstep. This point applies to the Gypsy Traveller community as well as to minority ethnic communities. We would like to appoint enumerators who represent the communities that they are enumerating. We had some success with that during the 2006 test when we had one enumerator and one census team leader from minority ethnic groups, which is in the order of the groups' size relative to the total population. We hope that we will be as successful, or more successful, in doing that in 2011. We will certainly try very hard to liaise with the communities themselves so that they can publicise the availability of opportunities to act as enumerators.

I was an enumerator in 1981—

Duncan Macniven:

Thank you very much. I hope that you will sign up again.

Bill Kidd:

I was obviously too young. [Laughter.] I got the job through the job centre; I was unemployed and I saw the prospect of a temporary job. How can you reach out to people from minority communities who might not go to job centres or whatever? How can you target them specifically?

Duncan Macniven:

The recruitment of enumerators is hugely important to the slightly old-fashioned method of enumerators on the doorstep that we propose to use to conduct the census. That is why I am particularly grateful to you for volunteering in 1981. We will proceed by a variety of means. We will continue to use job centres. We are able to get good enumerators by liaising with local authorities throughout Scotland. As I suggested in response to Sandra White's question, we are also able to take advantage of our links with a variety of communities, so that we can target them.

Peter Scrimgeour:

From our test, it struck me that, although we conduct a census only every 10 years, there is remarkable constancy. Some folk do the job once, for good reasons, and move on to other things, but our test in the west of Scotland showed that there are people who enjoy doing it, who did it 10 or 20 years ago and who want to do it this time. They will be quick to tell us if it is not as good as it was last time.

Sandra White:

That is an interesting question, because we need to know how people are recruited. Thank you for your answer.

One area of concern is recommendation 16 of the evaluation of the 2006 census test, which refers to the need for information on the people who took part. You suggest that

"a profile is created of the ethnic makeup of each Census District involved in the Test."

The committee would like you to explain the purpose of that. Why is such a profile needed?

Duncan Macniven:

I have not picked up exactly the reference that you cite, but I suspect that the answer to your question is that there is not much point in trying hard to recruit enumerators from the minority ethnic community in an area where the community does not exist. On the south side of Glasgow, by contrast, there is a fruitful recruitment ground. The recommendation is a slightly bureaucratic way of saying that we try to establish the size of the minority ethnic community in each area, which we can do from the previous census and from other sources of information.

Sandra White:

You are seeking information on where different ethnic groups stay. You suggest that it would be easier to recruit enumerators from the minority ethnic community in certain areas. What other purpose does the information serve? Perhaps I am looking at the recommendation in a sinister light, although it is not meant to be sinister.

Duncan Macniven:

Not much that we do is sinister—the recommendation is certainly not intended to be sinister.

Some people might construe it as such. You are seeking information so that you can create a profile of the ethnic minority groupings in each census district.

Peter Scrimgeour:

There is a particular focus on ethnic minorities at the moment, but that is part of something that we do across the board. When we have people employed locally, we need to be able to give them advice on the area that they are entering. They may not be working in their own area, but in a neighbouring area. We need to go back to our sources, so that if there are ethnic minority groups in an area we can tell people to ensure that they make contact with them, get out the word that a census is happening and offer them help. If there is a high proportion of older people, we must ensure that our employees know about that. If there are old folks homes, hospitals and prisons in an area, the people who are looking after the census for us there must have background information on that, so that they are better prepared to do the job that we have given them. That is why we are seeking to establish whether a particular group in the population is present in certain areas. We have no sinister purpose—our aim is to get the job done well.

Duncan Macniven:

If I may expand on that, we know where people live—

Big Brother is watching them.

Duncan Macniven:

However, in publishing census statistics, we are careful to preserve people's anonymity, especially when the number of visible minority ethnic people in an area is very small. I am not suggesting that they would be targeted but, because they are small in number, people might learn more about their personal circumstances than they should. We take care to ensure that outputs are anonymised.

Thank you.

I warn you that Bill Wilson is going to ask you some questions that are dear to his heart.

Bill Wilson:

Yes, I have three questions on language, which relate particularly to the Scots language. Which of questions 13 and 36 was used in the test census? Is there any reason why question 13, rather than question 36, is included in the "Possible content" section of the draft census?

Duncan Macniven:

Question 13 in the annex to the paper that I have submitted deals only with Gaelic and is, essentially, the question that we asked in 2001. That worked well. We have asked about Gaelic since 1881, and we will certainly ask about Gaelic in the next census. Perhaps you could remind me of the number of the other question that you mentioned.

It is question 36.

Duncan Macniven:

It is, essentially, the question that we used in the test. We were trying to find out information about languages other than Gaelic for a variety of reasons. There is a good user demand for that question, especially in identifying people who do not understand English or read, write or speak English fluently.

We have parked question 36 under the subheading "Person questions excluded from this draft" because it did not work very well in the test. People did not tick any box against English. We think that that happened because we did not design the question very well and people did not read across to the right-hand side and see the box marked "No ability". People did not tick in the places where we expected them to tick. We are reacting to that by trying to design the question better while continuing to ask about languages additional to Gaelic. We will certainly include question 13; we are keen to include question 36, and we are working on it.

