Scotland's New Ethnicity Classification for Scottish Official Statistics and Recommended for Scotland's 2011 Census

3. The Work Programme and Evidence Base

3.1 The review team undertook an extensive work programme to develop and test the new classification. The key aim was to ensure that the classification (and decisions underpinning it) was based on evidence. In broad terms, the work programme consisted of consultation and research (including question testing). The classification was developed and refined over time by repeated cycles of evidence gathering and assessment.

3.2 Similar work programmes have been conducted by ONS (in partnership with WAG) and these programmes were conducted in a 'joined-up' way, where possible, so that UK classifications could be developed in a broadly consistent way. The review team used some of the evidence collected by ONS and WAG (and vice versa) to help them make decisions, but as stated previously, these findings cannot be reported here. All evidence is listed in the table on page 20.

How was the work programme developed?

3.3 Scotland's work programme was developed with the following pre-requisites in mind:

• It was necessary to find out what information is needed by those who collect and use ethnicity statistics, in order that the classification meets these needs and is applicable in practice.
• It was important to address concerns about the existing classification, in particular to find out what terminology is acceptable to different ethnic groups and to investigate inconsistencies between categories.
• Methodological investigation and question testing were crucial for developing a robust classification that is clearly understood by respondents, works as intended and produces high quality statistics.
• Given that the census covers all ethnic groups and communities in Scotland, it was necessary to gather a wide range of views on ethnicity and to balance these views carefully, whether or not they were expressed by many or few people.

What evidence was collected?

3.4 Consultation was conducted with data users, stakeholders, representatives from different ethnic groups and other interested groups at different stages in the process (see the table on page 20 detailing the key consultations). Consultation involved written responses to consultation papers and a range of stakeholders meetings. This was supplemented by desk research and ongoing liaison with stakeholder organisations.

3.5 Qualitative research was conducted on ethnic identity using in-depth interviews and group discussions with stakeholders and community groups, data users and members of the public. Ethnicity is a complex issue and is therefore suited to qualitative research methods which are used to explore social issues and provide in-depth information. Further qualitative research was conducted using cognitive question testing, which is a widely used approach to test proposed survey questions with target audiences. This research was conducted with members of the public from a range of different ethnic groups. A benefit of this technique is that it mimics, to some degree, the situation whereby a census form is delivered to a person's home and they then complete it with minimal, if any, assistance. During cognitive testing, small changes are made to the questions tested as results come in and issues or problems are identified. The questions shown in this report represent the very final question versions tested.

3.6 Some larger scale quantitative testing was conducted in Scotland's 2006 census test which covered approximately 50,000 Scottish households. In addition fieldwork was undertaken with Glasgow City Council, The City of Edinburgh Council and NHSScotland.

3.7 Key findings from the evidence collected by SG and GROS is summarised in detail at Annex C. Links to published findings, where available, are also provided here for those seeking more detailed information. The full findings of the other UK census offices will not be available until their classifications are finalised and published towards the end of the year.

How was the evidence assessed?

3.8 The evidence was assessed according to three core criteria for the design of the classification:

• It must be statistically robust (based on sound methodological principles) and produce high quality statistics on ethnicity,
• it should meet the key information needs of data users who collect ethnicity statistics; and
• it should be broadly acceptable and meaningful to those people who answer it.

3.9 The challenge for the review team was to assess the whole evidence-base in a balanced way, whilst keeping each of these criteria in mind. The evidence also had to be assessed with due regard to a number of constraints which precluded certain options or had a significant bearing on the viability of certain options. These are explained in section 4 from page 21 onwards.

3.10 As part of the drive towards UK harmonisation, the SG and UK census offices developed and applied a set of UK Guiding Principles to ensure that evidence was assessed in a broadly consistent way across the UK and to facilitate robust classification design. Detailed principles are set out at Annex B and include five overarching themes:

• Strength of need for information on a specific ethnic group.
• Lack of alternative sources of ethnicity information.
• Clarity and quality of the information collected and acceptability to respondents.
• Comparability with ethnicity data derived from the 2001 Census.
• Operational considerations such as: length of question(s), speed and cost-effectiveness of data processing and ability to collect comparable information in other surveys.

3.11 SG and GROS used these principles as a general guide to help assess its evidence. ONS and WAG are applying these principles using a tool, to score and prioritise which additional ethnic groups should be included in their classifications. Like SG and GROS, they also took other factors into consideration.

Notes:
1.ONS & WAG findings will not be available until their classification is finalised and published towards the end of the year. Therefore they cannot be provided in this report.
2. Cognitive question testing is a research technique used by survey designers to trial questions with members of the public to see if they are meaningful to respondents and function as intended.