The present research investigated the impacts of L2 proficiency and gender on the facilitative effect of the presence of visuals in video-mediated L2 listening comprehension tests. An 18-item, multiple-choice English listening test was constructed and administered in two forms: ODD and EVEN. The ODD form presented the listening passages of the odd items with video, and the even items with audio only. The EVEN form reversed this pattern. Participants were 279 university students of English in Japan. The participants were randomly assigned to the test forms. The facilitative effect of the presence of visuals was investigated using bias analyses via many-facets Rasch modeling. TOEFL ITP scores were used to investigate the interaction between English proficiency and format (video or audio-only). Large differences of the size of the effect of the presence of visual cues based on individual differences would be indicative of differential item functioning (DIF), which would be a threat to the validity of video-mediated L2 listening tests. Although the video versions of the test items were easier than the audio versions in most cases, this facilitative effect of the presence of visuals did not significantly differ by proficiency level or gender.

研究論文