The Value of Longitudinal data: using the British Cohort Studies to understand women - PowerPoint PPT Presentation


PPT – The Value of Longitudinal data: using the British Cohort Studies to understand women PowerPoint presentation | free to download - id: 6845ed-YzlhO


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation

The Value of Longitudinal data: using the British Cohort Studies to understand women


Title: Introduction to the surveys, design, data coverage and methodology Heather Joshi & John Bynner Author: Rose Last modified by: Jane Created Date – PowerPoint PPT presentation

Number of Views:25
Avg rating:3.0/5.0
Slides: 45
Provided by: rose173
Learn more at:


Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: The Value of Longitudinal data: using the British Cohort Studies to understand women

The Value of Longitudinal data using the British
Cohort Studies to understand womens employment
from a life course perspective
  • Dr Jane Elliott
  • Research Director (NCDS and BCS70)
  • Centre for Longitudinal Studies
  • 29th June 2006

Structure of presentation
  • The four British Birth Cohort Studies
  • The value of cohort data
  • The 1958 cohort (National Child Development
  • Gendered aspirations at age 11 (combining
    qualitative and quantitative methods)
  • Occupation and Fertility
  • Womens employment behaviour after the birth of a

British Birth Cohort Studies
  • Fully representative samples of the British
  • Based on one weeks births - approximately 17,000
  • Followed up from birth into adulthood
  • Four British Birth Cohort Studies
  • 1946 National Survey of Health and Development
    (MRC funded)
  • 1958 National Child Development Study
  • 1970 British Cohort Study 1970
  • 2000/1 Millennium Cohort Study

Housed at CLS
1958 Birth Cohort Study
  • Representative sample of over 17,000 infants born
    in March 1958
  • Not initially planned as a longitudinal study
  • Sample followed at ages 7, 11, 16, 23, 33, 42, 46
    (prospective study)
  • Retrospective life history data collected at age
    23, 33, 42, 46
  • For example
  • work history
  • partnership history
  • fertility history
  • housing history
  • Approximately 12,000 individuals are still
  • Information on individuals can be linked from
    birth and childhood through into adult life
  • Now funded by ESRC with data collected every four

NCDS follow-ups and sources of information
Hypothetical life history
Age 16
Age 42
Age 46
Age 23
Age 33
Research questions best addressed by cohort data
  • Long term outcomes of experiences and decisions
    in early life
  • Medium and short-term outcomes links between
    different life domains (e.g. health and
  • Descriptions of individual trajectories
    careers, relationships, fertility, poverty and
  • The links between social change and the changing
    experiences of different cohorts
  • Intergenerational transmission of advantage and
    disadvantage and the processes involved

NCDS 11-year old Essays
  • At age 11, in 1969 NCDS Cohort members completed
    a short questionnaire (at school) about leisure
    interests, preferred school subjects and
    expectations on leaving school
  • They were also asked to write an essay on the
    following topic
  • Imagine you are now 25 years old. Write about
    the life you are leading, your interests, your
    home life and your work at the age of 25. (You
    have 30 minutes to do this).
  • 13669 essays completed, mean length 204 words
  • Copies of the original essays (in childrens
    handwriting) are available on microfiche at CLS
    and are currently being digitised.

Existing research on the essays
  • A small sample of 521 essays have been coded for
    word count
  • Boys 180 words
  • Girls 228 words
  • All essays have been coded for employment
    aspirations, over 90 give a classifiable
  • No other systematic coding and analysis of the
    essays has been carried out to date

(No Transcript)
Research project funded by the Nuffield
foundation (Elliott and Morrow)
  • Project is intended as a pilot study, if
    successful to be followed up by larger
    application to ESRC
  • Aim to type up and code a sub-sample of 560
    essays conduct preliminary descriptive analyses
  • Sample stratified to reflect gender ability
    social class family structure
  • Essays will be coded for themes such as
  • family life leisure employment housing
    expectations contact with parents pets
    transport and travel aspirations vs expectations
  • Both qualitative and quantitative analysis will
    be carried out using NVIVO.7 and SPSS to help
    organize, code, and analyze the data
  • Main research questions how do gender and social
    class shape childrens aspirations?

