Corrective Feedback on Pronunciation: Students’ and Teachers’ Perceptions

The aim of this research was to find out similarities and differences between teacher and student perceptions of corrective feedback (CF) on pronunciation for students’ presentations in advanced English class through a group interview and a questionnaire survey. Both teachers and students agreed that CF is not only important but necessary since junior and senior students still have pronunciation problems and the best time to provide CF is soon after presentation. However, they differed in concern about students’ self-respect, the types of errors that should receive CF and preference for the types of CF. In particular, students’ eagerness to learn exceeded their concern about self-respect. Teachers turned to offer CF to repeated errors, while students would like to receive more than teachers could offer. Moreover, teachers regarded prompt as being more effective, whereas students preferred recast to prompt considering the latter to be more demanding though they held similar views about explicit correction. It is suggested that taking into consideration both teachers’ and students’ perceptions of CF would help improve senior and junior students’ pronunciation.


Introduction
To an English major, a good English pronunciation is of great importance.However, some students no longer pay attention to their pronunciation when they reach the third or the fourth year.Meantime, foreign language schools of most universities attach less importance to their juniors and seniors' English pronunciation as well.In other words, though English major students generally receive systematic instructions in English pronunciation in their first year of university studies, their pronunciation and spoken English courses are limited.As a result, their problems with pronunciation cannot fully be solved in teaching.When they reach their third and fourth years, they no longer have pronunciation classes in which they have opportunities to be corrected by their teachers.The presentation section of advanced courses is, therefore, believed to be the most desirable moment for students to get help from teachers.
However, some teachers who are responsible for senior students' advanced courses think that at this stage of their university studies, students should focus more on their advanced learning rather than on their basic language skills since the number of hours for each advanced course is limited.In other words, teachers themselves have neither time nor energy to deal with students' pronunciation.When students are asked to make a presentation, the main purpose is to develop students' ability to study after class, such as finding information, digesting what have been taught in class, and finally presenting their work before the public during class.For these reasons, pronunciation is considered the last thing to be considered.
So, the question is: Should English teachers continue to give their students feedback if they find their students' pronunciation problems?
Corrective feedback (CF) has been practiced in many English as a Second Language (ESL) or English as a Foreign Language (EFL) classes.A large number of studies on CF have been conducted in morphosyntactic aspect (e.g., Lyster, 1998;Mackey et al., 2000;Kim & Han, 2007), grammatical aspect (e.g., DeKeyser, 1998;Ammar, 2003;Lyster, 2004;Ellis, 2007;Yang & Lyster, 2010;Sheen, 2011), lexical aspect (e.g., Mackey & Goo, 2007;Dilans, 2010;Elgort, 2011), and pragmatic aspect (e.g., Joan & Kaya, 2006;Takimoto, 2008;Nipaspong & Chinokul, 2010;Nguyen et al., 2012).Research has also been found in the area of phonology (e.g., Zhao, 1997;Jensen & Vinther, 2003;Sheen, 2004Sheen, , 2010;;Sato & Lyster, 2012;Saito & Lyster, 2012a, 2012b;Li, 2012;Lyster et al., 2013), however, few studies, if any, have been found to investigate teacher and student perceptions of corrective feedback on pronunciation, for students' presentations in particular.Therefore, it is of top priority to investigate what perceptions teachers and students have about corrective feedback on students' pronunciation so as to find solutions to the current problems of pronunciation teaching.Such research is of pedagogical importance because in the first place, by studying this topic teachers as well as students would pay more attention to English pronunciation teaching in advanced classes.In the second place, being aware of teachers' and students' different perceptions of the same problem, ESL or EFL departments would attempt to find effective measures to tackle students' pronunciation problems.

