Assessing L2 listening at a Japanese university: Effects of input type and response format

Abstract

Large-scale and classroom-based tests of L2 listening comprehension typically employ audio-only listening texts and a multiple-choice response format. However, the effect of these test task characteristics on test-taker performance remains elusive, and begs the question as to whether they are the best option in terms of construct validity. Furthermore, the interaction between these test task characteristics and their potential joint effect on performance has not been sufficiently explored. To address this gap, a study was conducted at a Japanese university which investigated the effect of input type (audio-only vs. video) and response format (multiple-choice vs. short answer) on L2 listening test performance. Participants were divided into four groups to take an academic listening test with one of four combinations of input and response format: (1) audio-only with multiple-choice questions; (2) video with multiple-choice questions; (3) audio-only with short answer questions; and (4) video with short answer questions. Results of a 2 x 2 factorial ANOVA revealed a statistically significant effect of response format on test-taker performance. No significant effect for input type was found and no significant interaction among the variables was detected. Results suggest that visual input and audiovisual literacy need to be more clearly articulated within the construct definition of academic listening, if they are to be included at all.