The British Household Panel Survey

bagimpertinentUrban and Civil

Nov 16, 2013 (3 years and 4 months ago)

192 views

ESRC UK Longitudinal Studies Centre

The British Household Panel Survey

Nick Buck

Institute for Social and Economic Research

University of Essex

ESRC UK Longitudinal Studies Centre

What is a Household Panel Study?


Household panel studies involve repeated data
collection about a sample of individuals


They are distinct from cohort studies and individual
level panels in following all members of sample
households


First wave is similar to cross
-
sectional household
survey (e.g. General Household Survey)


In subsequent waves follow individuals as they move
between different households, and interview new
household members

ESRC UK Longitudinal Studies Centre

Why use household panel surveys


Household panel especially useful where
household context is important (e.g.
demographics, income dynamics, employment
participation, social values)


Family relationships over time


Collects information on changing household units


Typically have short intervals between surveys,
allows collection of ‘continuous’ information


ESRC UK Longitudinal Studies Centre

The development of Household
Panel Studies


Household Panel Studies have become the leading survey type
for cross
-
national longitudinal research


Began with Panel Study of Income Dynamics in USA
-

analysis
of poverty persistence 1968


SOEP in Germany/ SEP in Netherlands 1984


BHPS in GB 1991
-

widening agenda


ECHP 1994
-

cross national comparison (also CNEF)


Understanding of transition countries (Hungary, Bosnia)


Understanding of differences within nation states: East/West
Germany, Scotland, Wales, NIHPS, NHPS

ESRC UK Longitudinal Studies Centre

British Household Panel Survey


Annual survey of members of initial sample of
5511 households interviewed in autumn 1991


Random sample representative of population of
Great Britain (south of Caledonian canal)


The sample frame was the Postal Address File;
The sample was clustered and stratified, 250
primary sampling units, stratified by region, and
census social indicators.


The sample is essentially equal probability
(Sampling information on file PSUs and Strata are
distinguished


though not identified)


ESRC UK Longitudinal Studies Centre

BHPS Interviews


Interviews with all adults, aged 16+, in each
households (40 minute interviewer
administered, plus 10 minute self
completion)


Household level interview (10 minutes)


From wave 4, interviews with young people
aged 11
-
15

ESRC UK Longitudinal Studies Centre

Following rules


Indefinite life panel study, ‘Following rules’
maintain representation of original population and
their descendants


these specify who should be
eligible to be interviewed at each wave.


The Longitudinal Sample consists of: members of
original households, and their natural descendants
born since the start of the panel


The above are eligible for interview each wave so
long as they remain in scope (i.e. in GB / UK from
1997)


Sample members are followed as they move.

ESRC UK Longitudinal Studies Centre

Following rules (2)


Original sample only included members of private
households. At subsequent waves interviews sought with
members in institutional accommodation.


At each wave the interviewed sample also contains co
-
residents of longitudinal sample members. These will not
in general be followed if they no longer live with a sample
member.


The panel sample will be reduced by:


Attrition


refusal and non
-
contact


Members becoming ineligible
-

Deaths and moves out
of scope


ESRC UK Longitudinal Studies Centre

The BHPS additional samples


Wave:

1

2

3

4

5

6

7

8

9

10

11

12

13

14
-
18

Original Sample

X

X

X

X

X

X

X

X

X

X

X

X

X

X

ECHP ‘low income’
and Northern
Ireland

X

X

X

X

X

Scotland and Wales
extension sample

X

X

X

X

X

X

Northern Ireland
extension sample

X

X

X

X

ESRC UK Longitudinal Studies Centre

Scotland and Wales Extension
Samples


First Wave in 1999/2000
-

Wave 9 BHPS,
approximately 1500 new households in each


Sample structure similar to BHPS wave one, sub
-
regional stratification, Highlands and Islands


Generally contain the same questions, except
where early wave life histories carried instead of
BHPS variable components.


Additional questions on national identity etc.


Released with main BHPS data, and to be
analysed in conjunction with those data

ESRC UK Longitudinal Studies Centre

Northern Ireland Household Panel
Survey


First Wave in 2001
-

Wave 11 BHPS,
approximately 2000 new households


Simple random sample across the whole of
Northern Ireland


Generally contain the same questions, except
where early wave life histories carried instead of
BHPS variable components.