Previous Studies
Research on CF first initiated in the 1970s (Long, 1977) based on the idea that learners benefit from teachers' feedback on errors they commit when they fail to figure out the correct form during communication (Lyster & Mori, 2006, p. 270).Then, corrective feedback was practiced frequently in the field of education and language learning, especially in immersion classrooms (Note 1).Lyster & Ranta (1997) carefully categorized corrective feedback into six categories, including recast, explicit correction, elicitation, metalinguistic clues, clarification requests and repetition (Note 2).Later, more studies have focused on the effect of three kinds of correct feedback: recast, prompt, and explicit correction (Mori, 2002;Tsang, 2004;Lyster & Mori, 2006).
By definition, recast involves teacher's reformulation of all or part of a student's utterance, minus the error (Lyster & Ranta, 1997, p. 46).In other words, the teacher corrects his student's error implicitly by repeating the correct form so that the student can instantly compare his answer with his teacher's correction and repeats his utterance in a correct way.
Prompt, on the other hand, means the use of a variety of signals -other than alternative reformations -that push learners to self-repair (Lyster & Mori, 2006, p. 217).Whatever signals are used, "they have one crucial feature in common: They withhold correct forms (and other signs of approval) and instead offer learners an opportunity to self-repair by generating their own modified response (Lyster, 2004, p. 405).
Explicit correction refers to "the explicit provision of the correct form.As the teacher provides the correct form, he or she clearly indicates that what the student said was incorrect" (Lyster & Ranta, 1997, p. 46).That means that the teacher would point out students' errors, offer the correct forms and provide explanations.
Over the past decades, a series of research were conducted to investigate the effects of these types of corrective feedback.For example, Ammar (2003) compared the effects of recasts and prompts in form-focused instruction in three grade six intensive ESL classrooms for four weeks with third-person possessive determiners in English (his and her) as an intervention target.Results show that the two feedback groups (the recast group and the prompt group) outperformed the control group in immediate and delayed oral posttests, although the latter did better on written posttests.Lyster (2004) also investigated the differential effects of recasts and prompts on grammar gender in French with a pretest, immediate posttest, and delayed posttest design.The analysis of eight proficiency measures showed that the group that received prompts distinguished itself by being the only group to significantly exceed the comparison group on all eight measures.The recast group significantly outperformed the comparison group on five of the eight measures.
At the same time, Sheen (2004) examined similarities and differences in teachers' corrective feedback and learners' uptake across instructional settings.Four communicative classroom settings -French Immersion, Canada ESL, New Zealand ESL and Korean EFL -were compared.The results indicate that recast was the most frequent feedback type in all contexts but was much more frequent in the Korean EFL and New Zealand ESL classrooms.The rates for both uptake and repair for recasts were also greater in these two settings.
Then, further studies were conducted to investigate the effects of CF (particularly recast, prompt and explicit correction) on learners' language development with similar positive results (e. g.Sheen, 2010;Lyster & Saito, 2010a;Sato & Lyster, 2012).
In China, Yang (2009) reported an observational study on CF using Lyster and Mori's (2006) feedback coding scheme to document the frequency of teachers' feedback moves and learner uptake rate following feedback.The findings show that the distribution of prompts and recasts was in reverse order compared to other communicative second language contexts.In other words, explicit correction and prompts had a higher uptake rate than recasts, suggesting that form-oriented Chinese foreign language context indeed mediates the pattern of feedback and uptake.Yang and Lyster (2010) carried out another quasi-experimental study comparing the effects of recast and prompt on 72 Chinese learners' use of regular and irregular English past tense.Their comparisons of group means across testing sessions using a repeated-measures ANOVA revealed significant gains by the prompt group on all eight measures, the recast group on four, and the control group on three.
It can be concluded from these studies that recasts are by far the most frequent type of feedback than prompts and explicit corrections in a range of classroom settings, though prompts have been proved to be most effective (Lyster & Mori, 2006;Lyster & Saito, 2010b;Lyster, Saito & Sato, 2013).
In another respect, Kartchiva and Ammar (2014) found that the effects of CF were associated with target types.For example, in their experiment they found that learners were quicker on the uptake of feedback on past tense errors than on problems with questions in the past.
However, since few studies, as mentioned above, have focused on investigating teacher and student perceptions of corrective feedback on pronunciation for students' presentations, this study is an attempt to do so.
As for the opinion of interrupting students' flow of speech, though Terrell (1977) and Krashen (1985) point out that correcting mistakes during the students' speech could make them nervous and that language is mainly responsible for transferring information, and being interrupted is not good for keeping the theme integral.Whereas Swain and Lapkin (1995) think that correcting students' errors could help them not only to avoid errors but also to prevent fossilization of them in their second language acquisition process.Doughty and Long (2003) also believe that feedback which is collected during communication helps the learner to better acquire the second language.
So, in this research, attempts were also made to find out what teachers and students think about the proper time for corrective feedback on pronunciation for students' presentations in advanced English classes.