Bill Wilson:

Various reasons have been offered in the past for the failure to include the UK's second largest native language, Scots, in the census. Did the test survey provide any indication of why Scots should not be treated equally with Gaelic in the coming census?

Duncan Macniven:

No. The reason why question 36 did not work had nothing to do with the inclusion of Scots. Earlier this year, we consulted on the questions and on other aspects of the census, and people expressed a demand for information about Scots for the kind of reason to which you allude, but they also pointed out that it was difficult to define. We continue to plan to include that question in 2011, but we need to work a bit more on the detail of it.

I have a final question. In 2001, translations of the main questions were provided in Gaelic. Will they also be provided in Scots in the next census?

Duncan Macniven:

We have taken no view on that.

I hope that you will give it consideration.

Duncan Macniven:

We will give it consideration.

Elaine Smith:

I have a couple of questions about the religion question that you mention on page 3 of your submission, under point 13. In 2005, the Equal Opportunities Committee discussed the wording of the religion question, as two groups had notified us that they did not like the wording of the question in the 2001 census—albeit that it was a voluntary question. Did you change the wording of it for the 2006 test? If so, what impact do you think that that had on the outcome?

Duncan Macniven:

The committee made a very constructive suggestion in 2005 and my memory is that we changed the question, but I will just check that.

In 2001, we asked:

"What religion … do you belong to?"

In the 2006 test, we changed that question to: "What is your religion?"

Peter Scrimgeour:

The full question was:

"What is your current religion, denomination, body or faith?"

Duncan Macniven:

Yes. It is hard to tell whether that change affected the results because, of course, we could not recapture what people thought in 2001, but I do not think that it significantly affected the results. We were happy to make the change because that was a neater way of putting things.

Elaine Smith:

That is fine. Thanks.

Your submission states:

"There seems to be limited user demand for the question on religion of upbringing."

What did you mean by that? Do people know that such a question is available? Have you promoted the possibility that it might be asked? Why is the question not in demand?

Duncan Macniven:

That takes us back to what I said in reply to the convener's first question. Space on the census form is limited. In 2001, we asked two questions on religion—one on current religion and the other on religion of upbringing—which produced interesting information. However, in the consultation earlier in the year that I mentioned, we asked census users and other interested parties whether they were absolutely sure that we needed to continue to ask two questions on religion, given that space is limited on census forms and that only one question on religion was asked south of the border in 2001. The responses to the consultation showed that there was limited support for asking about people's religion of upbringing, but there was absolutely no question but that we should be asking about people's current religion. Subject to the approval of Parliament, we will include a question on that, but there is limited demand for including an additional question on religion of upbringing. There are better uses for the space on census forms.

So you propose not to include in the form a question on religion of upbringing.

Duncan Macniven:

At the moment, we propose not to do so.

The committee may want to take soundings on that.

Peter Scrimgeour:

That was one of the 15 points that we highlighted when we did our consultation in the spring. We asked about the impact that dropping the question on religious upbringing would have, just as we asked about dropping several other questions. We also asked about what other topics people would be prepared to exclude from the form if they thought that a question on religious upbringing was a priority and that it should be kept in. We tried to bring to users the reality that there is limited space for questions on the form.

From an analytical and statistical point of view, as a result of asking both questions on religion in 2001, we got a measure of how people's original religions, their religious upbringing and their current religions tie up. We will lose something if we do not ask both questions, but religion of upbringing results would not change much from the census in 2001, whereas people's current religions will change a bit more. There are arguments about the statistical value of asking twice in 10 years about people's religion of upbringing. Perhaps the question is of less value than other questions.

Elaine Smith:

Somebody could be interested in whether the answers to a question on religion of upbringing change in comparison with those to a question on current religion; they could be interested in whether that could be due to discrimination under the new strand, for example. I wonder whether asking that question might be useful. I am speaking off the top of my head. That said, it is obvious that you have considered the question and that you think that you will have to drop it in order to fit in another question that might be more in demand.

Duncan Macniven:

There is no doubt that the answers to such questions are interesting. We asked about religion in 2001 for the first time, and it produced interesting results, as Peter Scrimgeour said. However, there is so much competition for space on the form that we need more than interest to justify including certain questions on it. We need use.

I suppose that one finding that could have been made is whether there had been a move towards a more secular society. The question is whether there is any value in determining that.

Duncan Macniven:

I think that we will be able to answer that question by asking about current religion.

Bill Wilson:

There may be risks. Some people may feel that the religion of upbringing question is irrelevant, because they decide what their religion is. What is important is what their religion is at that particular point in time. Perhaps they feel that the religion of upbringing question almost labels them as something that they may or may not be later in life. Was there any resistance on such grounds?

