Joshi, Heather and Plewis, Ian and Cullis, Andy and Sadigh, Mahmood and Mojaddad, Mos and Dodd, Katie and Woods, Joanne and Chapman, Aliy (2002) Mobile Families and Other Challenges in the Design of the Millennium Cohort Study. Centre for Longitudinal Studies, Institute of Education, University of London, London. ISBN 1-898453-50-0
Joshi2002MobileCLSWP5.pdf - Other
Download (292Kb) | Preview
The Millennium Cohort Study continues the tradition of Britain’s national longitudinal birth cohort surveys, but it has broken the design mould of being based on all births in a week. The specification that it should cover a year’s births provided the opportunity for some innovative design. This paper describes the disproportionately stratified design, based on a sample of electoral wards and the practicalities of implementing it. The need to sample offered the possibility of over representing populations of particular interest: children living in economically deprived areas, in areas where minority ethnic groups are concentrated, and in the smaller countries of the UK. Eligible families in sampled wards are selected from records held by the Department for Work and Pensions, who also subsequently administer an opt-out to families in sampled areas. Although the final analysis of the sampling process (and the substantive results) will not be ready until 2003, it will be possible to present an interim report on the experience of drawing the sample in this way, focusing particularly on England and Wales, where the survey started earlier. The number of children identified through the administrative records will be compared with those recorded in Birth Registration records, as far as ward boundary changes allow. The cases withdrawn before a sample can be issued to field are analysed by type of ward and type of claimant. A preliminary assessment of the coverage of the Child Benefit Register is offered in a comparison with data from vital registration. This exercise has so far only covered those wards in England where there have been no boundary changes between those used to define the sample in 1998 and those applying in 2000-2001 when the children were actually born. The extent of mobility in and out of target areas by families with children aged 6-9 months is also to be discussed. The strengths of using the Child Benefit register as a sampling frame have been confirmed. The difficulties have in some ways been less than anticipated, and the requirement of a geographically defined sample would have formed a challenge whatever method of sampling had been adopted.
|Divisions:||IOE Departments > Departments > Centre for Longitudinal Studies|
|Depositing User:||IOE Repository Editor (2)|
|Date Deposited:||10 Jul 2010 14:44|
|Last Modified:||29 Jan 2015 10:22|