Sample Attrition between Waves 1 and 5 in the European Community Household Panel

Publisher Website
Google Scholar
Abstract
Attrition in the first five waves of the European Community Household Panel is examined. Although overall attrition rates are high in several countries, multivariate analysis revealed that only a small proportion of the total variation in attrition is accounted for by a comprehensive set of substantive variables of the kind likely to be of interest in research on poverty and social exclusion. The loss of respondents is greatest among those adults who did not complete a personal interview in the initial wave (indicating reluctant participation in the survey), and individuals who are geographically mobile. There are also smaller, but statistically significant, differences between countries in the tendency to lose individuals either from the top (the Southern countries and Ireland) or the bottom (the remaining countries) of the income distribution. Nevertheless, even in the absence of compensating attrition weights, attrition has only a minor impact on the estimates of two key social exclusion indicators: the poverty rate and the quintile income share ratio (S80/20). The results indicate that fears that attrition has undermined the representativeness of the ECHP samples in later waves of the survey are largely unfounded.

This publication has 0 references indexed in Scilit: