Transcript Lecture 5. - University of Bath

Lecture 5.
Social Survey Research III:
Other Techniques Of Quantitative Data Collection And
Generation: Secondary Data, Experiments And Social
Indicators
Leah Wild
Overview
 Questionnaire Design: Response Format
 Pilot Questionnaires
 Using Existing Sets of Questions &
Secondary Data Sets
 Data Collection and Experimental Design
 Social Indicators
 Society at a Glance: OECD Social Indicators
2002 –Study
 BCS
Questionnaire Design: Response Format
Intensity scales.
A good way of writing closed-ended questionnaire items is to measure people’s attitudes
and opinions with intensity scales.
A commonly used one is the Likert scale, a scaling technique designed by
Rensis Likert in 1932.
It typically makes use of a 1 to 5 rating scale where
1 is ‘strongly agree’,
2 is ‘somewhat agree’,
3 is ‘neutral’,
4 is ‘somewhat disagree’ and
5 is ‘strongly disagree’
Pilot Questionnaires And Interview Schedules
 Clarity Check
 Adequacy of Instructions Check
 Gain Experience In Asking Questions (i.e.
Interview Schedules)
Secondary Data.
 secondary data as any information that has not





been generated personally (first-hand) by the
researcher who uses it.
highly quantified sources -statistics
to the more qualitative -personal documents and
diaries, government and business reports and the
mass media.
can be contemporary or historical.
saving of time, money and effort.
use of secondary sources may be a necessity if
historical and / or comparative research is being
carried-out.
Problems with Secondary Data
1. Authenticity
2. Credibility
3. Representativeness
4. Meaning (literal and metaphorical)
5. Purpose
(Scott 1990. ‘Documents in Social Research’: Social Studies
Review, 1990)
Reasons to Use Secondary Data.
 Availability.
 Examination of trends / changes over
time.
 Comparisons.
 “Before" and "after" studies.
(Bilton et al 1997)
USING META-DATA.
 Traditionally:
metadata has been understood as “Data about Data” - an
integral part of the dataset.
 Metadata can include such items as:
• Questionnaires;
• a keyword list of names and items;
• a record of how the data were collected;
• an inventory of related data and publications;
• information on principal investigators;
• geographical coverage.

 Metadata is important as it provides essential information on the
use and interpretation of information contained in a data file. It
outlines what datasets are available, where they can be found,
what format they are in, how they should be used, and can provide
a permanent record in catalogues and search facilities of the
resources that are available.
Data Collection: Secondary Data Sets











British Household Panel Survey (longitudinal)
5,500 Households,
10,000 Individual interviewees
75% response rate
Focus is on social and economic change at individual
and household level
Topics include:
Income
Employment
Housing
health
Attitudes.
Labour Force Survey
 1973 Onwards
 Up to 60,000 households
 80% response rate
 Topics include:
 Employment
 Education and Training,
 Health and Disabilities
General Household Survey














Annual since 1971 (except 97-98 and 99-00)
10,000 households, 25,000 people
72% response rate
Topics include:
household and family information
housing tenure and household accommodation
consumer durables including vehicle ownership
employment
education
health and use of health services
smoking and drinking
family information including marriage, cohabitation and fertility
income
demographic information about household members including
migration.
British Crime survey









a 'victimisation' survey,
experiences of property crimes and personal crimes
provides a record of the experience of crime
unaffected by variations in the behaviour of victims about reporting
Unaffected by variations in police procedure/practice
Between 1982- 2001 biennial. From 2001 annual.
Target population: All households in England and Wales
Sampling unit: People aged 16 or over living in private household.
Design/selection of sample: Stratified random sample of postcode
sectors. One quarter sector selected at random. Interviews conducted
at address
 Size:20,000 respondents in BCS (2000) core sample, plus 4,000 ethnic
minority booster sample.
 Enlarged sample size in 2001. 40,000 core sample
 Core crime counting questions remain unchanged since first sweep
(1982). Other questions asked on ad hoc basis. Moved to CAPI /CASI
1994.
Data Collection and Experimental
Design
 When interested in understanding how the





manipulation of a variable can explain specific
outcomes on another variable, experiments can also
be used.
Experimental Group
Some treatment is conducted on the subjects of the
experimental group and measures its effects in
comparison to another group.
Control Group
Control group might not receive any treatment or it
might receive a different kind of treatment
Social Indicators
 Society At A Glance: OECD Social Indicators


1.
2.
3.
4.
2002 Study
Context Indicators illustrate national
differences in social trends and social status
Response Indicators. Four areas of social
policy have been operationalised:
Self-sufficiency
Equity
Health
Social Cohesion
 UK official statistics www.statistics.gov.uk
 UN statistics
www.un.org/depts/unsd/global.htm
 OECD statistics www.oecd.org/statistics
 EU statistics europa.eu.int/comm/eurostat/