Scotland's People Annual Report: Results from 2011 Scottish Household Survey

1 Background to the survey

Introduction

The Scottish Household Survey (SHS) is a continuous survey based on a sample of the general population in private residences in Scotland. The survey started in 1999 and, since then, has been carried out by a team from Ipsos MORI and TNS-BMRB (formerly TNS System Three).

The SHS is designed to provide reliable and up-to-date information on the composition, characteristics, attitudes and behaviour of Scottish households and individuals, both nationally and at a sub-national level. It covers a wide range of topics to allow links to be made between different policy areas. The specific aims of the survey are:

• To provide household and individual information particularly to support the work of the Scottish Government's transport, communities and local government policy areas and the work of the Scottish Parliament;
• To permit disaggregation of information both geographically and in terms of population sub-groups (such as families with children or the elderly);
• To allow the relationships between social variables within households to be examined. This will support cross-analysis on a range of issues;
• To allow early detection of national trends;
• To allow detailed follow-up surveys of sub-samples from the main survey sample, if required.

The Annual Report

SHS results have been reported in a series of Annual Reports between 1999 and 2010. The annual report is designed to act as an introduction to the survey and to present and interpret some of the key policy-relevant results.

A comprehensive range of web tables is provided on the SHS website.^[1] A series of SHS 2009/2010 Local Authority Tables were published in October 2011, as part of the Scottish Government's ambition to extend the use of SHS data. Similar information is also being published through an experimental SHS Interactive Mapping tool which presents time-series analysis from the survey for key measures across all local authorities in Scotland^[2]. The user is able to select the measure of interest including ones such as neighbourhood rating, car availability and provision of unpaid care. All relevant analysis for the measure selected is displayed as a time-series in an accompanying chart along with thematic map showing differences in estimates between all local authorities. Relevant metadata is also built in to the system to offer contextual information, as well as additional notes on the questions or variables used in the survey.

Whilst this latest release only gives a single year data period, some analysis may be able to be re-produced for certain local authorities. Please contact the SHS Project Team with any enquiries.^[3]

Structure of the Annual Report

At the start of each chapter introductory paragraphs draw on key policy documents to set the results that follow into the policy context for the topic it covers. In most of the chapters, the introduction draws on the Scottish Budget Spending Review 2007.^[4] This document highlights the current Government's overall purpose of increasing sustainable economic growth, and five strategic objectives, which are designed to ensure the purpose is delivered - as defined by the National Performance Framework. The framework was updated in December 2011. The objectives that are most relevant to the subject of a chapter, as defined in the spending review, are identified. The five objectives are:

Wealthier and Fairer - Enable businesses and people to increase their wealth and more people to share fairly in that wealth.
Smarter - Expand opportunities for Scots to succeed from nurture through to life long learning ensuring higher and more widely shared achievements.
Healthier - Help people to sustain and improve their health, especially in disadvantaged communities, ensuring better, local and faster access to health care.
Safer and Stronger - Help local communities to flourish, becoming stronger, safer place to live, offering improved opportunities and a better quality of life.
Greener - Improve Scotland's natural and built environment and the sustainable use and enjoyment of it.

Additional policy documents, including more detailed strategies on particular policy areas, are drawn on as appropriate and are referenced in the text.

The SHS is the source of information on eight of the 50 national indicators in the Government's National Performance Framework.^[5] The two transport indicators^[6] will be reported on separately by Transport Scotland within their Transport and Travel in Scotland 2010 report^[7] and in the SHS Travel Diary Results publication^[8]. This report provides estimates for the remaining six national indicators:

Improve people's perceptions of their neighbourhood (Chapter 4);
Widen use of the Internet (Chapter 9);
Reduce the percentage of the adult population who smoke (Chapter 10);
Improve people's perceptions of the quality of public services (Chapter 11);
Improve the responsiveness of public services (Chapter 11);
Increase cultural engagement (Chapter 13).

In some cases the SHS is not the official source of statistics on a particular topic: such as income, employment or housing. The interview collects information on these topics to select the data of particular groups for further analysis or for use as background variables when analysing other topics. The results are included in order to set the context for, and aid interpretation of, the remaining chapters. Where results are not the official source, this is indicated in the chapter introduction.

The Scottish Government conducts several major population surveys that are used to inform the policy debate in Scotland, and in some instances the surveys can be complimentary. In particular, the Scottish House Condition Survey^[9] (SHCS) looks at the physical condition of Scotland's homes as well as the experiences of the householders. The SHS and SHCS were combined in to a single survey with fieldwork starting in January 2012. First results from the new survey will be released from summer 2013, for which a consultation for future dissemination options will be undertaken late 2012 to early 2013.