Released with main BHPS data, and can be
analysed in conjunction with those data


ESRC UK Longitudinal Studies Centre

BHPS Sample Sizes - Adult Respondents
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
Wave 1
Wave 2
Wave 3
Wave 4
Wave 5
Wave 6
Wave 7
Wave 8
Wave 9
Wave 10
Wave 11
Wave 12
Northern Ireland
Scotland
Wales
ECHP
Original
ESRC UK Longitudinal Studies Centre

BHPS topics


BHPS questionnaire consists of core questions repeated
each year and variable components. Core questions cover
the following areas:


Housing and consumption, neighbourhood characteristics


Household organisation, domestic work


Education and training


Labour market behaviour, current job, and annual job history


Health, limitation of activities, use of health services


Social and political values, social participation and networks


Income: current and annual measures, non
-
monetary
indicators

ESRC UK Longitudinal Studies Centre

Questionnaire changes across time


In general questions remain fixed across the panel waves,
and associated variables containing responses will be
found in each waves data set. Exceptions are:


Initial conditions questions, asked only once (e.g. school
leaving age)


though in this case the variable is present at
each wave, but only valid for some


Intermittent rotating questions


Variable components (e.g. wealth, life history)


Questions dropped as low priority


New questions

ESRC UK Longitudinal Studies Centre

Types of BHPS data


Initial conditions / life histories


Repeated annual measures


Continuous history information (work, income, family)


Irregular topics, collected from variable components (e.g.
wealth, ageing etc.)


Permits a range of different statistical methods, including panel
repeated measures models and duration models.


Data files contain weights for a range of analysis, imputation for
item non
-
response, and a considerable number of derived
variables.

ESRC UK Longitudinal Studies Centre

The BHPS and longitudinal research


BHPS supports a wide range of types of longitudinal research


Analysis of gross (individual level) change


inflow and outflow
measures (e.g. employment and family status)


Inherently longitudinal phenomena (eg unstable employment,
poverty persistence).


Controlling for unobserved characteristics


Causal inferences from temporal sequence


both short term and
long term


Relationships between attitudes, expectations, preferences and
behaviour


Analysis of quasi
-
experiments


impacts of policy


ESRC UK Longitudinal Studies Centre

Examples of Research Using BHPS


Poverty and income dynamics / impacts of family change


Welfare in old age and wealth accumulation over the life course


Impacts of class or human capital on life chances


Scarring effects of unemployment and poor quality jobs on later
employment


Impacts of life events on psychological well
-
being


Choice between marriage and cohabitation


Neighbourhood effects on social exclusion


Sharing of political attitudes within the household


Impacts of parental circumstances on child outcomes


ESRC UK Longitudinal Studies Centre

BHPS data design


BHPS data comes in a series of separate files which reflect
the way it was collected, and the main units of analysis at
each wave


e.g. a file containing most data collect from
each adult respondent.


These files can be linked by simple matching operations.
These operations are aided by some basic file and variable
naming conventions


common root name with wave
specific prefix


Data are released using the same file structures for all
software (SPSS, SAS, STATA, SIR


In addition to variables from the questionnaire there are
large numbers of derived variables some of which aim to
reduce the amount of file matching required.

ESRC UK Longitudinal Studies Centre

Documentation


Extremely comprehensive and easily accessible
documentation


PDF files are available from the Archive, but now easier to
use web version:
http://iserwww.essex.ac.uk/ulsc/bhps/doc


Volume A


descriptive information about survey and
data.


Volume B


detailed variable lists, questionnaire info,
marginals, links to other waves
-

this volume is indexed by
wave and record type, and by subject thesaurus


Questionnaires are also available as PDF files for
download.

ESRC UK Longitudinal Studies Centre

Future data plans: variable
components


Wave 14 (2004) National identity, views of UK
governance, additional health measures (SF36), and
measures of work attitudes and work stress.


Wave 15 (2005)
Wealth, Assets and Debts


second
measurement in Scotland


Wave 16 (2006)
Ageing, retirement, health, and quality of
life


Wave 17 (2007) Children and parenting, non
-
resident
parents, aspirations of young people



Wave 18 (2008) Neighbourhood, expectations of
relationships and marriage in the future

ESRC UK Longitudinal Studies Centre

Other issues for the future


Sample size, re
-
sampling


Representation of migrants to the population


Linkage to contextual data


neighbourhoods,
organisations (e.g. school, workplace)


how do we dal
with disclosure risks?


Linkage to individual administrative data


New agendas for data collection e.g. health measurement,
crime and victimisation


New methods of data collection (e.g. web
-
surveys)


ESRC UK Longitudinal Studies Centre

More information


Documentation, including lists of publications based on
BHPS available at
http://www.iser.essex.ac.uk/ulsc/bhps


Also provides information on BHPS user group


Data from Economic and Social Data Service (Data
Archive) http://www.esds.ac.uk