Microsoft PowerPoint - NCRM EPrints Repository

skillfulbuyerUrban and Civil

Nov 16, 2013 (3 years and 8 months ago)

46 views

Quantitative methods
for researching lives
through time

Heather Laurie

Institute for Social and Economic Research

University of Essex


laurh@essex.ac.uk

Overview


What do we mean by
longitudinal
?


Distinguish method from design


Cohort and panel designs


What can quantitative longitudinal studies


tell us?


Introduce a new study


Understanding Society



Defining ‘longitudinal’


Term that encompasses many forms of data
and many designs


Used as a catch
-
all for any study that has
some element of time within it


From oral histories through to qualitative
analysis of life
-
course events and transitions


And to quantitative panel or cohort analysis
of repeated observations of the same cases


Defining ‘longitudinal’



Who or what are we following?


For what time period?


How often?


Defining ‘longitudinal’


Common factors regardless of methodology


Information at or about more than one point in
time


Research questions that are inherently
longitudinal in nature


understanding the antecedents of events


process and lived experience


influences on decision
-
making


the timing of transitions from one state to another




A longitudinal research focus


Inherently longitudinal research questions e.g.


Educational and employment outcomes for
children depending on family and social
background


Labour market transitions and the impact on life
time earnings, career progression and well
-
being


Impact of ethnicity and gender on long
-
term
outcomes in the labour market and family life


Inter
-
generational transmission of attitudes,
values, deprivation, wealth




Analysis across the life
-
course


A longitudinal perspective on change in
individuals’ lives over time


Recognises complex interactions between life
course domains e.g. work, living
arrangements, income, leisure, health


Importance of social context e.g. household,
wider social networks, local environment, and
social and cultural norms.


Method or design?


The design of any study depends on the
research questions you want to answer


The method you use to collect those data
depends on many elements



cross
-
sectional or longitudinal


the population of interest


generalisation to a population
vs

case study


the type of analysis needed to answer your
research questions




Quantitative design approaches


Retrospective studies


Past events and transitions are recalled by
respondents e.g. life time employment history


Record linkage panels


From data collected for administrative purposes
e
.
g. taxation or social welfare system


Prospective studies



Produce ‘waves’ or ‘sweeps’ of measurement
collected over a period of years


Cohort studies


Often birth cohorts e.g. MCS but can be any age group/ ELSA
aged 50+


Follow up same cohort members at fairly long intervals
-

e.g.
NCDS now every 4 years; ELSA every two years


Birth cohorts include developmental data as well as social and
economic data


Triangulation of data from parents, teachers, schools
associated with cohort member


Allow observation of long
-
term outcomes with detailed
information on childhood and family circumstances

Household panels


Draw a sample at one point in time and follow those
sample members indefinitely


Collect individual level data in household context


Usually interview all members of the household


Repeated measures at fixed intervals (usually annual
data collection)


Annual data collection allows analysis of short
-
term
change


Indefinite life design allows longer
-
term outcomes to
be observed


What do quantitative longitudinal
panels offer?


Temporal information on sequence of events


Allows us to make better inferences about
cause and effect


Short term dynamics of change


Long term dynamics


Links between current events and outcomes
and past history


What do quantitative longitudinal
panels offer?


Repeated observations on the same individual
controls for the effects of unmeasured
heterogeneity between cases


Reduce recall error


Can better understand social change by
separating out age, period and cohort effects


To establish the effect of a treatment


Sampling techniques mean statistical findings
can be generalised to the whole population

Understanding Society


A new household panel for the UK

Key features:


Large
sample
size of 40,000 households


Includes all of UK


Household focus
with full
age range
sample


Annual interviews with all aged 10 and over


Innovation Panel for methodological research
and testing


Key features(cont)


Multi
-
topic design to meet a wide range of
disciplinary and inter
-
disciplinary research
needs


Ethnic minority research


Biomedical research


Data linkage to administrative records


Opportunities for qualitative linked studies



Annual repeating content


Basic demographic characteristics


Changes between waves
-

employment, fertility, partnering,
geographic mobility, health


Health status (e.g. SF12), disability,


Labour market activity and
employment status, job search


Current job characteristics, basic
employment conditions, hours of
paid work, second jobs


Childcare, other caring within and
outside household



Income and earnings


Life satisfaction


Political affiliation


basic
measures


Transport and communication
access


Education aspirations and
expectations


Consumption expenditure


Housing characteristics


basic


Housing expenditure



Household facilities, car
ownership

Understanding Society

rotating content


Family
and social networks outside the household


Attitudes and behaviours related to environmental issues


Illicit
and risky behaviour especially for young people


Psychological attributes


Cognitive ability measures


Health
outcomes and health related behaviour


Quality of sleep


Well
-
being


Quality of marital relationships


Risk and trust


Collection of data about younger children < 10


Transition
into young adulthood


Discrimination and harassment


Ethnic and national identity

Ethnic
minority research


Boost sample for five key groups in the UK (Indian,
Pakistani, Bangladeshi,
Caribbean
, Black African, Mixed
)


Increasing prominence of research into ethnic
difference for understanding the make
-
up of UK society


Focus on issues of diversity and commonality.


Common questionnaire content across the sample for
many
questions


Additional questionnaire content within the ethnic
minority boost


Linkage
to administrative records


Ask respondents for permission to link
to:


Health
records and hospital episodes statistics


Education
records


Pension
and state benefit records


Parents asked for permission on behalf of children < 16


Link to pupil level and school level education data


each
pupil in England has a Unique Pupil
Number
so can follow as they
progress through the school system


Link survey data to a range of geo
-
coded data, including
environmental data


Biomedical
research


C
ollecting a
wide range of biomarkers and health
indicators


Opportunity
to assess:


exposure and antecedent factors of health status,


understanding disease mechanisms (e.g. gene
-
environment interaction),


household and socioeconomic effects,


analysis of outcomes using direct assessments or data
linkage.


Opens up prospects for advances at the interface
between social science and biomedical research.

Research potential


Developed as a research resource for the
whole user community


First data available from the UK Data Archive
from early 2011


Like the British Household Panel Study(BHPS),
we hope it will be widely used


ISER publications and web contact


For further information about
Understanding Society
see
www.understandingsociety.org.uk



For ISER publications and Working Papers see
www.iser.essex.ac.uk/publications