Animation Videos in a Computer-Based Listening-Speaking Integrated Task: An Eye-Tracking and Mixed-Methods Study on Test-Taking Processes

Computer-based language assessment tasks are becoming more multimodal and integrated. This paper reports on a study on a computer-based listening-speaking integrated task which uses an animation video as a prompt to provide a virtual communication context that requires test takers to summarize what they have heard and offer a solution to a problem in a role-play contract negotiation task. The study aimed to investigate how animation video, a unique type of video and multimodal input, influenced test-taking processes, drawing on two sources of data: test takers’ eye movements when completing the task and post-task stimulated recall interviews. The study found that test takers tended to focus on watching certain elements of the video that could provide contextual information so that they could establish the communication context and purpose; however, test takers’ behaviour of viewing different areas of the video itself did not have significant correlation with test scores, although there were statistically significant negative correlations between test scores and test takers’ viewing of task instructions. Furthermore, according to stimulated recall interviews, it was found that the positive effects of video elements were largely on the higher-order test-taking processes. Implications of the findings are discussed with reference to enhancing communication authenticity of multimodal integrated speaking tasks.