Using video in college English listening tests aims to improve test validity, however, whether test-takers watch the video determine whether such intention can be realized. After analysing the recording of test-takers’ eye movement, follow-up questionnaire and interview notes, we find that when answering multiple choice questions, test-takers tend to watch video, while in completing blank-filling questions, they opt not to watch it. Further analysis on test-takers eye movement reveals that factors, including test design, audio and video features, influence test-takers’ viewing patterns. Finally, after analysing the questionnaire and interview data, we summarize the types of visual messages that test-takers really pay attention to and their functions.