Theoretical sampling designs for a UK birth cohort with potential accelerated design
Theoretical sampling designs for a UK birth cohort with potential accelerated design
Executive Summary
Accelerated Longitudinal Designs (ALDs) have several advantages over single cohort designs, specifically the ability to estimate and adjust for cohort effect differences and produce results in a shorter period. These are traded off against additional model complexity in analysis and challenges in recruitment of additional sample cases in the field.
There are several properties of ALDs which can be varied; analyses based on the power to detect cohort differences in a linear model for age effects show that small numbers of cohorts have the best chance to improve the properties of the design and speed up data availability across the age range of interest while remaining practical to implement.
Further work is needed on methods to determine sample sizes to meet variance constraints for a range of analytical parameters, and to determine how best to produce a general purpose design which can meet a wide range of user needs.
University of Southampton
Smith, Paul A.
a2548525-4f99-4baf-a4d0-2b216cce059c
Dawber, James
85c7c036-2ae3-4c57-a8b3-9f5223cd4da6
Van Der Heijden, Peter
85157917-3b33-4683-81be-713f987fd612
July 2019
Smith, Paul A.
a2548525-4f99-4baf-a4d0-2b216cce059c
Dawber, James
85c7c036-2ae3-4c57-a8b3-9f5223cd4da6
Van Der Heijden, Peter
85157917-3b33-4683-81be-713f987fd612
Smith, Paul A., Dawber, James and Van Der Heijden, Peter
(2019)
Theoretical sampling designs for a UK birth cohort with potential accelerated design
University of Southampton
13pp.
Record type:
Monograph
(Project Report)
Abstract
Executive Summary
Accelerated Longitudinal Designs (ALDs) have several advantages over single cohort designs, specifically the ability to estimate and adjust for cohort effect differences and produce results in a shorter period. These are traded off against additional model complexity in analysis and challenges in recruitment of additional sample cases in the field.
There are several properties of ALDs which can be varied; analyses based on the power to detect cohort differences in a linear model for age effects show that small numbers of cohorts have the best chance to improve the properties of the design and speed up data availability across the age range of interest while remaining practical to implement.
Further work is needed on methods to determine sample sizes to meet variance constraints for a range of analytical parameters, and to determine how best to produce a general purpose design which can meet a wide range of user needs.
Text
WP6 Theoretical sampling designs for a UK birth cohort with potential accelerated design
More information
Published date: July 2019
Identifiers
Local EPrints ID: 435299
URI: http://eprints.soton.ac.uk/id/eprint/435299
PURE UUID: 8d0bb8a6-6e0b-465a-81ca-cf1dee225544
Catalogue record
Date deposited: 30 Oct 2019 17:30
Last modified: 16 Apr 2024 01:46
Export record
Contributors
Author:
James Dawber
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
