A Study Of The Effect Of Year-Two English Majors' Oral Proficiency On Self-Repairs In Testing Context

The present study investigates Chinese Year-Two English majors' self-repairs in their story retelling in testing context, describing the general characteristics of self-repairs and exploring the influence of the test takers' oral proficiency on self-repairs. To be specific, this study addresses the two major questions: 1) What are the general characteristics of self-repairs in Chinese EFL learners' oral production in testing context? 2) How do different types of L2 speakers' self-repairs vary across the different levels of their oral proficiency?Materials employed in the present study were taken from "The Spoken English Corpus of Chinese Learners", known as SECCL, with 30 EFL learners of three different proficiency levels involved. The data analysis generated the following major findings:Firstly, the frequency of repairs in three different groups' narrative production accounts for 1.95%, in other words, the test takers made 1.95 repairs every 100 words in their story retelling. Of the three types of repairs identified in the material, form repairs constitute the highest percentage (68.11%), followed by content repairs (26.49%) and appropriacy repairs (5.41%). Of the form repair subcategories, grammatical repairs are more frequent than lexical ones, 76.98% and 23.02% respectively.The results indicate that form repairs tended to be more frequently made than content repairs, and that, of the two categories of form repairs, grammatical repairs constitute an obviously higher percentage than lexical repairs, which suggests that the test takers allocated more of their attentional resources to linguistic form than to content, considering the nature of the story retelling task in which the test takers would not have too much trouble with the content of the story they had listened to.Of all the repairs identified, 87.03% were labeled as successful, which partiallysupports Krashen's argument that the high success rate of self-corrections does not indicate the effectiveness of monitoring, but L2 learners' attempt to repair mistakes they find easy to correctSecondly, correlation analysis shows that the participants' oral proficiency is correlated positively with the frequency of the total repairs and that of form repairs, indicating that lower level speakers made more self-repairs in their speech production.One-way ANOVA analysis generated the following results: 1) Top, middle and bottom group display significant difference in the frequency of repairs regardless of the types. 2) The three groups show significant difference in the frequency of form repairs, but not in that of content and appropriacy repairs.Thirdly, the three types of repairs also vary across the three groups in terms of distribution and effectiveness. Of all the repairs, form repairs by the middle group account for the highest percentage. The possible explanation for this is that, the top group made fewer form repairs in their retelling, while the bottom group was frustrated in their attentional allocation to both the content of the story and the form correction.Interestingly, the effectiveness of self-repairs among the three groups indicates that the middle group (87.93%) and the bottom group (88.61%) were more successful than the top group (83.33%) in self-repair. This might be due to the varying degree of difficulty of the mistakes they repaired; the top group attempted to repair more difficult items, compared to the other two groups.The findings from this study have two pedagogical implications: 1) L2 learners are more sensitive to linguistic accuracy in their speech processing. Therefore, in oral language training, language instructors should not only encourage the learners' attempt to repair the mistakes they detect, but also raise their awareness of information organization. 2) In this study, high correlation between proficiency and self-repairs suggests that the learners of different levels should be given tasks with varying degree of difficulty in classroom instruction. The oral tasks should be designed following thedevelopment of the learners' interlang...