MALPRACTICE IN ONLINE VERSUS ONSITE COMPUTER-BASED LANGUAGE TESTS Reflections from the COVID lockdown experience

Computer-based tests provide a fast, efficient and cost effective means of assessing students, especially if they can be administered online (Chalhoub-Deville 2002; Suvorov, Hegelheimer 2013). Traditionally, they are administered in controlled conditions, where supervision and invigilation (or proctoring) can be conducted in much the same way as it is with paper-based tests. Around the world, the COVID emergency (2020-21) led to the sudden closure of educational institutions and a shift online of all teaching and testing. In this paper, we will examine the effects of such a change in the context of two specific university courses where a set of bespoke Moodle-hosted computer-based (CB) tests were adapted from a controlled and supervised onsite context to a much freer online context using a general business communication platform (Microsoft Team). We compare the results of over 4,500 tests (Lexis and Use of English, Translation, Reading, Listening and Writing Tasks: the first four marked automatically, the last manually) taken in both onsite and online contexts to identify differences in results between both testing environments. The aim is to provide some hard data to ascertain how far online distance-invigilated tests taken under only the most basic supervision can be adequate substitutes for more traditional formats especially in view of both their positive aspects (e.g. greater accessibility and inclusivity for students) and their negative ones (e.g. the greater opportunity for malpractice).