Large sampling errors when using the Unmatched Count Technique to estimate prevalence: A simulation study

Abstract

The Unmatched Count Technique (UCT) is a method for ensuring respondent anonymity and thereby providing an unbiased estimate of the prevalence of a characteristic in a population. I illustrate that under realistic conditions UCT estimates can have ten times more sampling error than estimates derived from direct questions, and that UCT estimates can take nonsensical negative values. Therefore, the UCT should be used with caution.