CCSR's Longitudinal Programme

CCSR Staff and Students
Valerie Antcliff, Shu-Li Cheng, Angela Dale, Shirley Dex, Clare Holdsworth, Andrew Pickles, Mark Tranmer & Yayoi Sugihashi

Associated University of Manchester Staff
Graham Dunn, Daphne Kounali & Shuying Yang (Biostatistics), Richard Harrington (Child Psychiatry), Mike Savage (Sociology).

With the appointment of Andrew Pickles to a joint chair in Epidemiology and Social Statistics our programme of work in this area has developed considerably. Our inter-linked interests and expertise now span theory and substantive research focussed around the life-course, methods of statistical analysis and a new interest in efficient sampling design methods. We see this strength in breadth as a key element in generating both novelty and coherence and in providing a rounded training programme for new researchers in the field.

Life-Course Research

Interest in life-course research grows apace, the concept gaining in use in such diverse fields as sociology, psychology, economics and medicine. Life-course research however, presents a considerable challenge both in terms of methodology and in terms of its fundamentally multi-disciplinary nature. CCSR occupies a key position in the centre of the network of skills necessary for the highest quality work in this area. We intend to follow our very successful one day Life-Course meeting by an ESRC Seminar Series and to use our key position to develop more thoroughly interdisciplinary research in the social and health fields.

 Our substantive research in this area includes:

Transitions to Adult Life

* European Patterns of Leaving Home [ESRC] (CH - see attached page)

* Adult relationship, employment, social support and mental health difficulties of individuals with mental retardation [MRC funded] and of adoptees (AP with Barbara Maughan, King’s College)

* Genetic and environmental mechanisms underlying developmental processes during adolescence including the rise in depression among girls, substance abuse and antisocial behaviour [US NIMH] (AP with Lindon Eaves et al., Virginia)

* Early adult outcomes for children raised in institutions in Greece (AP with Yiotta Vorria, of the University of Thessalonika)

* Relationship and mental health problems of women suffering physical or sexual abuse as children [MRC funded](AP with Jonathan Hill, University of Liverpool)

* Early adult psychological and social outcomes of children who had attempted suicide [PPP HealthCare Trust](AP with Richard Harrington, University of Manchester)

* Life-course psychopathology: a 30-year follow-up of the Isle-of-Wight Sample [MRC & Nuffield](AP& DK with Barbara Maughan, King’s College & Martin Knapp, LSE)

* Developmental trajectories in aggressive and oppositional behaviour and its relationship to early adult crime (AP with Barbara Maughan of King’s College and Adrian Angold of Duke University) 

Child Development

* Early language problems: outcomes for children and relatives [Wellcome Trust] (AP, EM and Patrick Bolton, University of Cambridge)

* A comparison of institutional and foster care (AP with Penny Roy, City University)

* Early identification of and change among children with autism [US NIMH] (AP with Cathy Lord, University of Chicago)

* Screen and observational identification of children with autism (AP with Michael Rutter, King’s College and Sibel Kazak-Berument of the University of Ankara)

* Psychopathology among children of Brazil [Wellcome] (AP with Robert Goodman, King’s College)

* Life-events and onset of disorder in children (AP with Michael Rutter, King’s College and Seija Sandberg, University of Somewhere in Finland)

 Gender and Employment

* Anglo-Japanese comparison of gender wage differences (YS - see attached page)

* Women in mathematics and science (SC - see attached page)

* Gender and casualisation of employment (VA & SD - see attached page)

 Statistical Analysis 

In addition to our new training programme, we have been developing a number of methods and programmes, primarily under the ERSC’s ALCD Programme 

* GLLAMM Generalized Linear Latent and Mixed Models. This STATA programmeme extends STATA’s existing random effects models to (i) random coefficients (ii) heteroscedasticity (iii) mixed response types (iv) factor loadings making it suitable for multivariate response data (v) and discrete mixture/non-parametric maximum likelihood random effects distributions (AP with Sophia Rabe-Hesketh of King’s College)

* Non-linear and multilevel structural equation models for longitudinal data (using the freeware programme Mx) (SY and AP)

* Stochastic expectation maximisation for complex models of multivariate categorical data using STATA and Mx (SY and AP)

* Mean score and multiple imputation for missing data (AP with David Clayton and David Spiegelhalter of Cambridge MRC Biostatistics Unit and Graham Dunn of University of Manchester)

 Sampling Design and Efficient Use of Multiple Datasets

For reasons as varied as cost, response burden and simple impracticality, study designs that involve deliberately incomplete data collection or build on partial data already available are becoming of increasing importance. In parallel to exploring methodologies for missing data we are examining the patterns of efficiency gains possible from such designs [MRC] (AP & DK) and how it might provide a framework within which quantitative and qualitative research might be brought together.