Factors Associated with Test-takers’ Listening Performance in a Video Listening Comprehension Test: Examining Item and Test-taker Characteristics

Factors Associated with Test-takers’ Listening Performance in a Video Listening Comprehension Test: Examining Item and Test-taker Characteristics

This study examines the degree to which item and test-taker characteristics have an effect on the test-taker’s listening performance in a video-based listening test. A total of 114 participants were allocated into two separate groups, and an identical listening test consisting of retired items of the CSAT English was developed in two different modes (audio-only and video) and was given to each group. The test performance of the two groups was compared by prompt type (dialogue and lecture) and proficiency (low-mid-high). Findings showed that test-takers in all three proficiency groups performed better in video mode, but the differences were not statistically significant. Findings of logistic regression models showed that the significant positive effect of visual input was constant over all initial listening performance levels, but it did not vary by proficiency. Test-takers performed better in the lecture items than in the dialogue items in both conditions. When cross-classified multilevel logistic models were fitted, a significantly larger random effect of the item was found over the testtaker random effect. This study expands our understanding of the factors associated with the listening construct when a video is included in the listening tests.