Health-related non-response in the British Household Panel Survey and European Community Household Panel: using inverse probability weighted estimators in nonlinear models

A.M. Jones; X. Koolman; N. Rice

doi:10.1111/j.1467-985X.2006.00399.x

Health-related non-response in the British Household Panel Survey and European Community Household Panel: using inverse probability weighted estimators in nonlinear models

Research output: Contribution to journal › Article › peer-review

Abstract

The paper considers health-related non-response in the first 11 waves of the British Household Panel Survey and the full eight waves of the European Community Household Panel and explores its consequences for dynamic models of the association between socioeconomic status and self-assessed health. We describe the pattern of health-related non-response that is revealed by the British Household Panel Survey and European Community Household Panel data. We both test and correct for non-response in empirical models of the effect of socioeconomic status on self-assessed health. Descriptive evidence shows that there is health-related non-response in the data, with those in very poor initial health more likely to drop out, and variable addition tests provide evidence of non-response bias in the panel data models of self-reported health. Nevertheless a comparison of estimates—based on the balanced sample, the unbalanced sample and corrected for non-response by using inverse probability weights—shows that, on the whole, there are not substantive differences in the average partial effects of the variables of interest. The main differences are between unweighted and one form of inverse-probability-weighted estimates for the average partial effects of income and education in those countries that have fewer than eight waves of data. Similar findings have been reported concerning the limited influence of non-response bias in models of various labour market outcomes; we discuss possible explanations for our results.

Original language	English
Pages (from-to)	543-569
Number of pages	27
Journal	Journal of the Royal Statistical Society: Series A (Statistics in Society)
Volume	169
Issue number	3
DOIs	https://doi.org/10.1111/j.1467-985X.2006.00399.x
Publication status	Published - Jan 2006

Keywords

attrition
British Household Panel Survey
European Community Household Panel
inverse probability weighting
non-participation
non-response
selection on observables
self-assessed health
INCOME-RELATED INEQUALITIES
SEMIPARAMETRIC REGRESSION-MODELS
SELF-RATED HEALTH
SOCIOECONOMIC-STATUS
SAMPLE ATTRITION
MISSING DATA
POOR HEALTH
DYNAMICS
TIME
MORTALITY

Access to Document

10.1111/j.1467-985X.2006.00399.x

http://dx.doi.org/10.1111/j.1467-985X.2006.00399.x

Cite this

@article{ba863cb2de2e40e9b85cc7c3afd1265f,

title = "Health-related non-response in the British Household Panel Survey and European Community Household Panel: using inverse probability weighted estimators in nonlinear models",

abstract = "The paper considers health-related non-response in the first 11 waves of the British Household Panel Survey and the full eight waves of the European Community Household Panel and explores its consequences for dynamic models of the association between socioeconomic status and self-assessed health. We describe the pattern of health-related non-response that is revealed by the British Household Panel Survey and European Community Household Panel data. We both test and correct for non-response in empirical models of the effect of socioeconomic status on self-assessed health. Descriptive evidence shows that there is health-related non-response in the data, with those in very poor initial health more likely to drop out, and variable addition tests provide evidence of non-response bias in the panel data models of self-reported health. Nevertheless a comparison of estimates—based on the balanced sample, the unbalanced sample and corrected for non-response by using inverse probability weights—shows that, on the whole, there are not substantive differences in the average partial effects of the variables of interest. The main differences are between unweighted and one form of inverse-probability-weighted estimates for the average partial effects of income and education in those countries that have fewer than eight waves of data. Similar findings have been reported concerning the limited influence of non-response bias in models of various labour market outcomes; we discuss possible explanations for our results.",

keywords = "attrition, British Household Panel Survey, European Community Household Panel, inverse probability weighting, non-participation, non-response, selection on observables, self-assessed health, INCOME-RELATED INEQUALITIES, SEMIPARAMETRIC REGRESSION-MODELS, SELF-RATED HEALTH, SOCIOECONOMIC-STATUS, SAMPLE ATTRITION, MISSING DATA, POOR HEALTH, DYNAMICS, TIME, MORTALITY",

