Evaluation of Linguistic Labels Used in Applications

Linguistic labels such as high, medium, and low are commonly used in different applications. Researchers in psychometrics argue that before adding new labels to applications, the labels must be empirically evaluated. In this paper, we explain the process of selecting labels for a security assessment application. We also show how we evaluate the labels empirically using a sample population from Amazon Mechanical Turk users. This research is supported by grants from the National Security Agency and King Abdul-Aziz University.