Title: Improving the item sum technique using auxiliary information in complex surveys
Authors: Pier Francesco Perri - University of Calabria (Italy) [presenting]
Beatriz Cobo - University of Granada (Spain)
Abstract: To collect sensitive data, survey statisticians have developed many approaches and strategies to reduce the rate of nonresponse and social desirability response bias. In the last years, the item count technique has been largely employed as alternative indirect questioning survey mode for qualitative characteristics and some variants have been proposed to face with new needs and challenges. The item sum technique (IST) is a recent variant introduced to estimate the mean of a sensitive quantitative variable when sampled units are asked to confront themselves with a two-list of items containing a question on the sensitive character under study and a number of innocuous questions. To the best of our knowledge, very few theoretical and applied works exist in the field of the IST. We, therefore, intend to discuss some methodological advances in order to spread the technique in real surveys. In particular, we discuss, under a generic sampling design, the problem of how to improve the estimates of the sensitive mean when auxiliary information on the population under study is available and used at the design and estimation stages. An Horvitz-Thompson type estimator and a calibration estimator are proposed and their efficiency evaluated on the basis of a simulation study performed on real data from the ``World Bank's Enterprise Surveys''. It is shown that estimates obtained by supposing that data are collected by the IST are nearly equivalent to those obtained using the true data.