author = "A.M. Jones and X. Koolman and N. Rice",

year = "2006",

month = jan,

doi = "10.1111/j.1467-985X.2006.00399.x",

language = "English",

volume = "169",

pages = "543--569",

journal = "Journal of the Royal Statistical Society: Series A (Statistics in Society)",

issn = "0964-1998",

publisher = "Wiley-Blackwell",

number = "3",

}

Research output: Contribution to journal › Article › peer-review

TY - JOUR

T1 - Health-related non-response in the British Household Panel Survey and European Community Household Panel: using inverse probability weighted estimators in nonlinear models

AU - Jones, A.M.

AU - Koolman, X.

AU - Rice, N.

PY - 2006/1

Y1 - 2006/1

N2 - The paper considers health-related non-response in the first 11 waves of the British Household Panel Survey and the full eight waves of the European Community Household Panel and explores its consequences for dynamic models of the association between socioeconomic status and self-assessed health. We describe the pattern of health-related non-response that is revealed by the British Household Panel Survey and European Community Household Panel data. We both test and correct for non-response in empirical models of the effect of socioeconomic status on self-assessed health. Descriptive evidence shows that there is health-related non-response in the data, with those in very poor initial health more likely to drop out, and variable addition tests provide evidence of non-response bias in the panel data models of self-reported health. Nevertheless a comparison of estimates—based on the balanced sample, the unbalanced sample and corrected for non-response by using inverse probability weights—shows that, on the whole, there are not substantive differences in the average partial effects of the variables of interest. The main differences are between unweighted and one form of inverse-probability-weighted estimates for the average partial effects of income and education in those countries that have fewer than eight waves of data. Similar findings have been reported concerning the limited influence of non-response bias in models of various labour market outcomes; we discuss possible explanations for our results.

AB - The paper considers health-related non-response in the first 11 waves of the British Household Panel Survey and the full eight waves of the European Community Household Panel and explores its consequences for dynamic models of the association between socioeconomic status and self-assessed health. We describe the pattern of health-related non-response that is revealed by the British Household Panel Survey and European Community Household Panel data. We both test and correct for non-response in empirical models of the effect of socioeconomic status on self-assessed health. Descriptive evidence shows that there is health-related non-response in the data, with those in very poor initial health more likely to drop out, and variable addition tests provide evidence of non-response bias in the panel data models of self-reported health. Nevertheless a comparison of estimates—based on the balanced sample, the unbalanced sample and corrected for non-response by using inverse probability weights—shows that, on the whole, there are not substantive differences in the average partial effects of the variables of interest. The main differences are between unweighted and one form of inverse-probability-weighted estimates for the average partial effects of income and education in those countries that have fewer than eight waves of data. Similar findings have been reported concerning the limited influence of non-response bias in models of various labour market outcomes; we discuss possible explanations for our results.

KW - attrition

KW - British Household Panel Survey

KW - European Community Household Panel

KW - inverse probability weighting

KW - non-participation

KW - non-response

KW - selection on observables

KW - self-assessed health

KW - INCOME-RELATED INEQUALITIES

KW - SEMIPARAMETRIC REGRESSION-MODELS

KW - SELF-RATED HEALTH

KW - SOCIOECONOMIC-STATUS

KW - SAMPLE ATTRITION

KW - MISSING DATA

KW - POOR HEALTH

KW - DYNAMICS

KW - TIME

KW - MORTALITY

U2 - 10.1111/j.1467-985X.2006.00399.x

DO - 10.1111/j.1467-985X.2006.00399.x

M3 - Article

SN - 0964-1998

VL - 169

SP - 543

EP - 569

JO - Journal of the Royal Statistical Society: Series A (Statistics in Society)

JF - Journal of the Royal Statistical Society: Series A (Statistics in Society)

IS - 3

ER -