The Current Study
The following was the research question that guided the study.
What are the similarities and differences between the perceptions of teachers and of students about corrective feedback on students' pronunciation for their presentations?

Participants
The participants of the study were 73 students (5 males and 68 females) of the School of Foreign Languages, Beijing Forestry University, among whom 40 were junior English-major students and 33 senior English-major students, aged from 20 to 24.They have all finished basic courses of English language skills, such as listening, speaking, reading and writing, especially English pronunciation.At the time of the study, they were taking more specialized courses, such as Introduction to Linguistics, English and American Literature, Culture and Society of English Speaking Countries, Interpretation and Translation, Academic Thesis Writing and so on.
Meanwhile, another group of participants were 25 teachers who have been teaching both basic and specialized courses in the Department of English or just specialized courses offered to college students in the Department of College English.In such classes, one of the most important modules is students' presentation, in which they are required to make PPT according to the requirements of the teacher (Note 3), present it in front of the class in English for about five to ten minutes.While the student is making the presentation, the teacher will sit in class, listening carefully.He/she may or may not make comments or offer the student corrective feedback during or after the presentation.

Instruments
In order to produce a well-designed questionnaire, first a group interview with several students was conducted to have brainstorming with regard to CF on pronunciation.The interviewees were asked to recall when and where they received corrective feedback, in which way and who offered it.They contributed much useful information which was noted down and categorized by the authors.
Based on the interviewees' ideas together with the data that the researchers had already collected, the first draft of the questionnaire was produced.Then, it was revised several times before the final draft was used for the study, which consists of 29 questions with 28 close questions using a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree) and one open-ended question.
The 28 items were divided into five categories which were sometimes put at random in the questionnaire to avoid order effect.These five categories include: (1) general perceptions of CF; (2) when to provide CF; (3) the expected effects of CF; (4) how to provide CF; and (5) preference for the type(s) of CF.Also, the questionnaire was designed in two versions, one for the students and the other for the teachers.But the content (the 29 question items) and the format were exactly the same for the purpose of comparison.

Procedures
It took almost one week for the authors to distribute the questionnaires to the students, and more than two weeks to retrieve the teachers' questionnaires.
The questionnaires were first distributed to the students when they were in their dormitories in the evening.Before they answered the questionnaire, the researchers offered some training to the respondents, explaining the purpose, clarifying some abstract terms (such as recast, prompt, explicit correction, word stress, sentence stress, and language transfer, etc.) and how they should write their answers.After they finished, all their questionnaires were collected by one of the authors.At the same time, the questionnaire was sent to male students through e-mail, asking them to complete the questionnaire at their earliest convenience.Fortunately, all the respondents were cooperative enough for this survey.Finally, 5 male students' and 68 female students' valid questionnaires were retrieved.
The teachers' questionnaires were delivered mainly through two channels, one is by delivering in person and the other is through e-mail.During the week, some teachers who were giving classes in Teaching Building No. 1 were visited and asked to finish the questionnaire during the 20-minute break, others preferred to hand it in at their convenience.Considering different working hours of the teachers and the difficulty in making an appointment with them, the questionnaire was sent to them through e-mail.Almost two weeks later, all 25 valid questionnaires were sent back to the researchers by the teachers.

