Exploring a Decade of Research: A Systematic Review of Computer-Based English Speaking Tests

Abstract

The rapid integration of technology into educational assessment has revolutionized the evaluation of English speaking proficiency. Computer-based English speaking tests (CBESTs) have emerged as scalable and efficient solutions, which offer enhanced consistency and accessibility in high-stakes and large-scale testing contexts. However, existing studies on CBESTs have primarily focused on specific aspects of their design, implementation, and impact, leaving a fragmented understanding of their broader implications. As such, this systematic review synthesizes empirical research on CBESTs published between 2014 and 2024 to identify overarching trends, challenges, and opportunities. Employing the PRISMA methodology, the review analyzed 36 studies identified from three databases: Web of Science, Scopus, and Google Scholar. The findings highlight diverse research foci, including advancements in automated scoring, test validity, and the influence of cognitive and affective factors on performance. Studies also explored test-taker perceptions and experiences, which revealed mixed attitudes toward fairness and authenticity. Research methodologies ranged from quantitative correlational studies and qualitative case studies to mixed-methods designs, reflecting a diverse yet fragmented body of work. The review highlights the need for continued innovation in CBEST design and emphasizes the importance of hybrid models that integrate automation with human judgment. For test developers and policymakers, the findings underscore the importance of equitable implementation, technical refinement, and alignment with pedagogical goals. Future research should explore underrepresented areas such as long-term learning impacts and broader inclusivity to enhance the utility and fairness of CBESTs.