Resumo

RESTREPO ESTRADA, María Isabel  e  MARIN DIAZARAQUE, Juan Miguel. Income Imputation in the 2010 Great Integrated Household Survey (GEIH). Desarro. soc. [online]. 2012, n.70, pp.219-243. ISSN 0120-3584.

The aim of this paper is to present the issue in managing surveys with missing data. In order to address this problem, it is reviewed a technique known as imputation. Some methodologies on the income imputation in the 2010 Great Integrated Household Survey (GEIH 2010) were implemented. Seven methods for the total sample and groups of housing strata were evaluated: listwise deletion, unconditional mean imputation, stochastic regression imputation, hot-deck imputation, regression hot deck imputation, multivariate normal imputation, and multiple imputation by chained equations. It was discovered that the results of the applied methods are relatively similar, since non-response percentages are not high and the missing data may follow an ignorable pattern.

Palavras-chave : Missing data; single imputation; multiple imputation.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )