UC Santa Barbara

This paper reports on a study of Chinese speakers who received sustained, systematic online training on English prosody (i.e., pausing, prominence and tone choice) in order to improve their oral speech quality. Twelve students in the experimental group completed the training in four weeks, and another 12 students in the control group did not receive the training. The training modules employed discourse-based materials, and consisted of instruction videos, as well as listening and speaking activities, including learner-created visual pitch contours as instant feedback. The students read a script and gave a one-minute speech on a given topic at the beginning and the end of the study. Four native English speakers judged the speech comprehensibility, fluidity, accent, confidence and attractiveness of the speech. The results show that both groups of students faced challenges of using appropriate prosody when speaking English. The online modules helped the experimental group of students improve on almost all aspects of their script-reading and spontaneous speech. The improvement in comprehensibility and confidence of the spontaneous speech was statistically significant. In contrast, the control group did not show improvement. The students gave high ratings on the helpfulness of the training modules. This study also conducted acoustic analyses of the speech samples, and compared the auditory measures to the acoustic measures of intonation in predicting the speech quality. While both approaches were able to predict some aspects of the speech quality, the auditory approach was able to predict more variables and explain more variances than the acoustic approach. The findings have implications for teaching pronunciation to English learners and developing computer assisted pronunciation training tools.