Conferences

The R-indicator is an instrument that evaluates to what extent a survey is representative of the population under investigation (Schouten & Cobben, 2007). It measures the variation of response propensities, predicted by a set of auxiliary variables that are available for the gross sample. This paper seeks to investigate the usefulness of such a monitoring tool to assess the quality of the contact process, as well as to give some typical examples. It is clear that many agents such as the fieldwork management, interviewers and sample units participate in the fieldwork process, so that many (unintended) decisions are made that possibly affect the representativity of the interviewed sample. Moreover, the contact process can be considered as multi-layered: first a sample unit should be attempted, then contact has to be made, etc. Of course, each of these fieldwork phases may have its own representativity issues. The presentation will focus (1) on the incorporation of paradata in the screening of the sample quality, (2) the development of best practices for the application to data collection monitoring and (3) development of graphical tools for the visualisation of representativity indicators. As an example one may think of the evolution of an R-indicator during the fieldwork. Sudden jumps in the curve may indicate a threat to the quality of the sample.

This paper is part of the RISQ project (Representativity Indicators for Survey Quality) and is financed by the 7th Framework Programme (FP7) of the European Union.

References

Schouten, B., & F., Cobben (2007) R-indexes for the comparison of different fieldwork strategies and data collection modes.