Shadowing and L2 Fluency: Is Shadowing Practice Effective for Preparing for Speaking Tests?

In recent years, Japanese high school students have been encouraged to take commercial English proficiency tests measuring four language skills. This demand indicates that Japanese students need to develop not only receptive but also productive skills in English. To date, effective methods to develop L2 fluency in EFL contexts have not been fully explored. In this respect, Kadota (2015, 2018, 2019) argues shadowing functions as a meaningful practice that connects input processing with output processing. The present study examines to what extent shadowing practice helps Japanese learners of English develop their speaking abilities. The participants of the study were nine Japanese college students. First, they performed four types of speaking tasks (shadowing, oral reading, a picture-description task, and an opinion-expression task). Then, they learned how to use Xreading, an online extensive reading program, for shadowing practice. They were instructed to listen to and shadow stories of their choice at least twice a week, for 15 minutes each time. They worked at their own pace for four weeks. Then, they completed the same tasks with different versions. The speaking tasks were analyzed in the following way: for the shadowing task, the number of syllables accurately reproduced was counted; for oral reading, the time to read the passage was calculated and it was transformed into words per minute; for the picture-description and opinion-expression tasks, complexity (words per clause), accuracy (the ratio of the accurate clause to the total number of clauses), and fluency (speech rate and mean length of runs [MLRs] ) were analyzed. To find statistical differences in the paired samples, all data set were submitted to the Wilcoxon signed ranks test. Furthermore, to examine how the hours of shadowing practice were related to speech performances in the second task, Spearman's rank-order correlations were performed. Major findings were as follows: (1) the participants' ability to shadow significantly improved with a large effect size; (2) oral reading skill improved significantly, but with a small effect size; (3) the complexity of speech production did not change significantly; (4) the accuracy decreased in the picture-description task but it increased in the opinion expression task, both with medium effect sizes; (5) speech rate and MLRs did not significantly improve in the picture-description task; however, those improved in the opinion task with large effect sizes; (6) there was a relatively strong correlation between shadowing time and accuracy in the second picture-description task. These results seem to indicate that shadowing is especially effective in facilitating the accurate reproduction of sound and articulation speed.