Data Analysis
The students' (N=73) and the teachers' (N=25) weighted means for their perceptions of corrective feedback on pronunciation for students' presentations in advanced English classes were calculated based on Griffiths (2007: 95).All the 28 closed questions were arranged to 5 choices, for example, from "totally disagree" to "totally agree" (A=totally disagree, B=disagree, C=no opinion, D=agree and E=totally agree).Meanwhile, in order to calculate the weighted mean for each of the questions for the teachers and the students, numerical value was given to each choice (A=1 point, B=2 points, C=3 points, D=4 points and E=5 points).
After the authors collected the questionnaires, the data from both versions were inputted into a database (SPSS 20.0) for analysis.First, Cronbach's alpha co-efficient of the reliability for the students' version was calculated.Then, the average reported frequency of CF on pronunciation for presentations across all students was calculated for each CF item and overall, and the number of CF items considered important at a high rate of frequency (defined as overall average level of frequencies across all items and participants= 3.37 or above) was worked out.To probe into the difference of perceptions by higher and lower level students, the student sample was divided into two groups: the junior student group (N=40) and the senior student group (N=33).The average reported frequency of CF on pronunciation for presentations for each CF item and overall, and the number of CF items considered important at a high rate of frequency was also calculated for both the junior and the senior student groups.
The data for the teachers' CF questionnaire (N=25) were also analyzed for reliability.Similar to the students' data, the average reported frequency of CF for each item across all teachers and overall, and the number of CF items regarded by the teachers as particularly important (using the same average = 3.37 or above as threshold to keep consistent with the students' data) was calculated.All these results were then compared with those from the students' data.

Results and Discussion
The Cronbach's alpha co-efficient for the reliability of the students' version of the CF questionnaire was 0.80, and for the teachers' version it was 0.77.The results of the reliability analyses are above the standard reliability threshold of 0.70, indicating that the questionnaire is basically reliable.Although the reliability of the two versions of the questionnaire was not very high, many aspects of the teachers' and the students' perceptions in question could still be reflected from it.

General Perceptions of CF
The results from both the students' version and the teachers' version of the questionnaire are shown in Table 1.
From the table, it can be seen that the average frequency of CF on pronunciation for presentations reported by all students (N=73) over all items was 3.38, with a higher average frequency of CF on pronunciation for presentations reported by junior students (3.42) than that reported by senior students (3.34).Seventeen items happened to be chosen by junior students, senior students and all students, respectively, at a high rate of frequency though these items chosen are not completely the same (see Items 9 and 13).The average frequency of CF on pronunciation for presentation reported by teachers (N=25) over all items was 3.35, with 18 out of 28 CF items showing stronger preference than other items.So, it is apparent that there is great similarity in teachers' and students' average frequencies of CF on pronunciation for presentation.Among the 18 CF items as indicated above which teachers strongly favoured, 14 (78%) were agreed by all students, meaning that there was a general consensus of opinions between teachers and students with regard to CF on pronunciation.This is a promising result because teachers and students' similar perspectives can facilitate CF practices, which implies potentially positive outcomes in terms of students' pronunciation improvement.
From the table, it can be seen that the four items at the beginning of the questionnaire (Items 1, 2, 3, 4) and two other items (Items 7, 16) were designed to elicit, directly and indirectly, the respondents' general perceptions of CF on pronunciation except Item 1, which was used as a warm-up question.For Items 1 and 3, teachers and all students (including junior and senior students) agreed that spoken English is important and presentation is a good time to provide CF for students.The disapproval of Items 2, 4 and 7 by both teachers and students denied some stereotypes that junior and senior students have few pronunciation problems, inferring that they still need the attention and help of teachers with their pronunciation.Surprisingly, Item 16 (listening to presentations carefully) was not regarded as highly important by teachers, although it received a high rate of frequency by junior students, senior students and all students.Perhaps teachers thought that the main purpose of asking students to do presentations was to check how well each of them learned about the course rather than the pronunciation of the students (Note 4).

When to Provide CF
Items 5 and 6 require teachers and students to describe facts as to when teachers usually provide CF.Both teachers and students reported that teachers provided CF as soon as students finished their presentation (Item 5) though the amount of feedback they gave was not clear.Interestingly, when asked whether teachers offered CF after class, teachers' response was more positive than students' (Item 6).It might be possible that in the minds of teachers it would be helpful to give students' CF after class, which was consistent with Item 8 though in fact they did not find time to do so.Comparing Item 8 with Item 9, it would be more interesting to find that while teachers were more concerned about hurting students' self-respect, students (particularly junior students) cared more about time though they also agreed that offering CF after class protects students' self-respect (Item 11).To obtain further information, Items 23, 24, 25 intended to find out what time the respondents thought was proper for teachers to provide CF.They highly agreed that CF should be provided after presentation or class instead of interrupting the student while he/she is doing the presentation.

