Pooled Sample from Population Surveys in Scotland
The Scottish Government, in consultation with stakeholders, developed the Long Term Strategy for Population Surveys in Scotland 2009-19, with the aim of taking a more strategic view of the substantial analytical resource that our surveys provide. One of the key aims in recent years has been a review of the effectiveness and efficiency of the surveys. This has included work to align the methodologies deployed in our large-scale surveys, and harmonise questions where possible – with a view to obtaining a pooled sample of core variables spanning all three surveys.
The three large-scale cross sectional population surveys currently run by the Scottish Government and each published separately are:
- Scottish Crime and Justice Survey (SCJS)
- Scottish Health Survey (SHeS)
- Scottish Household Survey (SHS)
Since the beginning of 2012 each of the surveys has included a set of 20 core questions – the topics covered are shown below. Responses on these questions from all three surveys have been pooled to provide a dataset with a sample size in excess of 20,000 responses.
This pooled sample allows for :
- Increased precision of estimates at a national level
- More accurate estimates of the numbers of people in small population groups such as specific ethnic groups or religions.
- Significant increases in the precision of estimates at Local Authority level.
This is the first year of dissemination of the pooled data and provides only a snapshot of the Scottish population. With subsequent publications it will be possible to develop time series to examine the change in variables.
Although the three large-scale surveys are all designated as National Statistics, the pooled sample is currently being released under the classification “Data Being Developed.” This is because, while we want to encourage analysis, exploration and feedback by stakeholders, care should be taken when analysing these data. While every effort has been made to harmonise the core questions across the surveys, different context or ordering effects could lead to inconsistencies in responses across the different surveys which may skew the results. In light of the above, and despite the larger sample size, the pooled data being release should not be considered as the primary source of data for the variables it contains. This will be reviewed as the pooled dataset is further tested and developed. The tables presented show the preferred sources for some of the core question topics, including where that is out-with the three surveys that give rise to the pooled sample (e.g. the Labour Force Survey).
The Scottish Government is keen that the pooled core is exploited as much as possible and the pooled sample is being released now to allow comment and testing from users, who are invited to engage with us to improve the quality of the data, our knowledge of the data and to influence future developments of pooled data. In addition to the illustrative tables presented here, the dataset will be made available subject to controls being in place.
The core questions which are included in all three surveys cover:
- Household relationships
- Age and date of birth
- Country of birth
- Ethnic group
- Sexual orientation
- Self perception of health
- Housing Tenure
- Economic activity
- Banded household income
- Qualifications held
- Car access
- Marital status
- Disability or long-term health
- Mental well-being
- Perception of local crime rate
- Perception of police performance
Full details of the questions are available here:
Further information on Population Surveys in Scotland can be found here:
In producing the pooled sample dataset a number of issues have been identified that affect the overall consistency of some of the specific variables, and these are discussed below. Further work is underway to explore such issues in more depth.
The estimate of the number of carers varies substantially between the three surveys. The Scottish Health Survey is currently the preferred source for estimating unpaid caring prevalence.
The differences are likely to be caused by a combination of factors. By comparing responses between the surveys, the most significant of these factors is thought to be the effect of proxy responses (i.e. where one member in the household answers the question on behalf of others). In the SHeS and SCJS, the results are based on people answering the question directly whilst for the SHS the highest earner in the household answers the question for everyone in the household. Hence, one third of the SHS adults estimate are proxy responses. This results in an expected underestimate of caring prevalence in the pooled sample currently.
Other effects include slight variations in question wording, specifically the inclusion of a discretionary interviewer prompt in SHeS and SCJS asking that paid caring is included, in contrast to wording in SHS which explicitly asks that paid caring is excluded from all responses.
Given these effects and the variation in results for direct and indirect question responses, the carers core question has been changed from April 2014 so that the estimate will be based entirely on direct responses, and to align question wording exactly across all three surveys.
The data in Table 9 includes responses from all three surveys.
b) Long Term Conditions
The results from the question on Long Term Conditions is affected by the fact that for the first 8 months of the 2012 SHS a different version of the question was used. The earlier responses have been excluded from the pooled dataset and this explains the smaller base in Table 11. SHeS estimates a higher proportion of adults with a long-term condition, which may be explained by an effect whereby participants are more likely to identify some health conditions in the context of other questions relating to their health.
c) Mental Wellbeing
The distribution of scores for the Mental Wellbeing questions in SHS includes no ‘refusal’ or ’don’t know’ responses. The questionnaire script from 2014 has been amended to allow these responses. Other slight differences in mental wellbeing scores between the surveys may be related to variations in questions format, self-completion in SHeS as opposed to face-to-face interviews in SHS and SCJS.
d) Household Income
The data for household income is affected by two issues. Firstly a high refusal rate – the number of respondents refusing to answer this question lowers the base for Table 14 from around 23,300 to around 20,100. Secondly, deductions - the SHS question for income asks respondents to give their levels of pay after deductions for tax and national insurance but the core question (and that used in the SHeS) is before deductions. Hence a set of assumptions have been used to calculate gross household for respondents from all three surveys.
e) Confidence in Policing
The results on confidence in policing in tables 19-24 in this publication have been calculated and presented differently to the equivalent results in the Scottish Crime and Justice Survey (SCJS). The main difference is that the pooled sample dataset does not include ‘don’t know’ responses for this question while the SCJS analysis does. Comparative analysis on the underlying SCJS data on the same basis as used in this publication (i.e. not including SCJS responses of ‘don’t know’) provides results that are broadly consistent with those derived from the pooled sample.
