The Scottish Government

« Previous | Contents | Next »

Listen

C hapter three the views of data users

Introduction

3.1 Clearly, there will be varied perspectives on approaches to data collection in relation to ethnicity, and more broadly identity. The major part of this research has focused on people's views in relation to people answering questions on their own ethnic identity. In addition to this it was considered important to explore the impact of any proposed changes on organisations that could potentially be using such data to inform policy and service provision. Within the remit of the research it was understood that these findings would inform a much wider consultation process, and as such the number of interviews was limited, but informative.

3.2 In collaboration with the project steering group, the research team selected a range of organisations to reflect the different types of uses made of ethnicity data, encompassing both local and national perspectives. Previous research undertaken by the Scottish Executive was also used to identify some organisations that had developed ethnicity classifications that were different from the 2001 Census question. Steps were also taken to ensure that we captured the views of professionals involved in the processing and analysis of data as well as those who were more concerned with services and policy.

3.3 This chapter sets out the perspective of data users on ethnicity classifications. It begins by exploring the reasons why respondents were collecting data on ethnicity and how the data was being used. The chapter then moves on to look at the classifications respondents were using in relation to identity and the ways in which ethnicity classifications could be improved (with a particular emphasis on the 2001 Census question). It finishes by exploring what impact any changes to the Census question might have on respondents' organisations.

Reasons for collecting data on ethnicity

3.4 Respondents gave a number of reasons for collecting data on ethnicity. The previous research undertaken with data users by the Scottish Executive identified a number of organisations who saw the main reason for such data collection as associated with compliance with legislation. This was certainly mentioned by respondents in this research study, however, a host of other reasons were also cited that showed that organisations were interpreting the data and applying their findings to working practices. In addition, whilst compliance with race relations law was seen as important, a few respondents thought that the spirit of the law meant that organisations and bodies ought to be striving to go beyond data collection:

'We've decided to go beyond that [the legislation] and we saw the spirit [of the law as]… not just monitoring but you know to understand what the position is in the sector in terms of how well different ethnic groups need to be met and so on.' (Male, Age 18-25, Group No.5)

3.5 A number of the organisations interviewed talked about monitoring for discrimination and devising policies to promote equal opportunities, which included setting targets and developing outreach work. In particular, respondents involved in policy making gave some examples of how this had been done. Examples included:

• An organisation wanting to ensure that their recruitment practices were promoting equal opportunities surveyed staff on a regular basis including a question on ethnicity. The results were then used to inform the development of a strategy to increase the numbers of staff from ethnic minority groups. In addition, they had recently begun to collect data on the ethnicity of line managers, as they were particularly interested in the career paths and promotion opportunities for staff from minority groups.
• Another organisation identified that the specific population they worked with included higher numbers of people from minority ethnic groups when compared with the Scottish population as a whole. This had led them to plan for more detailed data collection to explore whether there was a different level of service usage among minority groups so that planning could be better informed.

3.5 Similarly, one organisation used data on ethnicity and other specific information, like language and religion, to tailor their services directly to the needs of individuals by identifying needs for interpretation services and spiritual support and then putting services in place as appropriate.

3.6 In another instance, an organisation saw a role for collecting data on ethnicity as awareness-raising as they felt it signalled to the public that the actions of staff were being monitored and, they hoped, highlighted to staff that awareness of racism was important to their employers. Similarly, another organisation conducted trends analysis to study variations in health and disease in relation to ethnicity.

3.7 A number of the respondents also mentioned the need to respond to requests from parliament, government bodies and the Cabinet Office for information on ethnicity among staff and/or service users.

3.8 Resources were also mentioned, with some respondents talking about using ethnicity data to see if additional resources were needed, for example for interpretation services, or whether existing resources could be targeted more effectively.

How ethnicity data was being analysed

3.9 As mentioned in the previous section, organisations had a variety of reasons for collecting data on ethnicity and this was often related to how they used such information. More specifically, respondents talked about how their organisation's analysed and interpreted the data and the implications this could have for dissemination.

