ABSTRACT
Educators and policy-makers tend to trust particular forms of assessment as being more secure than others. However, there is little in the way of evidence with which to evaluate these claims, and no common metrics to compare different approaches. This chapter argues that we need a variety of metrics to make informed decisions about assessment security. Four types of metrics are proposed, based on: cheating difficulty; cheating detection accuracy; ability to prove instances of cheating; and prevalence of cheating; alongside counterbalance metrics focusing on learning, teaching, assessment and the student experience. The chapter concludes by arguing the need for minimum standards in assessment security, and proposes a hypothetical set of standards as a starting point.