The question on self-identified sexual orientation was introduced to provide statistics to underpin the equality monitoring responsibilities of public sector organisations and to assess the disadvantage or relative discrimination experienced by the lesbian, gay and bisexual population. Despite this positive step in collecting such information, it is felt that the figures are likely to under-report the percentage of lesbian, gay or bisexual (LGB) people within society due to a number of reasons, including the following:
- Asking about sexual orientation/identity is a new development in national surveys and such questions can be seen as intrusive and personal.
- There is still significant prejudice and discrimination against LGB people in society. In a context where some LGB people will not have told friends and family about their sexual identity, there is a real question about whether LGB people generally would want to be open with an interviewer.
- The default option for being uncertain about one's sexual orientation may be to respond 'straight/heterosexual' rather than to say 'Don't know / not sure'.
- Particular LGB people are still less likely to be open where they belong to groups or communities where an LGB identity is less acceptable.
All of the responses have weighted to make them representative and the weighting for the pooled data set is based upon that used for the three individual surveys.
The unweighted base in each table shows the number of respondents for each row whilst the weighted base shows the effective number of respondents used for calculating Scotland level estimates.
Other than the issues identified above, the base varies for each table mainly due to differences in individual response rate for each question . Respondents refusing to answer questions or answering don’t know are excluded from all tables except for Tables 6 and 18.
The underlying pooled dataset is available for further analysis subject to consideration of legal and ethical factors – contact details can be found at the end of this release. A number of the tables in this release have had categories collapsed and data aggregated to a higher level due to existence of very small figures. This is to allow further time for the Scottish Government to consider disclosure control options for this data.
The data presented below mostly shows local authority level data but it is possible to analyse all of the variables by the following geographies:
- Health Boards
- 2012 SIMD Quintiles
- 2012 SIMD 15% most deprived indicator
- 6 category urban-rural indicator
The potential to provide sub-local authority estimates was considered. However, it was decided that the current sample size was insufficient for the next level down, which would have been Scottish Parliamentary Constituencies. Over time, it is intended that with subsequent releases of pooled sample data a larger dataset will allow the release of data at lower geographies.
List of Tables
The tables are availiable to download as an excel file.
Local Authority Tables
Characteristics of Adults in Scotland
Table 1: Characteristics of Adults in Scotland - Age
Table 2: Characteristics of Adults in Scotland - Country of Birth
Table 3: Characteristics of Adults in Scotland - Ethnicity
Table 4: Characteristics of Adults in Scotland - Marital Status
Table 5: Characteristics of Adults in Scotland - Religion
Table 6: Characteristics of Adults in Scotland - Sexual Orientation
Table 7: Characteristics of Adults in Scotland - Highest Qualification Held
Table 8: Characteristics of Adults in Scotland - Employment
Table 9: Characteristics of Adults in Scotland - Providing Regular Care
Health of Adults in Scotland
Table 10: Health of Adults in Scotland - Smoking
Table 11: Health of Adults in Scotland - Long Term Condition
Table 12: Health of Adults in Scotland - General Health
Table 13: Health of Adults in Scotland - Mental Wellbeing (Average Score in Short Warwick-Edinburgh Mental Wellbeing Scale)
Characteristics of Households in Scotland
Table 14: Characteristics of Households in Scotland - Gross Household Income
Table 15: Characteristics of Households in Scotland - Household Composition
Table 16: Characteristics of Households in Scotland - Housing Tenure
Table 17: Characteristics of Households in Scotland - Car Access
Crime in Scotland
Table 18: Crime in Scotland - Perception of the Change in Local Crime Rate in Past Two Years
Table 19: Crime in Scotland - Confidence in the Police to Prevent Crime
Table 20: Crime in Scotland - Confidence in the Police to Respond Quickly to Appropriate Calls and Information from the Public
Table 21: Crime in Scotland - Confidence in the Police to Deal With Incidents as They Occur
Table 22: Crime in Scotland - Confidence in the Police to Investigate Incidents After They Occur
Table 23: Crime in Scotland - Confidence in the Police to Solve Crimes
Table 24: Crime in Scotland - Confidence in the Police to Catch Criminals
Table 25: Employment by Highest Qualification Held
Table 26: Highest Qualification Held by Age
Table 27: Religion by SIMD
Table 28: Place of Birth by SIMD
Table 29: Perception of Change in Crime by Urban Rural Classification
Table 30: Household Income by Urban Rural Classification
DATA BEING DEVELOPED
The statistics contained in this release are new official statistics that are undergoing evaluation and are published in order to involve users and stakeholders in their development as a means to build in quality at an early stage.
For enquiries about this publication please contact:
Office of the Chief Statistician,
Telephone: 0131 2443992,
For general enquiries about Scottish Government statistics please contact:
Office of the Chief Statistician, Telephone: 0131 244 0442,
Complaints and suggestions
If you are not satisfied with our service or have any comments or suggestions, please write to the Chief Statistician, 3WR, St Andrews House, Edinburgh, EH1 3DG, Telephone: (0131) 244 0302, e-mail firstname.lastname@example.org.
If you would like to be consulted about statistical collections or receive notification of publications, please register your interest at www.scotland.gov.uk/scotstat
Details of forthcoming publications can be found at www.scotland.gov.uk/statistics
You may use or re-use this information (not including logos) free of charge in any format or medium, under the terms of the Open Government Licence. See: www.nationalarchives.gov.uk/doc/open-government-licence/