Scottish Household Survey User Day
26th November 2007
Mairi MacAskill
SHS Project Manager
Structure• Introduction to the SHS
• Survey Management
• Contract 2007-2010
• Questionnaire Content
• Sampling
• Dissemination
• Future Developments
Why was the SHS commissioned?
• Ministers agreed that devolution preparations should include improving information sources in Scotland
• Information needs were expected to increase with the new Scottish Parliament
Aims of the SHS
• Provide household and individual information to support transport, local government and communities policy areas
• Permit disaggregation of such information both geographically and in terms of population subgroups
• Allow the relationships between social variables within households to be examined
• Allow early detection of national trends• Allow detailed follow-up surveys of sub-
samples from the main survey
Background
• Survey of private households in Scotland– c15,000 interviews per year
• Continuous since February 1999
• No geographical exclusions
• Face to face using Computer Assisted Personal Interviewing (CAPI)
Process
• Once sample has been drawn, respondents are contacted by letter before the interviewer calls
• Interviewer must call up to 6 times to secure the interview, with one evening and one weekend call
• If non-contact or a “soft” refusal, the addresses may be reissued to more experienced interviewers
• Response rate has been stable at 70%
Interview Structure
• Two part interview– Highest Income Householder or spouse/partner
• Information about the households in Scotland• Characteristics of HIH help characterise the
household as a whole
– Randomly Selected Adult• Information about adults in Scotland
• 45 minutes long
Survey Management
• Management of the SHS within the Scottish Government has moved from Social Research to Statistics
• Management contained within Housing and Household Surveys Branch
• Joint working with the Scottish House Condition Survey
Survey Team
• Mairi MacAskill – Scottish Household Survey Project Manager
• Jonathan Waite – SHS Research Officer• Vacant – SHS Statistical Database
Administrator• Alan Sloan – SHS Assistant Statistician
Survey Management Groups
• Technical Group– Main management group– Includes senior analysts and analysts from
key priority areas, as well as the contractors– The Project Manager seeks advice on issues
before making decisions
• Fieldwork Group– Monitors fieldwork progress– Discusses detailed data issues
New Contract
• Contract for 2007-2010– Let in August 2006
– Consortium of three companies• Ipsos MORI• TNS System Three• Scottish Centre for Social Research
Questionnaire – Household Section
• Socio-economic details of householders– Name– Gender– Age– Relationship to Highest Income Householder– Economic Status
• Accommodation– Property Type– Tenure– Number of Bedrooms
Questionnaire – Household Section
• Internet Access• Recycling• Household Transport
– Number of cars available– Whether bicycles are available for use– Blue Badges/Thistle Cards/Taxi Cards– Access to Bus Services
• Young People in the Household– Childcare– Travel to school
Questionnaire – Household Section
• Health and Disability– Long-term illness and disability– Help or care
• Householder and Spouse Employment
• Householder and Spouse/Partner Income
• Financial Services
• Mortgage and Rent
Questionnaire – Random Adult Section
• Additional Socio-economic information– Marital Status– Ethnic Origin– Religion
• Accommodation– Time at address and tenure history– Homelessness
• Neighbourhoods and Community Safety– View of neighbourhood– Anti-social behaviour– Family Cohesion
Questionnaire – Random Adult Section
– Green Space– Discrimination– Neighbourhood Cohesion– Safety whilst walking and using public transport
• Education and Training– Qualifications held– Internet Use
• Travel and Transport– Travel Diary– Driving – Road Rage– Car dependency
Questionnaire – Random Adult Section
– Park and Ride– Traffic Growth and Congestion– Travel to Work– Car Sharing– Motorcycling– Bus and Train Services– Travel Information– Traffic Calming– Road Accidents
• Council Services– Use of and Satisfaction with services
Questionnaire – Random Adult Section
• Volunteering• Travel to Health Services• Culture and Sport• Health and Disability
– General Health– Illness and Disability– Specialist Equipment– Limitations– Caring
• Random Adult Employment• Random Adult Banded Income
Analytical Potential
• Geographical variables are added to the data to “add value” to analyses– Urban/Rural– Scottish Index of Multiple Deprivation
quintiles, deciles and 15% flag
Analytical PotentialHouseholds with access to the internet: 2006
0
10
20
30
40
50
60
70
80
1 2 3 4 5 6 7 8 9 10 ALLSCOTLAND
Scottish Index of Multiple Deprivation deciles
Per
cen
tag
e o
f h
ou
seh
old
s
Analytical Potential
• The wide range of topics collected in the SHS means there is huge potential for cross-cutting analysis– Transport and Health– Housing and Financial Inclusion– Education and Internet Use– etc
Analytical PotentialConvenience of Doctor's Surgery - % saying they found it "very" or fairly
convenient: 2006
50 55 60 65 70 75 80 85 90 95 100
None
One
Two or more
ALL SCOTLAND
Nu
mb
er
of
ca
rs a
va
ila
ble
Percentage
Analytical Problems
• Large, complex dataset
• 2005/2006 Dataset – almost 2,500 variables
• Weighting
• Complex sample design
Questionnaire Reviews
• Background – why do we carry out these reviews?
