A comparison of speech samples of monologic tasks in speaking tests between computer-delivered and face-to-face modes

Abstract

The present study attempted to discover plausible differences in examinees' linguistic performance on monologic tasks between computer-delivered and face-to-face modes. The study also examined the relationship between test mode and examinees' oral proficiency. Seventy-nine Japanese EFL students responded to two monologic tasks delivered by the two modes. The examinees' speech samples were compared on measures of fluency, accuracy, and complexity. Results indicated that significant but varied effects were produced by the two test modes only in terms of fluency: examinees used more repetition words in the face-to-face mode but more filled pauses in the computer mode. Further, the effects of test mode on speech samples were not found to relate to examinees' proficiency. Implications for the field of language assessment and second language acquisition research were discussed, and directions for future studies proposed.