Modelling multilevel data under complex sampling designs: an empirical likelihood approach
Modelling multilevel data under complex sampling designs: an empirical likelihood approach
Data used in social, behavioural, health or biological sciences may have a hierarchical structure due to the population of interest or the sampling design. Multilevel or marginal models are often used to analyse such hierarchical data. These data are often selected with unequal probabilities from a clustered and stratified population. An empirical likelihood approach for the regression parameters of a multilevel model is proposed. It has the advantage of taking into account of the sampling design. This approach can be used for point estimation, hypothesis testing and confidence intervals for the sub-vector of parameters. It provides asymptotically valid inference for small and large sampling fractions. The simulation study shows the advantages of the empirical likelihood approach over alternative parametric approaches. The approach proposed is illustrated using the Programme for International Student Assessment (PISA) survey data.
Design-based inference, generalised estimating equation, regression coefficient, two-stage sampling, unequal inclusion probability, uniform correlation structure
1-16
Oguz Alper, Melike
02d5ed8a-e9e3-438a-95c0-709acd83a5f8
Berger, Yves
8fd6af5c-31e6-4130-8b53-90910bf2f43b
May 2020
Oguz Alper, Melike
02d5ed8a-e9e3-438a-95c0-709acd83a5f8
Berger, Yves
8fd6af5c-31e6-4130-8b53-90910bf2f43b
Oguz Alper, Melike and Berger, Yves
(2020)
Modelling multilevel data under complex sampling designs: an empirical likelihood approach.
Computational Statistics and Data Analysis, 145, , [106906].
(doi:10.1016/j.csda.2019.106906).
Abstract
Data used in social, behavioural, health or biological sciences may have a hierarchical structure due to the population of interest or the sampling design. Multilevel or marginal models are often used to analyse such hierarchical data. These data are often selected with unequal probabilities from a clustered and stratified population. An empirical likelihood approach for the regression parameters of a multilevel model is proposed. It has the advantage of taking into account of the sampling design. This approach can be used for point estimation, hypothesis testing and confidence intervals for the sub-vector of parameters. It provides asymptotically valid inference for small and large sampling fractions. The simulation study shows the advantages of the empirical likelihood approach over alternative parametric approaches. The approach proposed is illustrated using the Programme for International Student Assessment (PISA) survey data.
Text
Oguz Berger 2020
- Accepted Manuscript
More information
Accepted/In Press date: 21 December 2019
e-pub ahead of print date: 28 December 2019
Published date: May 2020
Additional Information:
Funding Information:
Melike Oǧuz-Alper was supported by the Economic and Social Research Council, United Kingdom . The authors wish to thank the reviewers and Professor Li-Chun Zhang (University of Southampton) for helpful comments. They are also grateful to an anonymous reviewer who suggested adding Section 6 . Appendix A
Publisher Copyright:
© 2020 Elsevier B.V.
Keywords:
Design-based inference, generalised estimating equation, regression coefficient, two-stage sampling, unequal inclusion probability, uniform correlation structure
Identifiers
Local EPrints ID: 437069
URI: http://eprints.soton.ac.uk/id/eprint/437069
ISSN: 0167-9473
PURE UUID: 72b10ae2-18c2-4ef1-b043-9c1dac335ec5
Catalogue record
Date deposited: 16 Jan 2020 17:32
Last modified: 17 Mar 2024 05:12
Export record
Altmetrics
Contributors
Author:
Melike Oguz Alper
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