There are also a number of GB or UK surveys that include a Scottish dimension. The Integrated Household Survey^[10] (IHS) is a composite survey combining questions asked in a number of Office for National Statistics GB-wide social surveys. The IHS is currently designated as "experimental statistics" so while the results should be considered with some care, in some instances the IHS may be particularly useful for making cross-GB comparisons. Please contact the Survey Methodology and Co-ordination team (0131 244 3339) if you have any queries.

The results are presented in the 12 main chapters covering: household composition; housing; neighbourhoods and communities; economic activity; finance; education; transport; internet; health and caring; local services; volunteering; culture and sport.

Guidance on using the information in the report and a glossary with detailed definitions of some of the key terms are included as annexes. Additional annexes present results on the main classificatory variables used in this report and provide guidance on assessing confidence intervals and the statistical significance of the results.

Additional SHS Reporting

Full details of the survey will also be made available through the companion Technical Reports. There are two parts to the technical reporting detailing the methodology and fieldwork outcomes^[11] and, separately, the questionnaire^[12] used.

A number of other Scottish Government publications covering previous years are also available. A comprehensive listing of all publications is available from the SHS website.^[13]

Changes to the SHS in 2011

Due to the high level of demand for space in the SHS, the questionnaire undergoes a substantial review every two years and a light refresh in the years in between to ensure that the information collected is relevant to current policy interests and is making the best use of the time in the survey. The existing questionnaire was updated prior to the start of 2011 and a number of sections were revised. The key changes made were:

• HA3 (household relationship matrix) new response options added
• RE10C (bus services) items D, E and F deleted and item M added
• RE10E (train services) items D, E and F deleted and item M added
• HH57 and HH59 (benefits) item T added
• HB2 (property type) old question reinstated
• HB3 (lowest floor level) old question reinstated
• HB4 ('other' property type) old question reinstated
• RAND_ID (sexual identity) question added
• RECYC (recycling) item F deleted
• AREA4 (greenspace in area) question deleted
• AREA5 (greenspace used) question deleted
• GREEN1 (distance to greenspace) question added
• GREEN2 (greenspace used) question added

Further information on the SHS Questionnaire can be found via the relevant technical report on the SHS website^[14].

Sampling

The sample for the survey meets a number of criteria. It is designed to provide nationally representative samples of private households and of the adult population in private households. This is achieved by splitting the interview between a household respondent and an adult selected at random from the permanent residents of the household.

In order to meet the reporting requirements, the sample is structured to be nationally representative each quarter and to provide a representative sample for larger local authorities each year. The sample is also designed to provide data for every local authority, regardless of size, over a two-year period. This is achieved by disproportionately sampling to achieve a minimum sample equivalent to a simple random sample of 500 interviews in each local authority area.

This report is based on data collected in the one year sampling period (2011). A new SHS was developed throughout 2011 and went out in to the field from January 2012. The new survey incorporates most of the elements from the SHS 2011 though at the same time incorporates the Scottish House Condition Survey (SHCS)^[15] also. The new survey was designed to provide certain high-level estimates at a local authority level on an annul basis and as such a decision was made to run the SHS 2011 as a single year survey only, rather than as the first year of a two-year SHS 2011/2012.

The sample design for 2011, like the one used from 1999 onwards, uses a multi-stage stratified design with a mix of unclustered and clustered sampling. For the current contract, starting in 2007, the sampling strategy was revised to achieve a higher proportion of interviews from unclustered sample. In general, reducing the level of clustering in a sample increases its statistical efficiency, allowing the same level of precision to be achieved with fewer interviews than would be required from a clustered sample. However, unclustered sampling is generally more expensive, particularly in rural areas because of the larger distances between addresses. The revised sampling strategy was designed to achieve the optimum balance between these approaches. As a result, the cost-effectiveness of the interviewing has improved compared with the previous design because it now requires fewer interviews to achieve the same level of precision.

The SHS sample is selected from the small user Postcode Address File (PAF) for Scotland, expanded to take account of addresses which might only be listed once but actually contain multiple dwellings, such as tenement blocks and multi-storey flats. Although the small user PAF excludes many institutional addresses such as student halls of residence or nurses' homes, there are no geographical exclusions from the survey, which covers all parts of Scotland, including the Highlands and Islands.

The main features of the design are:

• First stage, disproportionate stratification by local authority;
• Within each local authority, second stage stratification by the Scottish Government Urban Rural Classification^[16] with large urban and other urban areas combined into an 'urban' stratum and all other areas combined into a 'rural' stratum;
• Unclustered sampling is used in the 'urban' strata, with addresses sorted by Scottish Index of Multiple Deprivation (SIMD)^[17] and selected systematically from a random starting point;
• Clustered sampling is used in the 'rural' strata, with datazones used as primary sampling units which are selected with probability proportionate to size and, within each, a systematic sample selected from a random starting point.

There are some variations to this overall design:

• In local authorities with 80% or more of the household population in 'urban' areas, the sample is wholly unclustered;
• In local authorities with 80% or more of the household population in 'rural' areas, the sample is wholly clustered;
• The local authorities of Western Isles, Orkney and Shetland have wholly unclustered samples even though the urban rural classification suggests they should be wholly clustered. The reason for this is that sample size in these areas means that between 1 in 6 and 1 in 8 households should be sampled. Clustered samples in these areas would be no more efficient than an unclustered sample but would require larger samples for the same level of precision.

The SHS Interview

Interviewing is conducted in respondents' homes using Computer Assisted Personal Interviewing (CAPI) with data collected by interviewers on handheld or laptop computers.

The survey questionnaire is in two parts. The household reference person, who is the Highest Income Householder (HIH) or their spouse/partner completes Part 1 of the interview. Details of all members of the household, including children, are collected during the household interview. This section asks about the people in the household, access to services like internet and recycling, and opportunities in the community for children and young people. Subsequently a child is selected from all household members under 16 (the 'random child') and the household respondent is asked questions about childcare for that child. A child who is at school is also selected (the 'random school child')^[18] and the household respondent answers questions about the school that child attends and the journey they make to go there.

Once the composition of the household has been established, one of the adults in the household is randomly selected by the computer to complete Part 2.^[19] In all households with a single adult the same person completes both parts, but as the number of adults in the household increases, the probability of the random adult being the same as the household respondent declines.^[20]

The household section of the interview deals with topics such as household composition and current economic situation of household members; accommodation, access to the internet and broadband connection; recycling; cars available to the household, employment details of the highest income householder; household income, savings and use of financial services; housing costs; childcare and schooling. The random adult section deals with marital status, ethnicity and religion, individuals' accommodation change; experiences of homelessness and housing problems; neighbourhoods and community safety; transport modes, car dependency, congestion and road safety; travel planning; use of the internet; public services; income and employment; participation in culture and sport. Further information on the topic coverage is available on the SHS website.^[21]

Response Rates

After excluding addresses that were outwith the scope of the survey^[22], the overall response rate for this sweep of the survey was 68.7% which continues a fairly steady response rate year on year within the SHS. There was significant variation in response between local authorities. The highest response rate was achieved in Orkney (78.9%) and the lowest response was achieved in Edinburgh (59.9%). Further information on response rates and other such information is available in the accompanying SHS 2011 Methodology and Fieldwork Outcomes report^[23].

Weighting

Post-survey weighting takes account of the disproportionate sampling between local authorities, the differential response between authorities and any residual mismatch between the profile of responding households/adults and the profile of the population.

The data presented in the report have been weighted in one of two ways.

Household data (collected in Part 1 of the interview) are weighted to take account of the disproportionate sampling and response between local authorities. The profile of household occupants within each local authority sample is then compared with the age/sex profile of the population, as published by the National Records for Scotland (NRS).^[24] The SPSS software module g-Calib is used to produce calibration weights that match the survey sample to these population estimates. The procedure produces weights that provide both survey estimates and grossed up population estimates. This means, for example, that as well as being able to provide survey estimates (the percentage of households in owner-occupation), the survey can provide population estimates of the total number of households in owner-occupation.

Random adult data (from Part 2) are weighted to reflect both the disproportionate sampling and response to Part 2 between local authorities and the different probabilities of selection within households.^[25] The profile of participating adults is then compared with the profile of adults produced by the National Records for and corrective weights calculated that provide survey and population estimates.

The random child and the random school child are dealt with in a similar way - weighting to correct for disproportionate sampling is undertaken and then residual weights, aligning the sample profile with official estimates, are calculated. The only exception to this is that in the case of the random school child, there are no official estimates of the age/sex profile of school children within each local authority. In this case, the population estimates are calculated within the survey itself using the information on the economic status of all household members and household grossing weights. Estimates of the number of school children in each age group are used as the basis for comparing the profile of the random school child and the profile of all school children.^[26]