Assessing L2 speaking in virtual worlds: Validity, mode effects, and implications for group oral proficiency testing

The purpose of this study is to explore the potential of virtual worlds(VWs) as a platform for assessing second language (L2) oral proficiency. Specifically,it examines VW group oral tests in comparison to face-to-face (F2F) tests in termsof concurrent validity and face validity. Data were collected from 64 ESL learners' grouporal test scores, surveys on their perceptions of both testing modes, observer field notes,and interviews. Results show that while a mode effect was observed in the firstadministration, students' scores in the VW condition became comparable to those inthe F2F condition when the test was repeated. Students' perceptions of the VW modewere also positive in relation to reduced anxiety and ease of use, although turn-takingwas reported to be more complex in VW than in F2F tasks. Qualitative analyses furtherrevealed that VWs provide several advantages, including more engaging and less stressfultesting conditions as well as a stronger sense of copresence. Taken together, thesefindings suggest that VW-based assessments can complement traditional testing byoffering more flexible and accessible options for evaluating oral proficiency, particularlyin distant or online learning contexts. The pedagogical implications include the needto account for interactional challenges such as turn-taking, while also taking advantageof the unique affordances of VWs to create more engaging, scalable, and learner-centeredassessment environments.