Transition from computer-based testing of national benchmark tests to adaptive testing: Robust application of fourth industrial revolution tools

Abstract

In 2020, the COVID-19 pandemic strongly affected all sectors, including education. Across the globe, governments enacted policies restricting people's movement, affecting the educational assessment industry. During the lockdown, South Africa's National Benchmark Tests (NBTs) were cancelled. In contrast to paper-pencil and linear tests deployed for NBTs, computer-adaptive testing (CAT) is a modern alternative. CAT technology is widely used for licensing and certification exams in most developed nations. CAT is expected to be used by many educational assessment companies in sub-Saharan Africa for their high-stakes exams due to its reduced testing time and algorithms that produce stable and reliable results. This paper aims to provide an overview of the 4IR tools that will enable creation of a comprehensive CAT, such as feasibility studies, item bank development, test-and-calibrate item banks, and specifications for the final CAT. To successfully implement CAT, these steps are essential. The importance of thorough research and documentation in each stage of the CAT cannot be overemphasized. Based on Fourth Industrial Revolution (4IR) artificial intelligence tools, we conclude that CAT is an excellent modality to adopt for ensuring accurate evaluations of examinees' abilities within high-stakes exams.      

Keywords: CAT framework, Computerised adaptive testing, Fourth industrial revolution, Linear testing, National benchmark tests