The present thesis employed a mixed-methods research design spanning over two studies and an intermediate instrument development step to investigate the interactions between the presence of nonverbal and other visual cues with examinees, individual items, and item task types in video listening tests. The first study employed eye-tracking methodology to determine the specific visual cues to which examinees attend when interacting with a video-mediated listening test. The findings of Study I then informed the development of a new video-mediated listening test for Study II, which investigated the effect of the presence of visual cues on item and item task difficulty via many-facet Rasch modeling and qualitative item analysis. Individual examinee differences were also explored in both studies with respect to gender, proficiency, and perceptions of the two formats.

Study I found that examinees view facial cues an average of 81% of the time regardless of task type, but spend significantly more time oriented toward the listener (i.e., the character who is not speaking) when presented with an implicit item. Direct viewing of gestures, despite their prominent place in the nonverbal communication literature, only accounted for approximately 1.35% of the total video time. Study II found that the presence of visual information exerted a facilitative effect on all items, but that it was significantly more pronounced with implicit items, despite the fact that they were more difficult than the explicit items in the audio format. Finally, no substantive interactions between proficiency, gender, or perception of the formats were observed between either viewing behavior or the facilitative effect of video.

The thesis ultimately raises questions about the usefulness of video listening tests for listening comprehension assessment, as the effect appears to do little more that raise scores, and, in the case of implicit items, may obviate the need to comprehend the linguistic input in order to make correct inferences.