The Expected Effects of CF
In respect of the effects of CF, both teachers and students turned to be very objective.While they (teachers and all students, including juniors and seniors) admitted the effectiveness of CF (Item 17) despite students' nervousness (Item 19), they all agreed that CF could not solve all the problems.For example, teachers and all students (including juniors and seniors) thought that CF could be of great help to their word stress (Item 12), but limited help to sentence stress (Note 5) (Item 13, as was viewed only by seniors).Perhaps in their teaching, teachers found word stress problems more easily than sentence stress problems.Perhaps they did give CF to both word stress and sentence stress problems ( X =3.36) even though they might not have deliberately classified them.As for students, especially junior students, they might not be that clear of the two concepts even though they had taken English Pronunciation class not long ago relative to senior students ( X =3.41).Or because of that, they might not have realized that they were given CF in the two aspects by teachers.On the whole, both teachers and students thought that though CF is helpful in improving students' pronunciation, it does not mean that they will not make the same errors again after CF (Item 18).With regard to the comparison, implicit though, of the effectiveness of CF in and out of class (Item 10), all respondents reported that in-class CF helps memorize correct pronunciation since each student has his/her own errors corrected and they benefit from the CF given to other students as well.

How to Provide CF
Before coming to the specific point, a question of why students commit pronunciation errors (Item 20) was asked to see whether they had any knowledge of the causes of errors based on which they were supposed to give more practical solutions to their problems.Obviously, teachers were knowledgeable enough about language transfer, so they took it for granted that CF should be given to negative transfer problems (Item 21) though the function of presentation might not be considered that great in this regard (Item 22).Whereas students' average frequencies for the three items were fairly low.Two reasons might account for this.One might be that they did not fully understand the concepts even though the researchers explained these to them before they filled in the questionnaire.The other might be that they understood the concepts, but they did not think there was any negative transfer from Chinese since it is a tone language while English is a highly stressed (stress-timed) language.The most interesting is that in Item 14 all teachers and students agreed that CF should be offered to repeated errors, while in Item 15 only students expected that all mispronounced words would receive equal treatment.In other words, teachers might think it was not necessary, nor could they find time to do so.

Preference for the Type(s) of CF
Since a complete comparison between recast and prompt and explicit correction would require six question items which might cause ambiguity, three (Items 26,27,28) were designed instead.At the first sight, it seems that all respondents gave almost equal weight to the three types of CF.Making a closer scrutiny, one would find some differences.Teachers preferred prompt ( X =3.76) to recast ( X =3.52) possibly because they believed that it enables learners to do self-repair, which can facilitate internalization of their knowledge.As for their preference for explicit correction ( X =3.60), the reason might still be the same: students should not just know what, but why.Junior students turned to like recast ( X =3.85) more than prompt ( X =3.58) and explicit correction ( X =3.58), while senior students had similar preference for recast ( X =3.79), but they like explicit correction ( X =3.67) more than prompt ( X =3.61), which held true with all students ( recast X =3.82, licit X exp =3.62 and ( prompt X =3.59).Or to put it another way, recast was students' first choice, followed by explicit correction and then prompt.The reason might be that prompt is too challenging for the students who have been used to teacher-fronted or Chinese way of teaching, waiting for the teacher to give correct answers.
The last item, Question 29 , is an open-ended question, which requires respondents to give further advice about CF on pronunciation for students' presentations.The most useful suggestion from teachers is that students should pay close attention to CF related to both segmental and suprasegmental aspects of pronunciation.While students suggested: (1) teachers summarize students' common pronunciation problems before the end of each class, which would take only 3-5 minutes but would be much help; (2) teachers give different treatments to different students.For example, sensitive students should not be publically corrected, while outgoing students can be given CF even during the presentation, if it is necessary.