Home experiences
  • 46 of the eleven-year-olds were living in
    owner-occupied accommodation while 42 were in
    council housing
  • 44 of children had their own bedroom
  • 19 of girls and 16 of boys shared a bed with
    another member of the family
  • 61 of mothers reported being in work at some
    time since the child was seven (only 3.2 were
    in professional or managerial occupations,
    compared with 20 of fathers)
  • 66 of mothers reported that the father took an
    equal role in managing the child and a further
    24 described the father as having a significant

School experiences (1969)
  • The majority of children were in primary schools
    when they wrote the essays
  • Only 4 of children were at independent schools
  • At age 11 the median class size was 36 pupils
    (mean 34.3), while at age 7 the median class size
    had been 37 with a mean of 35.25
  • 82 of children were in a school with a male
  • 45 of children had a female class teacher

(No Transcript)
(No Transcript)
(No Transcript)
Preliminary analysis (139 essays coded)
  • Girls are slightly more likely than boys to say
    they will be married at age 25 (56 vs. 48)
  • Girls and boys are equally likely to say that
    they want children (45 vs. 42)
  • Girls are more likely than boys to write about
    domestic labour (65 vs. 25)
  • Girls are more likely than boys to mention their
    husband/wifes occupation (20 vs. 10)

282044A I am making dinner for Paul my husband
and Sally my young daughter. I think I will make
us all Shepherds pie and carrots. I car hear him
coming down the garden path, I had better hurry
up. Hello Kathyryn what are we have-ing for
dinner?, asked Paul has he walked into the
kitchen. I brought Sally in from the garden and
wash her hands, then sat her in her high chair.
Has every-thing been aright at the shop? Paul
works in a butchers shop and he has just been
promoted to the manger of five will known shops.
After we had eaten the shepherds pie and Sally
had eaten her bacon and chicken baby food I
started serving the blomonge.
Serial number 110335Y I am 25 years of age, I
live in a big house it has five bedrooms, dining
room, lounge, and kitchenet. I have a wife and
two children, a big jagur car which is just big
enough to hold the family. The children are two
boys, very energetic, they love swimming and
rugby. My wife is a type that can mix with
anybody, she is a good bridge, and very pasent,
she is a good nurse when the children hurt
themselves. The names of the family are Brian,
Kevin, the twins, Edith my wife and me, Bobby.
The doors of the house are orange and round the
windows of the door white, the window frames are
painted cream. We have oil fired central heating
and an electric cooker, and strip lighting in the
kitchenet. My wife does not work because I bring
plenty money home from were I work, I work at the
catterpiller factory at Birtleys I work as a pot
spot welder, because I do the work of putting the
holds through the bits of steel that make the
catterpiller tracks. Each piece of steel weighs
1 ton each. I get payed 31 a week and the firm
payes my petrol bill.
223004D I have two children one 4 the other 6. I
am working in a school as a teacher. My husband
is an Estate Agent. In the evenings we have tea
and each person tells the rest about any
interesting avents which has happened to them.
Then we have a sit and talk about what we are
going to buy or if we are going to buy a house or
not. We then watch Calendar and the news. At half
past seven the children will have a bath and be
put to bed. Meanwhile my husband will be seeing
that the cars engine is all right so that we
shall not be stranded any where is something goes
wrong. After that I shall do the ironing or
washing or any other jobs that want doing. At
Eleven oclock we shall go to bed and get up at
seven oclock. At eight oclock I would take the
children one to school with me the other to her
granmas. My husband goes to work at nine
oclock. At school I illegible teach first
English, Maths and Art. In the afternoon P.E.,
Projects and History.
The salience of occupation
  • Understanding heterogeneity of women
  • Recognise differences between women in terms of
    fertility and working patterns
  • Move away from voluntaristic accounts and notion
    of innate differences
  • Crompton Harris suggest that occupations may
    provide a major social filter through which the
    lives of individuals and families are
    structured(1998 p 299)

Professionals vs. Managers
  • Professionals
  • License to practice/credentials obtained
  • Long term planning of career possible
  • More flexible working arrangements
  • Educated to at least degree level
  • Managers
  • Careers forged in organisation
  • Careers less well planned
  • Fewer opportunities for part-time work
  • May not be so highly educated

Cross-national work by Crompton and Harris
focusing on bankers and doctors
  • Occupation impacts on the way that women combine
    paid employment and family life
  • Compared with professionals, managers
  • Have fewer children
  • Adopt less traditional division of labour

Results from Crompton and Harris cross-national
  • of women with two or more children
  • Doctors 54
  • Bankers 32
  • of women with a traditional domestic division
    of labour
  • Doctors 73
  • Bankers 46

Limitations of Crompton and Harris empirical work
  • Relatively small (non-random) sample
  • Focus on case study of retail banking vs medical
  • Differences between the two groups in terms of
  • Professionals vs. managers
  • Public vs. private
  • Levels of feminisation
  • Levels of qualifications
  • Cross-sectional study
  • Analysis of NCDS data can ameliorate many of
    these problems

