A New Version of the Item Count Technique for Asking Sensitive Questions: Testing the Performance of the Person Count Technique

Abstract

This paper presents empirical evidence on a recent advancement of the item count technique (ICT, a survey technique for asking sensitive questions), namely, the person count technique (PCT; Grant, Moon, & Gleason, 2014). PCT utilizes person lists instead of lists of filler questions, as is the case in the classic ICT design. This simplifies the questioning procedure, but leads to some methodological challenges such as floor and ceiling effects. The main part of this paper presents empirical evidence stemming from an experimental postal survey in Germany (N = 580) investigating how well PCT performs as compared to standard direct questioning (DQ) with regard to alleviating misreporting for questions on attitudes towards refugees. PCT prevalence estimates for hostile attitudes towards refugees are significantly higher than DQ estimates for one item, and non-significantly higher for three items. Although not consistently significant, the differences are substantial, amounting to a threefold increase of the proportion of respondents expressing negative attitudes towards refugees. Even though the findings are not unequivocally in favor of PCT, this new ICT variant still deserves consideration in the future and warrants further development. Specifically, more knowledge is required with respect to its statistical properties and the best practices of its implementation.