3.10 Taking analysis first, organisations described using standard statistical techniques to compare groups within the data-set. This could occur at two distinct levels: large-scale and localised. Large-scale comparisons meant comparing their population with that of the Scottish and/or UK population as a whole. Here data users were asking questions like 'is our staff profile broadly representative of the population as a whole?' this would include a focus on ethnicity. Clearly, any comparisons would be limited to ethnicity as defined in the national statistics that organisations were using to measure against, for example the last census.

3.11 In terms of localised analysis, respondents were looking more closely at how their organisation operated. This might mean looking at the numbers of people from ethnic minorities in different sectors. Respondents mentioned studying staff recruitment and student registrations to see whether changes needed to be made (for example, by reviewing their strategy to encourage more people from minority groups to join), and also exploring how people from minority groups were progressing in their organisation.

3.12 Another approach was to look at staff behaviour towards the general public by collecting data on ethnicity during interactions (an example being the use of police 'stop and search' forms). It was suggested that such data could then be used to see if people from minority groups were being unfairly targeted and would also allow them to investigate any allegations of racism more effectively. However, further research would be required with operational staff to explore how this was being done in practice.

Statistical analysis

3.13 For any meaningful statistical analysis, data needs to be organised into significantly sized groups. Current ethnicity classifications suggest that the numbers of people from minority ethnic groups are very small when compared to the Scottish population as a whole; this is particularly pertinent outside the two largest cities since most people from minority ethnic groups live in either Glasgow or Edinburgh. When organisations have analysed the data they hold on ethnicity it can be frustrating because they cannot compare groups, the number of people involved being too small. To help with this, respondents mentioned aggregating (combining) ethnic groups. By way of example, this might mean bringing together people into one 'umbrella' category, such as 'Asian', regardless of whether they describe themselves in a more detailed way - Pakistani, Indian, Bangladeshi and so on. By doing this they can then make comparisons with other ethnic groups and look at issues of gender and so on, more closely. This is magnified for very small groups in the population that might be classified under 'other' or where all groups not classified as, for example, 'White-Scottish', are labelled 'minority groups'. An example was given of one organisation using four aggregate categories, namely: 'white', 'black', 'Asian', and 'other'.

The importance of confidentiality and anonymity

3.14 Data user respondents had a desire to make public the statistics they held on ethnicity, and in some cases data was made available. However, this could only be done in aggregate form and respondents were extremely mindful and concerned over issues of anonymity; they feared identifying individuals and the potential for harm that could result from this. They explained that this was of particular pertinence to Scotland as the numbers of people from minority ethnic groups was very small. However, it should be pointed out that comments made by respondents suggested that data users, particularly among those with an equal opportunities remit, were studying confidential data in more detail. A few respondents remarked that just by looking at statistics collected for their jobs they could identify a particular member of staff because of their ethnic description and gender. They felt that their responsibility towards keeping that information confidential was paramount but that this still allowed them to consider any issues of discrimination internally as an organisation.

Dissemination of statistics on ethnicity

3.15 Whilst there were concerns over anonymity in relation to publishing ethnicity data, respondents talked about a variety of ways in which aggregated data was disseminated. This might be because they were required to by law, for example, with civil service statistics or because special requests had been made by Parliament (in the form of Parliamentary questions) or by a member of the public. There was also mention of data sharing between government departments and bodies. One example took this a stage further with two organisations who were pooling resources to run a survey among local residents to better inform strategy for their local area; collecting data on ethnicity was seen as part of this process.

3.16 Overall there was a sense that data dissemination outside an organisation was in an 'official' capacity and in response to external requests. There was no evidence among our (limited) sample that information on ethnicity statistics was being disseminated more widely, or that respondents saw a need for this. As mentioned earlier, there certainly were worries over ensuring that such statistics would be handled sensitively so that anonymity would be protected and this may have led data users to be particularly cautious. Equally, though there was nothing to suggest that data users would be averse to disseminating data so long as confidentiality needs were completely addressed.