Analyses using NCDS (1958 cohort)
  • Advantages of using NCDS
  • Bigger more representative sample
  • Longitudinal data ability to control for
    fertility aspirations
  • Multivariate analysis predicting having children
  • Logistic regression including occupation, views
    on having children at age 23, level of highest

Occupational variations in number of children
  • Data from Crompton and Harris study
  • women with two or more children
  • Doctors 54
  • Bankers 32
  • Data from NCDS (1958 cohort)
  • of graduate women with children
  • Professionals 75
  • Managers 64

Predicting who has had children Predicting who has had children Predicting who has had children Predicting who has had children Predicting who has had children
B S.E. Sig. Exp(B)
Want kids (age 23)?     0.00  
Don't know -1.37 0.16 0.00 0.25
No -2.12 0.14 0.00 0.12
SOC90 Major group     0.00  
Managers (ref cat) ref ref ref ref
Professionals 0.34 0.17 0.04 1.40
Assoc prof tech 0.43 0.15 0.01 1.54
Clerical 0.31 0.13 0.02 1.37
Craft related 0.72 0.32 0.03 2.05
Personal service 1.33 0.18 0.00 3.78
Sales 0.81 0.20 0.00 2.26
Machine operatives 0.87 0.30 0.00 2.39
Other 1.40 0.26 0.00 4.05
Not working 1.02 0.15 0.00 2.78
Degree by age 42 -0.43 0.12 0.00 0.65
Constant 1.32 0.11 0.00 3.76
N4934 women aged 42 from NCDS sweep 6

N361 graduate women aged 42 from NCDS sweep 6
In summary
  • NCDS (1958 cohort study) shows
  • Differences in family building between managers
    and professionals
  • No differences in domestic division of labour

Women and employment previous research
  • Major change in British society since the 1950s
    has been increase in numbers of women in the
    labour market
  • A growing body of research has focused on the
    length of time between childbirth and returning
    to paid employment.
  • Joshi and Hinde (1993)Dex, Joshi, Macran,
    McCulloch (1998) Blank (1989)Joesch (1994)
  • Emphasise that the age of the youngest child is
    the most important determinant of womens
  • Highly educated women display the greatest
    continuity in employment across the childbearing
  • Growing polarisation between women

  • The majority of existing research has
  • not distinguished between part-time and full-time
  • not fully exploited the temporal features of
    longitudinal data
  • not explicitly looked at duration effects

Key Research Questions
  • Are the processes of returning to part-time work
    and full-time work different?
  • Is part-time employment a bridge into full-time
  • Are there significant duration effects such that
    the longer a woman stays out of the labour market
    the more difficult it is for her to return?
  • Are there significant interaction effects between
    duration and other variables. i.e. do the
    influences of some factors change over time
    within individual careers?

Research approach
  • Uses retrospective life history data from NCDS to
    focus on womens employment and fertility careers
  • Examine transitions between episodes of
  • Full time employment
  • Part-time employment
  • Non-employment
  • Emphasis on duration effects and time varying
  • Focus on whole work history since leaving full
    time education, not just on transitions after
    child birth

State Transition Matrix based on womens work
histories from the National Child Development
State Transition Matrix based on womens work
histories from the National Child Development
Approach to Event History analysis
  • Data is available for each month of womens life
    histories between age 16 and 33.
  • This data was discretised to form a data matrix
  • Logistic regression was used to estimate models
    to indicate which factors were associated with
    transitions back into the labour market.
  • Logistic regression models do not allow for the
    link between woman/months at the level of the
    individual woman.
  • More sophisticated mixture models i.e. random
    effects models were therefore estimated using the
    SABRE software to exploit the structure of the

Models estimated
  • Models estimated focused on moving back into the
    labour market
  • not employed -gt full-time
  • not employed -gt part-time
  • part-time-gtfull time
  • Logistic regression models initially estimated
    showed negative duration effects (i.e. pooled
    cross sectional models)
  • Mixture models estimated allowing for unobserved
    individual heterogeneity indicated that the
    initial models were miss specified

Summary of main results from mixture models
(No Transcript)
(No Transcript)
(No Transcript)
Narrative elements of cohort studies
  • Allow us to trace lives through time understand
    how childhood circumstances may impact on adult
  • In the 1958 cohort study, essays at 11 provide
    insights into childrens own narratives about
    their lives
  • Potentially allow for the construction of
    individual case studies based on detailed
    information collected over the years (while
    preserving confidentiality)
  • Allow for a focus on the historical context which
    has helped shape individual experiences
  • Comparisons between cohorts can enable the
    development of a narrative about social change

Website Please register for regular