Abstract
The traditional paradigm of language assessment, historically dominated by psychometric paper-and-pencil testing, is increasingly falling out of alignment with the communicative and technological realities of twenty-first-century language learners. Conventional language testing methods often treat language proficiency as a static set of grammatical rules and lexical items evaluated under artificial, high-stakes time constraints. This static approach creates significant construct underrepresentation; it fails to measure the dynamic, multimodal, and interactive nature of contemporary digital communication.
References
digital assessment, computer-adaptive testing (CAT), formative evaluation, learning analytics, automated essay scoring (AES), automated speech recognition (ASR).