Current classifications being used by data users

3.17 Data users interviewed for this study described three different reasons for using the ethnicity classification they were using ⁸:

• The 2001 census question on ethnicity or something very similar
• Data users chose to use the 2001 Census question so that they could compare their population with national data. Where there were differences in the numbers of people from ethnic minorities ( i.e. less people than might be expected), staff used the data to review their practices in areas like recruitment.
• It was pointed out that using the same question also made it easier for organisations to develop joint initiatives.
• An ethnicity classification tailored to specific needs

• One organisation asked a series of questions in order to match services to individuals. Questions were asked about ethnicity, nationality, language and religion. However, the services available were orientated towards language and religion rather than ethnicity or nationality.

• An ethnicity classification developed many years ago
• As with organisations choosing to use the Census question on ethnicity in order to make comparisons, one organisation chose to stay with a question they had developed many years earlier because they wanted to compare statistics. However, in their case it was so that they could carry out trend analysis over time so they could study changing patterns in their population and adapt the organisation accordingly, for example, by engaging new language interpreters. It was also commented that the question(s) they used worked for the organisation and hence change was not thought necessary, something we return to later in this chapter.

Data users' views on the ethnicity classifications they were using

3.18 Data users' views on the ethnicity classifications that they were using at the time of interview varied. They ranged from being happy with the current classification to welcoming a radical overhaul. Those who were content mentioned that they had received no complaints from people responding to their question on ethnicity, although it should be noted that there was no evidence of data providers being actively asked to comment. Also, it was said that the small percentage of people from ethnic minorities meant that any refinement of categories was basically 'meaningless' for one respondent, as data needed to be aggregated to such a high level in Scotland.

3.19 Conversely, another respondent felt that it was timely for questions on ethnicity to be reviewed. This belief came from a suspicion that people were simply responding to existing categories, rather than reflecting on how they really felt. He suggested that identity was something that people often had not thought about, and if they had it was rarely with any great depth. It was felt that the Scottish Executive ought to take this reconsideration of the Census question as an opportunity to encourage people to think about their identity.

3.20 Generally though respondents were of the opinion that the current classification being used by their organisation was useful but could benefit from being improved or fine-tuned. This stemmed from a level of uncertainty about whether the current ethnicity categories reflected best practice, as it was not known what data providers thought of the options offered to them. Consequently, data users were interested in the findings from this research project. Below we describe the key points made in relation to how ethnicity classifications could be altered and/or improved.

White minority groups

3.21 A number of the data users we interviewed commented that there needed to be a much better understanding of white minority groups. They referred to a 'white mass' in the statistics that hid minorities with needs, and among them people who could be experiencing discrimination or prejudice. Examples were given of a Portuguese fishing community and people from Eastern Europe.

The use of 'Other' categories

3.22 Ethnicity classifications generally make allowance for the list of categories not being exhaustive by including an 'Other' box to tick. Similarly, on the 2001 Census question on ethnicity there are a number of 'other' boxes where people are offered the option of ticking and then writing in a description for themselves, such categories include:

• Any other White background, please write in
• Any other Asian background, please write in
• Any other Black background, please write in
• Any other background, please write in

3.23 Many comments were made in relation to how the 'Other' categories were handled and responded to by data providers. Data users were aware that some people avoided 'other' categories because they resented being classified as 'other'. They expressed concern that this led to such people being mis-classified.

3.24 There were mixed views on whether introducing a longer list of categories for people to choose from would be helpful or desirable. It was suggested that having a longer list might be confusing and one respondent remarked that there was little point as such a list would 'never be perfect'. Having a long list could also present problems for processing the data and one organisation reported that they had experienced problems when using a long list of options in the past. Staff entering the data was found to be pressing the wrong keys and mis-coding people. The errors were spotted and the anomalies were corrected during a subsequent audit to rectify the problem. This suggests that careful attention needs to be paid to how a long list is handled during the processing of ethnicity data to ensure that responses are correctly coded. Expanding on this point, it was noted that the quality of any statistics on ethnicity were also dependent on the skills of the people inputting the data, and the quality of IT provision when it came to designing user friendly computer programs that guarded against errors.

