Reflections Colin H Williams School of Welsh Cardiff University, UK [email protected].
Dr Scott Orford WISERD Cardiff University [email protected]
description
Transcript of Dr Scott Orford WISERD Cardiff University [email protected]
Understanding Wales: Opportunities for Secondary Data Analysis
Living in Wales / National Survey for Wales
Dr Scott OrfordWISERD Cardiff University
• Main source of general statistical information about households in Wales outside of the census
• Undertaken in 2004 until 2008
• About 12,000 households samples with an average annual response of around 7,500 households (60%)
• Different households surveyed each year
• In 2004 and 2008 a property survey (HCS) also undertaken by qualified surveyor (2,500 properties)
What is the Living in Wales Survey?
QuestionsThemes and questions varied each year (some core)Examples of themes:
• Current accommodation• Disability and caring responsibilities• Equalities• Financial• Home / neighbourhood• Household composition• Housing history• Internet usage• Quality of life• Tenure and housing costs• The environment• Transport• Values and opinions• Volunteering• Welsh language use
Geographical breakdowns
• One year: national breakdown• Two consecutive years: regional breakdown• Three consecutive years: LA-level breakdown• Four or more consecutive years: lower-level
geographies? (special license)• Weights so data can be nationally representative
Identifying ‘localities’ for WISERD research
•Analysis of secondary data to identify possible localities for in-depth research – empirically informed
•Several metrics considered (e.g. housing markets / labour market areas / areas differentiated by socio-economic characteristics eg. cultural and identity signifiers)
•People’s attitudes to neighbours and neighbourhoods in which they live
•Living in Wales survey
Values and opinions
Eight attitudinal questions have been asked with a likert (1-5) scale response
1. The degree to which the respondents trust the people in their neighbourhood
2. How they rate their neighbourhood as a place to bring up children
3. How often they talk to their neighbours
4. Whether they feel that they belong to their neighbourhood
5. Whether friendships with people in their neighbourhood means a lot to them
6. Whether they can get advice from their neighbours, borrow things and ask favours of their neighbours
7. Whether they can work together to improve their neighbourhood
8. Whether they see themselves as similar to their neighbours
Pooling data from all the surveys creates a large enough sample size for analysis at the Middle Layer Super Output Area (MSOA) level (c. 35,000 records)
They have a minimum of 5,000 residents and 2,000 households with a mean population of around 7,200.
Wales has 413 MSOAs and these will be used in this analysis to represent neighbourhoods.
Average 74 households (sd 36) per MSOA in the pooled survey (2004-07)
Min 21 hhs per MSOA to max of 225 hhs
10th - 90th percentile range is 36 to 118 households per MSOA, the distribution is not heavily skewed.
MSOA unit of analysis
•A statistical technique which classifies cases into groups based on similarity of each case with respect to selected variables (8 attitudinal questions) – data driven classification
•Cases in a group are more similar than cases in different groups
•Z-scores were calculated for each attitudinal variable and these were aggregated to a population weighted average z-score for each of the 413 MSOAs
•Neighbourhood level z-scores were used in a k-means cluster analysis based on Euclidean distance (as a measure of similarity)
•Five clusters emerged as being optimal in terms of maximising difference between clusters and parsimony
Cluster Analysis
Cluster ID
Distances between cluster centres
No. of MSOAs1 2 3 4
1 90
2 0.98 76
3 0.70 0.64 136
4 1.51 0.83 0.84 93
5 2.59 1.88 1.95 1.14 18Cluster 5 is the most distinctive and has the fewest members
Cluster 3 has the most similarity to the other clusters and has the largest membership
Cluster ID Map LabelGeographical description
Conventional Map Area (km-
sq)
Cartogram Map Area (km-sq)
ProportionalDifference in
Map AreaPopulation
2001
1 RuralRural mid- west- and north-west
Wales15552 9214 0.59 608186
2 ValleysValleys and
former mining areas
1223 2045 1.67 566472
3Semi-rural hinterland
Suburban and semi-rural areas
2942 3808 1.29 952310
4 DeprivedDeprived – non-former mining
areas627 1840 2.94 656505
5Urban Mobile
Inner-city neighbourhoods
37 303 8.29 119612
What is the National Survey For Wales?
• Successor of LiW Survey• 2009-2010 pilot survey• face-to-face interviews in a randomly chosen sample
of 4,600 households across Wales. • The survey covers the head of household and a
randomly selected adult. • Short self-completion questionnaire for every adult
in the household aged 16 or over. • Main themes of the survey are public services and
wellbeing
First Wave NSW
• Begin January 2012 and fieldwork will run continuously all year around
• Two specialist research companies: TNS-BMRB and Beaufort Research Ltd
• Aim –random sample of 25,000 households across wales with successful response of c.14,500 (660 per LA).
• The target response rate across Wales 70%
Three sets of questions
• Public services - local authority services, the education system, health services, and the transport system
• Wellbeing - five wellbeing questions by ONSrespondents’ proximity to green and blue spaces community cohesion questions about local area perception of quality of local neighbourhood people are keeping up with bills, credit and debt
• Core demographic / household questions
Dates of Statistical Releases
• September 2012: headline Wales-level results• Late summer 2013: detailed Wales-level results
Wales-level results broken down by age, gender, etc.LA-level headline results
• Annually from summer 2014: detailed results at Wales and local authority level
Wales-level results based on interviews in the previous financial yearLA-level results based on interviews in the previous 2 financial years
Data Access
• Micro-data at UKDA (including pilot survey data)
• End-user license• Special License• Individual data records linked to SAIL through
The Health Information Research Unit (HIRU) based at Swansea University
Further Resources
• ESDS http://www.esds.ac.uk/
• WG LiW http://wales.gov.uk/about/aboutresearch/social/ocsropage/living-wales/?lang=en
• WG NSW http://wales.gov.uk/about/aboutresearch/social/ocsropage/nationalsurveyforwales/?lang=en
• WG Surveys [email protected] tel: 029 2082 3035