Authors
Abstract
Household surveys remain a key data source for measuring living standards and inequality. Survey data on incomes are known to suffer from reporting and non-response errors. In this paper, we implement and test a method that attempts to reduce errors directly in the data collection process, rather than mitigating them econometrically at the analysis stage. We experimentally test two versions of an Editable Summary Screen. We find that both versions elicit improvements in data quality that are large enough to change measured inequality. The experiment also provides new evidence on the nature of errors in survey income data.
Subjects
Link
https://www.understandingsociety.ac.uk/research/publications/525973
Paper download