• Outcomes of questionnaire review for 2007
• Looking forward – questionnaire review for 2009– Scope– Other considerations
Questionnaire Review
• Background– SHS is revised at the beginning of each two
year sweep (odd years)– To ensure that the content is used and
relevant to policy interests– It is National Statistics Best Practice to
regularly review data collections, after consulting with users of the data
Questionnaire Review for 2007
• Questionnaire was reviewed throughout 2006
New topics introduced into the SHS, for example:• Discrimination• Culture and Sport• Family Cohesion• Road Rage• Travel to Health Services
Questionnaire Review for 2007: Sub-sampling within the survey
The SHS Review (2005) advocated modularisation of the survey:– Flexibility allows many more questions to be
asked of sub-samples
• Considerations:• Geographic Disaggregation• Size of change being monitored• Error around estimates
Sample Size
45 minutes 25 minutes
Socio economic
core
Tra
nsp
ort C
ore
Local G
overn
men
t Core
Com
mu
nitie
s C
ore
Culture and Sport Module
Local Government Module One
Transport Module One
Com
mu
nitie
s M
od
ule
Tw
o
Transport Module Two
Other Modu
le Two
Communities Module One
Socio economic
core
Rota
ting
Core
Con
ten
t ch
an
ges p
erio
dic
ally
Socio
econ
om
ic
rota
ting
core
Oth
er M
od
ule
On
e
Other Core Questions
Questionnaire Review for 2007: Sub-sampling within the survey
• In practice the extent of sub-sampling was less than that envisaged
BUT…
• Many full-sample questions were reduced to three-quarter sample in June 2007
Questionnaire Review for 2009
• Within the next couple of weeks you will receive information about how to get involved
• The revision for the start of the 4 year contract period is traditionally much more extensive
• This revision is in the middle of the 2007-2010 contract period and so will be relatively less extensive than the previous revision
Scope of questionnaire review
• Priority areas– Will work within their present time
• Non-priority areas– Will have to justify existing and new
questions, to ensure this is the best use of the time in the survey
Questionnaire Review – other considerations
• Harmonisation Agenda– Scottish Population Surveys Coordinating
Committee’s Harmonisation Working Group is putting together a harmonised “core” of questions
– These questions will be owned by the SHS TG and may take up more time than at present
Questionnaire Review – other considerations
• Survey good practice issues– Length of pre-code list– Length of show cards– Whether write-ins required– Risks of different types of attitudinal questions
• These issues will be evaluated for as many questions as possible to ensure good quality data
Questionnaire Review – other considerations
• Interview Coherence– Some interviewers complain about the
interview “jumping about”– The vision for the survey for 2007 was a core-
modular structure, so that certain interview types were dominated by certain topics
– Interviewer feedback session 7th February 2008
Culture and Sport Module
• Dedicated Culture and Sport Module
• 7,500 interviews over the two year sweep
• Spread across Scotland
• Socio-economic questions used from the main survey
Sampling• The sample is representative for:
– Scotland as a whole for every quarter of every year
– Each of the larger local authorities for every year
– Every local authority area (regardless of size) for each two-year sweep
– For each category of the six-fold urban-rural classification for every year, and
– For the SIMD 15% areas taken together as a group, for every year
Minimum sample size• Simple Random Sample Equivalent:
– Scotland as a whole – 2,500 per quarter
– Each of the larger local authorities – 500 per year
– Every local authority area (regardless of size) – 500 for each two-year sweep
– For each category of the six-fold urban-rural classification - 500 every year, and
– For the SIMD 15% areas taken together as a group – 2,000 every year
Sampling• The survey sample design was changed
for 2007;– Lesser proportion of the sample is clustered– Primary sampling unit has changed
• Postcode Address File (PAF) still used as the sampling frame
• The design is more efficient than before – the desired number of SRS equivalent interviews is achieved through fewer interviews
Implicit Stratification of addresses by a measure of class/deprivation
Implicit Stratification of PSUs by a measure of class/deprivation
Stratification by LA
Areas divided into clustered & unclustered
UNCLUSTERED CLUSTERED
Systematic Sampling of PSUs from Random Starting Point
Systematic Sampling from Random Starting Point within PSUs
Systematic Sampling from Random Starting Point
Sampling Strategy: Changes
2005/2006: 22 LAs were clustered, 10 unclustered – 42% of the sample was unclustered.
2007 onwards: urban areas within each LA treated as unclustered, rural areas as clustered – 70% of the sample is unclustered
2007 onwards: Datazone PSUS stratified by Scottish Index of Multiple Dprivation ranking
2005/2006: Enumeration District PSUs stratified by MOSAIC
Dissemination Strategy
• Introduced as part of the new contract 2007-2010
• To increase identity, awareness and use of the SHS
• Instead of fixed outputs for a fixed cost there is now a “pot” of dissemination money
• Consulting you on your views today
Future Developments
• Weighting– Non-response (i.e. under-representation of
certain groups, such as young males)
• Income Imputation– Possible Feasibility study into using Family
Resources Survey as donors for those in large adult households who are not the HIH or spouse/partner
Contact Details
• Mairi MacAskill, SHS Project Manager– 0131 244 0824
• Jonathan Waite, SHS Research Officer– 0131 244 5415
www.scotland.gov.uk/shs
Communities Analytical Services 1F(N) VQ
Questions?