Thesis (M.Sc.) - Cairo University - Faculty of Economics and Political Science - Department of Statistics

Many survey organizations pay a great attention to survey quality and focus on the response rate as being the quality indicator for the impact of non-response bias. However, Response rate is an insufficient indicator of response quality. A higher response rate may not imply a smaller non-response bias. In some cases Response Rate could be a weak indicator for survey quality; and therefore there is a need for additional indicators that provide more insight in the possible risk of biased estimators. Those indicators should measure how representative the survey response is. A good quality indicator should reflect how well the composition of the survey response represents the population (or complete sample). Different representativity indicators are offered in the literature and are known by R-indicators for short. R-indicators are not only applied in establishing the quality of registered data as supplementary indicator to Response Rate, but it could be also used in many different ways (data collection, comparing survey data quality for different surveys that share the same target population, comparing surveys quality over time and controlling the survey process through Representativity analysis for a previous version of the survey. Representativity indicators have a great importance in monitoring or evaluating survey quality. At the international level, there are some papers that focused on the Representativity indicators for measuring survey quality

Issued also as CD