Poverty comparisons with dependent samples

Abstract

Standard inference procedures for poverty comparisons require samples to be independent. For many commonly used income samples, however, this requirement is not fulfilled since samples are rotated. This article introduces an easy-to-use method of correction for sample dependency. We also apply the method to test changes in US poverty in the 1990s and to evaluate the marginal effects of public assistance on poverty before and after the recent welfare reform. Copyright © 2004 John Wiley & Sons, Ltd.

The Foster–Greer–Thorbecke (FGT) poverty measures: 25 years later

“…Complementing these approaches are tests that explicitly account for the sampling design of survey data. Kakwani [79], Cowell and Victoria-Feser [37], and Zheng [142] among others, have modified the estimation techniques to account for the difficulties that stem from using sample data rather than census data.…”

Accounting for trends in health poverty: a decomposition analysis for Britain, 1991–2008

“…The dependence of two BHPS samples taken from two different survey waves is only partial owing to sample attrition and inclusion of new entrants after wave 1 (see [ 13 ]). An appropriate method of accounting for partial sample dependency was proposed by Zheng [ 16 ] in the context of the inference for continuous additively separable poverty measures (including the continuous FGT indices). 5 In this paper, we use Zheng’s [ 16 ] approach to calculate the covariance term in Eq.…”

Accounting for Trends in Health Poverty: A Decomposition Analysis for Britain, 1991-20088

“…The dependence of two BHPS samples taken from two different survey waves is only partial due to sample attrition and inclusion of new entrants after wave 1 (see Taylor et al 2010). An appropriate method of accounting for partial sample dependency was proposed by Zheng (2004) in the context of the the inference for continuous additively separable poverty measures (including the continuous FGT indices). 4 In this paper, we use Zheng's (2004) approach to calculate the covariance term in (6).…”