Duncan Macniven:

We did not really pick up any great hostility. Because, as you will appreciate, religion is a sensitive subject, it was a voluntary question in the 2001 census; indeed, it was the first time that a voluntary question had been added to what is otherwise a compulsory census. Nevertheless, many people responded to it. It certainly did not stick out as a question with a poor response rate.

That is very interesting.

You have said that you are considering a question on national identity separate from ethnicity and a question on citizenship. What will those questions cover, and what will they seek to achieve?

Duncan Macniven:

As the citizenship question is a relatively simple matter, I will dispose of it first. It will basically ask the person what passport they hold. As you might know, the formal definition of citizenship relates to the passports that a person holds.

Peter Scrimgeour:

That a person is entitled to hold.

There is a difference.

Duncan Macniven:

Absolutely.

Such a question has limited demand in Scotland. However, the European Union has recommended to member states that it should be included in a census. In an ideal world, we would like to respect that recommendation. Of course, by 2011, it might even be a little more formal than a recommendation.

The question of national identity is more complex and seeks to get a sense of a person's Scottishness, Britishness, European-ness or whatever. In the census test, we used the question to preface the ethnicity question, from which we had removed the sense of Scottishness or Britishness that it had had in 2001. However, we have since developed the ethnicity question to allow people to express their national identity; as a result, a separate national identity question does not seem quite so necessary any more and is diminishing in importance and user demand. That said, we have not yet bottomed out whether there is sufficient user demand for other questions to justify its inclusion.

Bill Wilson:

It is important to allow people to determine their ethnicity and nationality. People from all ethnic groups have the right to say that they are Scottish, English, British or whatever, and we must make it clear that a person's ethnicity does not affect their nationality.

Duncan Macniven:

I absolutely agree.

Your submission says that although a question on income might have value, some people might find it intrusive. How do you intend to overcome that tension?

Duncan Macniven:

We have tried to assess the depth of feeling on that partly through our 2006 census test. I have to say that although the income question received the lowest response rate in that test, it was not horribly low. However, when we asked for comments at the end, the question received the most criticism. Obviously, there is a negative element that we need to address.

We are also looking at equivalent evidence from the tests in England, Wales and Northern Ireland, which were carried out this spring. In those tests, the question asked about the income of each person in the household, while we asked about the income of the household as a whole. Those results will give us a cross-bearing. Moreover, we are carrying out some specialised cognitive tests to see what smaller groups feel about the question.

We would like to include an income question because there is clear user demand, and the use to which users propose to put the information is important. However, we do not want to ask it at the risk of putting people off responding to the rest of the census form.

You are now making the census form available for completion online. What impact do you expect that to have on return rates? Can you guarantee the security and confidentiality of returns?

Duncan Macniven:

At the moment, we have not absolutely decided to make it available online, but it is our intention to do so. The spending review gave us a sufficient budget that should enable us to afford it, because it costs rather than saves money.

International experience—other countries have used the internet for their censuses—is that online completion sadly does not make much of an impact on the rate of return from underenumerated groups, such as young men. It would be nice if it did. However, it results in a more thoroughly completed form and, because one can guide people through the form on the internet, it makes mistakes less likely. It is helpful in relation to quality rather than quantity of response, but it also opens the way to making the form easier to complete for groups such as people who want to complete it in Gaelic—or perhaps Scots—and people with sight difficulties, because the computer can enlarge the typeface.

We are aware that you will carry out a further test in 2008 and then a rehearsal in 2009. What are your key priorities at the moment, and will you do anything differently in the 2008 test?

Duncan Macniven:

The tests that we are doing in 2008 are less thoroughgoing than the 2006 test. Our eye is mainly on what we call the dress rehearsal in 2009 and the crucial thing that we will test then is internet completion.

If I may hark back a moment to a part of Bill Wilson's question that I did not answer, the confidentiality of the form is vital. However, we have a handle on that. Other users of the internet have the same kind of demand, if you like—for example, HM Revenue and Customs has online tax return completion—and I think that we are into relatively settled territory, because the public is getting more and more confident about confidentiality and doing things online. We will test that in 2009 and, if it does not work, we will not proceed with it further.

We will try a few other things of less importance. For example, as a result of what we found in 2006, we propose to deliver by post to rural areas in Scotland rather than having the enumerator drive up innumerable farm tracks. We want to test whether that works—we think that it will.

There are a few other matters to do with how we run the business of taking the census. For instance, we will try out having local offices in each area—probably each local authority area. We have never done that before, but it seems to us that it would facilitate the conduct of the census. However, the internet is the big one.

That concludes our questions for this really interesting evidence-taking session. On behalf of the committee, I thank both witnesses very much for coming. We look forward to progress on the content of the 2011 census.

Duncan Macniven:

Thank you very much for the interest that you show in it. The decision on the content and conduct of the census will be taken in this building, and such interchanges are important if we are to avoid unpleasant surprises when we lay the regulations and the order before the Parliament. I welcome the interchange and am happy to repeat the process at any stage that is convenient to the committee.

Thank you very much. That is helpful.

Meeting closed at 13:25.