The research comprised two consecutive strands:
1) a mixed-mode survey of employers from across the social care sector
2) follow up qualitative interviews among a selection of employers who took part in the survey, as well as a small number of workers
Survey of employers
As described above, the over-arching objective of the survey was to provide robust information on the number of non- UK EU workers in specific social care sub-sectors (adult day care, care at home, day care of children, childminding, housing support, and nurse agencies).
The Scottish Government provided Ipsos MORI with details of services from the Care Inspectorate database of providers. The sample frame included the basic public-domain details of the in-scope service providers such as email addresses and telephone numbers. It also included information on employer type (public, private, voluntary) and details on registered capacity and number of staff, all of which were used for setting quotas for the telephone fieldwork. Services in the sample frame were categorised to match the categories used in the Scottish Social Services Council workforce data 2015  in order to set survey targets ( Table 2.1).
Table 2.1: Typology of specified care sectors
|Adult Day Care||Support service subtype other than care at home|
|Care Home for Adults||Care Home for Adults|
|Day Care of Children||Childcare agency, Day Care of Children Service|
|Housing support/Care at home||Adult Placement Service, Housing Support Service, Support Service subtype (care at home)|
|Nurse Agency||Nurse agency|
Further cleaning and editing of the sample information was carried out, with duplicate entries (as identified through email addresses and phone numbers) removed. Survey recipients were invited to respond in relation to a specific named service only.
The following variables were created:
- Employer type (public, private, voluntary).
- Nuts2 region (Eastern Scotland, South Western Scotland, North Eastern Scotland, Highlands and Islands). This was derived based on local authority.
- Threefold urban/rural classification (Urban, Accessible Rural, Remote Rural). This was matched into the data using postcodes of services. In some cases, there was missing information on postcodes which meant not all services could be classified and there were limitations on the associated analyses possible.
The sampling strategy balanced the need for precision across all sectors with the ability to analyse results by sub-sector and subgroup variables such as employer type and size of employer. Targets were set based on trying to ensure confidence intervals of no more than +/-3% at the sub-sector level for a finding of 10% (with a 95% confidence level). Population figures were taken from the latest available Social Services Workforce Data Report 2015  . Given the small number of organisations in two sub-sectors – Nurse Agencies and Adult Day Care – it was recognised that this would equate to a very high response rate.
Table 2.2: Population information by sub-sector and ideal sampling targets
|Sub-sector||Staff ( SSSC 2015)||Services ( SSSC 2015)||Average staff per service||Ideal number of interviews|
|Adult Day Care||8,080||492||15.8||218|
|Care Homes for Adults||53,980||1,149||46.7||290|
|Day Care of Children||33,460||3,710||9.0||348|
|Housing Support/ Care at Home||69,690||2,041||33.8||322|
The survey used a mixed mode approach to maximise the number of returns and the precision of the estimates among each sub-sector. Organisations were first invited to participate online, with a telephone survey following afterwards.
The online survey was sent to 50% of all services. Reflecting the sampling approach, a higher proportion of services were sampled in the sub-sectors with the most challenging response rate targets, namely Adult Day Care, Care Homes for Adults, and Nurse Agency. All services with available contact details were included in the issued sample for these sub-sectors in order to maximise the number of surveys returned. The survey was issued to comparatively less of the total available sample for the remaining sectors e.g. 26% of childminders were issued the survey.
The scripts for the online and telephone survey were developed by Ipsos MORI’s scripting team in IBM Dimensions. The telephone fieldwork was run from Ipsos MORI’s Edinburgh Telephone Centre, using Computer Assisted Telephone Interviewing ( CATI). Fieldwork was targeted to try to achieve the target number of interviews by sub-sector overall and two additional quotas were set to help ensure the achieved sample was representative.
- average number of staff per organisation (above and below the median values) for all sub-sectors bar childminding
- employer type (public, private and voluntary) for the three largest sub-sectors, namely Housing Support, Day Care of Children, and Care Homes
In all, the survey was open for a period of just over five weeks, from 10 January to 16 February 2018. An e-mail address and phone number for the survey was in operation throughout, so that the research team could answer queries and resolve any technical problems.
The survey questionnaire was developed by Ipsos MORI in consultation with the Scottish Government. It was deliberately kept as short as possible to help encourage participation among the target audience and thus help ensure the robustness of the results.
The core of the questionnaire comprised a series of pre-coded items to collect the total number of staff currently working in the services sampled – including, separately, the numbers of auxiliary staff, care staff/practitioners, NMC registered nurses and managers – and the proportions in each case who were non- UK EU nationals.
In addition, the questionnaire included a small number of questions to collect information on:
- the extent to which employers had found it easier or more difficult to recruit and retain workers over the preceding 12 months
- the extent to which employers had found it easier or more difficult to recruit and retain non- UK EU workers specifically over the preceding 12 months
- whether any non- UK EU workers had left the service in the last 12 months and the reasons for this
The questionnaire was piloted among a small sample of employers from across the care sector to ensure it was clear and easily understood. In light of the pilot findings, minor revisions were made to the questionnaire in advance of the main stage fieldwork commencing.
Response rate and achieved sample profile
The survey achieved 1,572  responses from services, covering 39,999 staff. Table 2.3 shows the profile of the issued over the achieved sample. The population figures are based on the 2016 SSSC data – the latest available at the time of analysis.