3.25 Even so, other respondents felt there would be value in having a wider choice of categories as some people could benefit from being counted differently and, it was hoped, this would enable people to provide more accurate responses. In addition, it was reported that some data providers had found exclusions upsetting (the example of Asian Caribbean was given).

Religion

3.26 A number of the data users wondered how religion should be handled and this led them to query the relationship between religion and ethnicity, the main examples being Judaism and Islam. There was a distinct lack of clarity over whether being Jewish, Muslim, or indeed of another religion, constituted being part of a distinct ethnic group.

Nationality and/or colour

3.27 Some respondents commented on the combination of nationality and colour often used in ethnicity classifications (including the 2001 Census question). It was remarked that some young people wanted to identify as primarily Scottish, regardless of ethnicity. Consequently, one respondent felt that ethnicity classification that did not make reference to nationality would be 'flawed'. Conversely though, another respondent considered it wrong for nationality to be used synonymously with colour, as in the 2001 Census. Although he did appreciate political concepts of colour and felt that the difference this had made in other countries like the United States needed to be acknowledged.

Flexibility

3.28 A number of points were made about ethnicity classifications in relation to flexibility. Firstly, it was felt that some classifications did not easily reveal a changing population. This was felt to be particularly pertinent to Scotland and the point made earlier about hidden white minority groups and changes in migration in recent years. One organisation used a classification that did evolve over time as staff was able to insert fields as new categories/self-descriptions arose. However, a curb had been placed on this as it had led to 'new categories' being added that were actually on the list; they were simply misspellings.

3.29 Secondly, it was noted that some people had a fluid approach to their ethnicity/identity. This might be dependent on the situation and/or who they are speaking to, but also changes over time. An example being changing nationality (such as receiving a British passport), as well as how long a person has been living in a country and feels more and more able to identify with that culture / way of life (both points that we take up later in this report). Some organisations recognised the fluid nature of identity and sought to update ethnicity data on a regular basis. However, it was also commented that a person's ethnicity did not change and consequently there was no need to request ethnicity information more than once.

3.30 Little comment was made in relation to self-definition (being asked to describe you own ethnicity). One respondent wondered whether asking people to describe their ethnicity left classifications open to misrepresentation but she did not suggest that the alternatives (like probing for details and then interviewers/officials deciding the most appropriate code) were any better.

The impact of changing to a new ethnicity classification

3.31 Data users were asked to consider what kind of impact any changes to their current ethnicity classification might have on their work and that of their organisations. This might be if they were asked to start using the Census question on ethnicity or because the 2001 Census question was itself changed in some way.

3.32 A number of concerns were raised by data users in relation to any changes to their current systems:

• Inhibiting trend analysis. One organisation expressed concern over having to go back to many thousands of data providers to update their records to allow for comparability over time (trend analysis). This was important to this particular organisation because they wanted to study how their population changed from year to year.
• Disruption. Another respondent was very worried about the disruption her organisation might experience; this was particularly pertinent to her as they had recently introduced an expensive computer system. Such worries could be alleviated so long as plenty of lead-in time was allowed for. Time was needed so that any changes could be introduced as part of normal review periods. Periods varied from a few weeks to a few years depending on how computer systems were organised and the ease with which they could be adapted (something that would need to be explored further with those responsible for implementing such changes). Allowances also needed to be made for training to be designed and implemented as appropriate and, more generally, communicating the need for change. In one case it was remarked that public bodies might need to enter into consultations of their own, for example with staff, the public and trade unions.

• Budget. Having a budget to manage any change was also highlighted by some respondents. Conversely though, one organisation mentioned that there were always unplanned for changes and they made contingencies in their budgeting to allow for this. This did not seem to be an option for some organisations working to tight spending plans.