Conclusion
The findings of this research have provided us with a good picture of teachers' and students' perceptions about corrective feedback on students' pronunciation for their presentations.In terms of similarities of their perceptions, results show that despite some stereotypes that pronunciation is not the focus of teaching for junior and senior students, there is a general consensus among teachers, junior and senior students (78%) that CF is not only important but necessary since the students still have pronunciation problems which need teachers' help.Their presentation is, therefore, considered to be a good opportunity for students to get that help.
It is also encouraging to find that both teachers and students agreed that it is inappropriate to interrupt students when they are making a presentation.The best time to provide CF is after the presentation or class.
As for the effectiveness of CF, though teachers and students were aware of the limited functions of CF, they still believed that they would benefit from it.For example, providing CF in class, or after students' presentations in particular, was considered more effective.
With regard to differences of teachers' and students' perceptions, one of the most important discovery is that students' eagerness to learn exceeds their concern about their self-respect.This is a good news for teachers who are too obsessed about it when they are trying to help their students.
Also important is that teachers thought CF should be offered to repeated errors, while students would like to receive as much CF as possible (repeated errors as well as all mispronounced words).
Still another major discovery is the respondents' preference for different types of CF.Teachers and students held similar opinions about explicit correction, but they differed greatly in the other two types of CF: teachers favoured prompt which they might consider more effective since it "forces" students to make use of their knowledge they have previously acquired to make self-repair, creating a condition conducive to learning; students, on the other hand, preferred recast probably because prompt is more demanding than the other two types of CF, especially when they are under stress.
In spite of some mismatches between teachers' and students' perceptions with regard to CF on pronunciation, the pedagogical implications of this research are obvious.Firstly, since the findings of the survey show that both teachers and students agreed that CF was not only important but necessary, ESL or EFL departments should attach more importance to junior and senior students' pronunciation even if they no longer have pronunciation classes and find effective measures to tackle their pronunciation problems.Secondly, the survey demonstrates that this cohort of students have a strong desire to improve their pronunciation before they graduate, therefore, it is suggested that teachers take advantage of their motivation for learning and provide timely and necessary CF for them.Furthermore, for errors that students think difficult, recast and explicit correction might be used, while for errors that teachers think students have the ability to make self-repair, prompt might be more appropriate.In summary, through a joint effort between teachers and students, junior and senior English majors' pronunciation will surely bear fresh fruit.Zhao, Y. (1997).The effects of listeners' control of speech rate on second language comprehension.Applied Linguistics, 18, 49-68.https://doi.org/10.1093/applin/18.1.49

Notes
Note 1. Immersion programme is a form of bilingual education in which children who speak only one language enter a school where a second language is the medium of instruction for all pupils.For example, there are schools in Canada for English-speaking children, where French is the language for instruction (Richards et al., 2002, p. 217).
Note 2. Lyster and Ranta (1997) define the four interactional moves in this way: Clarification request: the teacher indicates to the student, by using phrase such as "Pardon me" and "I don't understand," that the message has not been understood or that the utterance is ill-formed in some way, and a repetition or a reformulation is required.Repetition: the teacher repeats the student's erroneous utterance, adjusting the intonation to highlight the error.Metalinguistic clues: the teacher provides comments, information, or questions related to the well-formedness of the student's utterance, without explicitly providing the correct form (e.g., "Do we say 'goed' in English?").Elicitation: the teacher directly elicits correct forms from students by asking questions such as "How do we say that in French?"; or by pausing to allow students complete the teacher's utterance; or by asking students to reformulate their utterance (e.g., "Try again").(p.243).
Note 3. The teacher of a courses, such as Introduction to Linguistics, English and American Literature, Culture and Society of English Speaking Countries, Interpretation and Translation, Academic Thesis Writing and so on.
Note 4. Technically, this item was not well designed in terms of the validity of the questionnaire, which is one of the limitations of the research.
Note 5.Here only two (suprasegmental) aspects of pronunciation were set as question items since it is not possible, neither is necessary to cover all of it.

Table 1 .
Average levels of importance of CF on pronunciation for presentation considered by junior students, senior students, all students and teachers