Table 2.3: Issued versus achieved sample profile
|Population||Issued sample||Achieved sample (all)||Response rate|
|N||N||% of issued sample||Number of responses||% of overall response||%|
|Adult Day Care||492||333||5%||117||7%||35%|
|Care Home for Adults||1,149||1,004||16%||305||19%||30%|
|Day Care of Children||3,710||1,472||23%||360||23%||24%|
|Housing Support/Care at Home||2,041||1,276||20%||449||29%||35%|
|Highlands and Islands||1,517||784||12%||165||10%||21%|
|Urban-rural threefold classification|
|Rest of Scotland||10,618||3,379||53%||1,010||64%||30%|
The achieved sample was close to the issued sample for all of the key variables covered. The proportion of childminders covered within the survey is smaller due to the disproportionate sampling of sectors and services with larger number of employees (childminders have an average staff size of 1).
Data cleaning and analysis
Given the number of logic checks that were incorporated into the online and telephone scripts, the data cleaning required was minimal. A small number of cases were dropped from the final dataset due the high proportion of missing data. Outlier values were manually checked and key analysis variables were derived.
Total staff for each service was derived from the counts given at the breakdown of staff within each category (auxiliary, care staff, nurses, managers, other staff). This was checked against the total staff figure given within the survey. Similarly, the total number of non- UK EU staff was calculated in the same way. Weights were created to correct for disproportionate sampling by sector during the analysis.
The analysis approach was straightforward. The percentage of staff who were non- UK EU staff was calculated from the survey data. With the exception of staff from nurse agencies, the population figures from the Social Services Council Workforce Data Report 2016  were used to create the estimate of the total number of staff.
The differing approach for nurse agencies was due to a discrepancy between the two datasets, with the 28 nurse agency services surveyed reporting a total headcount of 2,718 – higher than the 2,400 headcount reported for 71 services in the SSSC data. The discrepancy reflects differences in the questions asked in the two studies. In the survey, respondents were asked to include all workers, whereas the Care Inspectorate analysis collects data on nurses only  , as registration is only with respect to agency nurses. To address this, the average number of staff in the nurse agencies surveyed was multiplied by the number of services in the SSSC data to produce the total number of nurse agency staff.
The estimates presented in this report include lower and upper confidence intervals.
The objective of the qualitative research was to complement the survey findings with insights into the role and contribution of non- UK EU workers in the social care sector. More specifically, the work sought to provide a deeper understanding of employers’ and workers’ circumstances, experiences and perspectives, including any specific challenges they were facing, or concerns they had, in light of the UK’s decision to leave the EU.
Sampling and recruitment
Prospective employees for interview were identified from among a pool who had taken part in the survey and agreed at that stage to be recontacted for follow up research. Further, they were purposively selected based on their working in sub-sectors that the survey (and wider research) had indicated were particularly reliant on non- UK EU workers; namely, nurse agencies, care homes for adults and children’s day care services.
A total of 20 employers were interviewed from across these sub-sectors; 2 from nurse agencies, 10 from care homes and 8 from children’s day care services. The achieved sample also represented a mixture of service types (local authority, third sector and private sector), size and locations, as set out in table 2.4.
Table 2.4: Profile of achieved qualitative sample of employers
|Nurse agency||Care home||Children’s day care||Total|
|Highlands and Islands||2||1||0||3|
The employers interviewed were invited to identify a non- UK EU worker from among their staff who might be willing to participant in a similar interview. In the event, a total of 10 workers were interviewed, with the achieved sample again representing a mixture of sub-sectors, service types, locations and sizes.
All of the qualitative interviews were conducted by telephone with a view to minimising any burden on participants. The workers were offered a £15 high street voucher to encourage their participation and as a ‘thank you’ for their time. Fieldwork took place between 19 March and 25 April 2018.
The qualitative interviews were digitally recorded with participants’ permission and in most cases transcribed for analysis purposes. The transcripts (along with raw interview notes) were then systematically analysed to identify the substantive themes which emerged in relation to each question in the discussion guide, along with key points and illustrative verbatim comments. This ensured that the analysis of the data was rigorous, balanced and accurate, and that key messages or concepts were brought out. It was also flexible enough to allow links and connections across different themes or sub-themes to be made, and for moments of interpretive insight and inspiration to be recorded.
Interpreting the data
Survey findings represent the views of a sample of the population concerned, and not the entire population, so they are subject to sampling tolerances, meaning that not all differences will be statistically significant. Throughout the report, differences between sub-groups are commented upon only where these are statistically significant, i.e. where we can be 95% certain that they have not occurred by chance.
Where percentages do not sum to 100%, this may be due to computer rounding, the exclusion of ‘don’t know’ categories or multiple answers. Throughout the report, an asterisk (*) denotes any value of less than half a percent and a dash (-) denotes zero.
Unlike survey research, qualitative social research does not aim to produce a quantifiable or generalisable summary of population attitudes, but to identify and explore the different issues and themes relating to the subject being researched. The assumption is that issues and themes affecting participants are a reflection of issues and themes in the wider population concerned. Although the extent to which they apply to the wider population, or specific sub-groups, cannot be quantified, the value of qualitative research is in identifying the range of different issues involved and the way in which these impact on people.