Public sector - understanding equality data collection: main report

3 Research methods and approach

3.1 This section describes the research design and data collection methods used in the study. The section also covers (i) the limitations of the research, and (ii) the terminology adopted in the report. First, however, the approach taken to define the scope of the project is described.

Defining the scale and scope of the project

3.2 There are over 150 public sector bodies in Scotland involved in various capacities in the design and delivery of public services. In order to make the project manageable and to maximise its value, it was agreed that the research should:

• Be based on (a subset of) the data collections undertaken by around 30 public sector bodies
• Focus disproportionately on public sector bodies whose remit is associated with addressing structural inequality (specifically those involved in Health and Social Care; Education; Justice; and Employment, Labour and Welfare)
• Focus on data collections which are 'public facing' and which support policy or service development and delivery
• Exclude consideration of equality and socio-economic disadvantage data collected in relation to the workforce in the selected public sector bodies
• Exclude consideration of UK-wide data sources (HMRC, DWP, etc.).

Research design

3.3 The research involved exploring the collection of equality data within a sample of public sector organisations and networks, gathering descriptive information about a range of identified data collections, and exploring the views and experiences of individuals working with selected data collections in a variety of capacities.

3.4 The research design employed a three-stage stratified approach to sampling. Whilst this design did not provide formal statistical representativeness, it allowed coverage of a range of (i) types of organisation (non-departmental public bodies, health bodies, executive agencies, etc.), and (ii) types of data collection (administrative, survey). It also covered all (nine) protected characteristics as well as socio-economic disadvantage to some degree.

3.5 The three stages of the sampling approach are described in further detail below, followed by a discussion of the approach to 'substitutions' where the original selection was not viable.

Stage 1 - A sample of organisations

3.6 Approximately 170 Scottish public sector bodies were identified as 'in scope' for this project.^[24] These bodies were organised into six groups (or 'strata'). Four of these groups reflected to specific topic or policy areas, namely: (i) Health and Social Care (ii) Employment, Labour and Welfare (iii) Education and Skills, and (iv) Justice. The fifth group was reserved for local authorities, whose work spans multiple policy areas. The final group contained all 'other' in-scope organisations.

3.7 Twenty-eight (28) organisations were initially selected for inclusion in the research as follows:

• 5 Health and Social Care organisations
• 3 Employment, Labour and Welfare organisations
• 6 Education and Skills organisations
• 5 Justice Organisations
• 5 Local Authorities (including a mix of large and small, and urban and rural)
• 4 'Other' Organisations.

Stage 2 - A sample of departments within selected local authorities / NHS Boards

3.8 A second-stage selection process was implemented in relation to the local authorities (LAs) and NHS Boards.^[25] The original intention was to select two departments in each local authority and NHS Board - although it was not possible to achieve this in every case (see paragraph 3.11 below).

Stage 3 - A sample of data collections

3.9 At the final stage, relevant data collections in each of the selected organisations or departments were identified. Between one (1) and four (4) data collections from any particular organisation or department were included in the research. In cases where there was more than one relevant data collection, the selection was guided by (i) a wish to focus on services provided to individuals (rather than organisations), and (ii) a wish to explore datasets of different kinds (e.g. administrative datasets, surveys, 'one off' exercises, etc.).

Substitutions

3.10 If it was not possible to identify - or confirm the participation of - relevant individuals within the selected organisation within an agreed timeframe, a substitution was made based on the stratified design. (For example, if it was not possible to identify suitable participant(s) within a selected health and social care organisation then a substitute organisation from within the health and social care grouping was selected as a replacement.) Substitutions were also made in cases where, following initial research, the original selection was deemed to be 'out of scope'.^[26]

Achieved sample

3.11 The final achieved sample comprised 27 organisations, including five (5) local authorities and one (1) NHS Board.^[27] As explained above (paragraph 3.8) there was an intention to select two departments in each of these organisations. However, this did not prove to be possible in every case. Overall, eight (8) departments (or service areas) were included (comprising seven (7) local authority departments and one (1) NHS Board department).^[28]

3.12 Annex 1 lists the 27 organisations that participated in the research.

3.13 Annex 2 lists the 55 datasets which were included in the research.

Data collection

3.14 The data collection for the research comprised four main elements:

• Desk-based review activities including:
  • Exploring the websites of selected organisations
  • Reviewing documents and publications for relevant material about the organisation's approach to the collection and use of equality and socio-economic disadvantage data
  • Identifying relevant individuals / departments for follow up.
• Collection of descriptive information on selected data collections. For each data collection considered, this involved:
  • Describing the main features of the data collection (type of data collection, frequency of collection, population coverage, geographic coverage, publication arrangements, methods of data collection, etc.) (See Annex 3 - Template 1)
  • Confirming which - if any - protected characteristics and socio-economic disadvantage data are collected (See Annex 3 - Template 1)
  • Recording (i) the questions used to elicit information about protected characteristics and socio-economic disadvantage, and any relevant definitions, (ii) the categories offered for responses, and (iii) the response (or completion) rates for specific items (if available). (See Annex 4 - Template 2).
• Qualitative interviews with staff in selected public bodies which explored (see Annex 5):
  • Reasons for collecting (or not collecting) specific equality and socio-economic disadvantage data
  • How data are used, and whether data are used for the stated purposes
  • Approaches to publication
  • Challenges encountered in collecting equality and socio-economic disadvantage data
  • Issues related to quality and completeness of the data
  • Improvement work and shared learning with regard to the collection of equality and socio-economic disadvantage data
  • Future considerations, including the potential for greater legislative requirements.
• Discussion with the Non-Departmental Public Bodies (NDPB) Equality Forum, and the Scottish Council Equality Network (SCEN) about all aspects of the collection and use of equality and socio-economic data.^[29]

Limitations of the research

3.15 As set out above, the research is based on (i) a purposive sample of 27 public sector organisations, and (ii) discussion with two public sector equality networks. The research participants were employed in a wide variety of roles and included senior managers, policy specialists, analysts, information managers, equality officers, statisticians, researchers and administrators. The issues discussed were complex and wide-ranging - and were often quite specific to the organisation concerned. Thus, the findings reported here should be regarded as exploratory and indicative, rather than definitive or comprehensive.

A note on terminology

3.16 From this point on in the report, the abbreviation 'equality data' is used to refer to data concerned with both protected characteristics and socio-economic disadvantage. Where only one of these elements (either protected characteristics or socio-economic disadvantage) is specifically being discussed, this is made clear in the text.

3.17 The term 'customer' is used to refer to the wide range of service users, clients, patients, grant recipients, training participants, applicants, students, pupils, members of the public, etc. who are asked by public sector organisations to provide information about their equality characteristics.