Digital?first assessments: A security framework

Background Digital-first assessments leverage the affordances of technology in all elements assessment process: from design and development to score reporting evaluation create test taker-centric assessments. Objectives The goal this paper is describe engineering, machine learning, psychometric processes technologies a security framework (part larger ecosystem; Burstein et al., 2021) that can be used systems protect integrity scores. Methods We use Duolingo English Test exemplify are presented. This includes methods for actively detecting deterring malicious behaviour (e.g., custom desktop app). It also passively large item bank created through automatic generation methods). certification process each administration undergoes, which both automated human review. Additionally, we our quality assurance dashboard leverages data mining techniques monitor inform decisions about pool maintenance. Results Conclusions As developers transition online delivery approach places taker at centre, it becomes increasingly important take advantage tools methodological advances different fields psychometrics). These essential maintaining so reliability sustained interpretations uses scores remain valid.