3.33 Whilst respondents expressed some concerns, there was a general view that organisations would 'just get on with it'; that they were used to handling change, such as taking on board new policy initiatives, as part of their work. A change to their ethnicity classifications would be similar to any other change. It was commented that it would be helpful if the 'main headings' (for example: black, white, Asian, other) were maintained to aid trend analysis, but another respondent was keen to stress that the emphasis of change should be on what works for data providers rather than meeting administrative needs.

The ideal approach

3.34 As described in the previous section, data users had mixed views on changing the ethnicity classifications they were using. Whilst the sample of respondents was small, there was a preference for not changing too much, but equally a desire to 'get it right' too. As part of the interview, data users were asked to put aside issues of trend analysis and concerns over change in order to consider a blank sheet approach: what would the ideal ethnicity classification look like from their perspective. Respondents did not sketch out a particular framework; as we mentioned earlier they were generally content with the current situation. However, they did flag up points that they felt would be useful to consider, and for the most part these reflected issues they had already highlighted to the researchers. This section explores their suggestions.

Should there be one question or multiple questions?

3.35 Generally there was a desire for a tiered or nested approach. This could be similar to the 2001 census question in that there is one question with main categories/headings and sub-categories or could be a series of questions that filter to a sub-question or questions. A nested approach was preferred because it had the potential for easily aggregating the data to standardised categories and, again if the current main categories were maintained, provided for the possibility of comparing data historically.

3.36 One respondent did feel that there should be just one question for simplicity and speed, but others said that using more than one question would enable them to capture more detail, be more accurate and, they hoped, overcome the problem of imposing labels that people could find inadequate.

3.37 In terms of how a series of questions might work, it was suggested that nationality and ethnicity be separated out. One respondent also mentioned that additional questions on language and religion would be useful in any classification related to ethnicity, as these factors also had a bearing on some people's feelings about ethnicity (it should be noted that a question on religion was included in the 2001 Census).

3.38 It was also suggested that the Census ethnicity question could focus on 'main headings', leaving organisations to define the more detailed sub-categories that were best suited to the population they were working with. However, adopting this approach would present two issues. Firstly, not all organisations collect data on ethnicity directly from users; they use the Census as their main source of information to look at needs. This was not an issue for the respondents in this study but cannot be ruled out. Secondly, data could be misleading if there was no way of checking how data providers were identifying with the main headings, with the possibility that organisations are grouping people differently.

The list of categories - 'pre-codes'

3.39 Whilst there were still concerns over no longer being able to compare data over time, should the Census ethnicity question change in 2011, data users were keen for any new approach to better identify changes in the Scottish population. For some this meant having a much longer list of pre-codes (ethnic groups/categories) to choose from, which it was hoped would allow people to be more accurate in their responses. Even so, they wanted the data to still be easily aggregated, confidentiality and anonymity to be protected, and careful testing/piloting to ensure that having a long list did not put people off answering the question.

3.40 Another aspect of having a better understanding of a changing population would be to look more closely at how the 'Other' categories are processed and analysed. Of course, this still might present problems in terms of publishing sensitive data that could identify individuals, but there may be scope for GROS to make public how they go about handling the data without compromising on confidentiality.

Ideas for ethnicity categories

3.41 Apart from the overall approach to collecting data on ethnicity, data user respondents also commented on the specific categories that should or should not be used. Their suggestions are described below:

• Ensuring that all the British nationalities are listed, including 'Northern Ireland'. Adding in 'British' and 'Republic of Ireland' as separate categories.
• It was felt that just having 'Scottish' and 'British' was inappropriate: 'screams inequality'
• It was also suggestion that 'White European' should be listed as a separate code as some people saw themselves as European.
• Having 'Chinese' as a separate category from 'Asian'.
• Adding in new categories, for example, 'Travellers'. Although there was some debate over how some groups should be presented / defined and to what extent any further distinctions should be made, for example by asking people who identify as 'Travellers' whether they are people of Roma, Scottish, Irish descent and so on.

3.42 Generally, though, data user respondents were unsure of making adjustments to the list of categories and felt that this was something that research among data providers should determine.

« Previous | Contents | Next »