Corrective Feedback in Oral EFL Learning Environments : A Study on Swedish Teachers’ Awareness of Corrective Feedback Strategies and Effects

(1)

1

Examensarbete

Grundnivå

Corrective Feedback in Oral EFL Learning

Environments

A Study on Swedish Teachers’ Awareness of Corrective

Feedback Strategies and Effects

Författare: Josefine Westerberg

Handledare: Iris Ridder Examinator: Jörgen Dimenäs

Ämne/huvudområde: Pedagogiskt arbete Kurskod: GPG22K

Poäng: 15 hp

Examinationsdatum:

Vid Högskolan Dalarna finns möjlighet att publicera examensarbetet i fulltext i DiVA. Publiceringen sker open access, vilket innebär att arbetet blir fritt tillgängligt att läsa och ladda ned på nätet. Därmed ökar spridningen och synligheten av examensarbetet. Open access är på väg att bli norm för att sprida vetenskaplig information på nätet. Högskolan Dalarna rekommenderar såväl forskare som studenter att publicera sina arbeten open access.

Jag/vi medger publicering i fulltext (fritt tillgänglig på nätet, open access):

Ja ☒ Nej ☐

(2)

2

Abstract:

The purpose of this study was to investigate four Swedish upper secondary teachers' reasoning about their own corrective feedback strategies in oral EFL learning environments, and how their reasoning correlated to previous research in the subject. The research was carried out by using the qualitative method of interviewing a selection of teachers. Four teachers of English as a foreign language from one Swedish upper secondary school participated in semi-structured pair-interviews.

The study found that the participating teachers’ perceptions and use of corrective feedback corresponds to a large extent with the findings of previous research, saying that teachers in various contexts prefer more implicit means of corrective feedback, but that they would adapt their strategies depending on the learner, error type, and the focus of the lesson. The result also suggested that although the teachers reportedly make conscious choices when selecting a corrective feedback strategy, they sometimes have to choose between using a strategy that promotes learning or using a strategy that will cause the least amount of inconvenience for the learner.

The conclusion was made that the teachers of the study are aware of their own corrective feedback strategies to some extent, which also corresponds with the results of previous research, but that their awareness is not always sufficient, and their strategies might not be as effective as the teachers think.

Keywords: corrective feedback, EFL, foreign language acquisition, English,

(3)

3

Abstract

1. Introduction

In the process of second language acquisition there are various factors that will determine the success and proficiency level of the learner. Things such as motivation, background, disabilities and other circumstances can make a substantial difference to the outcome. In a nation such as Sweden, where the English language is present everywhere and functions as a second language to some extent, an important target of the educational system is to enable learners to develop their language proficiency in order for them to be a part of the language community. Skolverket and the curriculum of the English subject describes the aim of the English courses

Teaching of English should aim at helping students to develop knowledge oflanguage and the surrounding world so that they have the ability, desire andconfidence to use English in different situations and for different purposes. Students should be given the opportunity, through the use of language in functional andmeaningful contexts, to develop all-round communicative skills (Skolverket, 2011).

Teachers do of course play an important role in exposing learners to knowledge, and they have access to strategies and tools to apply when circumstances call for it. One of the methods used by teachers to increase learners' proficiency in the target language is to provide feedback. Feedback can be given written or orally; immediate or delayed; and within these different categories there are several subcategories. This study will devote attention to feedback of a corrective nature, provided immediately and orally when the non-target like speech occurs by the learners of English as a foreign language (EFL), which from here on will be referred to by the more general term Corrective Feedback (CF for short).

According to Ellis, Loewen & Erlam (2006)

Corrective feedback takes the form of responses to learner utterances that contain an error. The responses can consist of (a) an indication that an error has been committed, (b) provision of the correct target language form, or (c) metalinguistic information about the nature of the error, or any combination of these (p. 340)

Any learner non-target like production can be considered an error, concerning pronunciation, lexical choices, grammatical features, etc. and can be responded to by the teacher through CF.

The advantages and disadvantages of different kinds of CF have been the subject of many studies, and many researchers disagree, not only on what method should be used in what context, but also on whether or not CF should be used at all. Section 2 will summarize some findings from a selection of previous studies and, based on that, attempt to reach a conclusion in the matter.

Sheen (2004) suggested that “investigating teachers' perceptions and beliefs underlying CF moves is [...] needed” (p. 295) which became the main inspiration for conducting this study, but my own experiences as an EFL teacher and the difficulties concerning CF that I have encountered has also inspired me to research the subject further.

(5)

5

1.1 Aim of the study

The purpose of this study is to investigate four Swedish upper secondary teachers' reasoning about their own corrective feedback strategies in oral EFL learning environments, and how their reasoning correlates to previous research in the subject. This will be done by attempting to answer the following research questions

• How do teachers describe CF and its effects, and how do teachers practice CF in the EFL learning environment?

• How does previous research about CF correlate to how Swedish upper secondary EFL teachers use CF?

1.2 Theoretical perspective

The current study is built upon the ideas of Ellis, featured in his article Corrective Feedback and Teacher Development (2009), which will be described in more detail in the next paragraph, but also upon sociocultural theory (Vygotsky, 1978). The sociocultural theory explains that learning is a social process, i.e. humans learn through interaction with others. One part of the theory is called the zone of proximal development (ZPD), which states that a person can learn on their own up to a certain point of which they are cognitively prepared, but with help of social interaction they can develop further. A teacher, or a peer with more experience or knowledge in the subject, can provide “scaffolding”, a support structure, to enable the learner to acquire more complex skills. In other words: the ZPD is the distance between the cognitive abilities that a learner as already acquired from going through certain developmental processes, the actual developmental level, and what a learner can acquire through interaction with others (Ibid p. 85-86). Applying sociocultural theory when studying corrective feedback implies that learners should be able to acquire more advanced language skills through support (scaffolding) from their teacher. Within the subject of corrective feedback such scaffolding could consist of linguistic clues, target language input, etc.

Ellis (2009) presents some discrepancies concerning corrective feedback and says that researchers “have frequently disagreed about whether to correct errors, what errors to correct, how to correct them, and when to correct them” (p. 3). Ellis also claims that teacher educators are usually not very keen on suggesting what CF strategies teacher students should use, partly because they simply do not have enough knowledge about the effectiveness of different methods, but also because they know that the outcome of CF depends on many factors (2009, p. 10). Because of the difficulties involved in the CF decision making process Ellis emphasises the importance of reflection, evaluation, and development: “reflecting on CF serves as a basis both for evaluating and perhaps changing existing CF practices and, more broadly, for developing teachers' understanding of teaching and themselves” (2009, p. 15).

Ellis (2009) set out to provide some guidelines for teachers to consider when attempting to use CF (p. 14), but even though Ellis himself makes these suggestions, he still maintains that teachers “should not accept pedagogic proposals without submitting them to their own empirical enquiry” (p. 15). Firstly, Ellis argues that teachers should inform their student of the advantages of CF and make sure students and teacher agree on its purpose. The CF must also be easy to recognize, i.e. if learners do not understand that they are being corrected they are less likely to repair their output. Teachers must give the learner a chance to repair their output (oral and written) but should not force them to do so. The learner must have time to think

(6)

6

following CF, so it is important that the teacher does not interrupt or move on with the conversation. Secondly, teachers should not hesitate to give CF, in both accuracy and fluency focused tasks, but the CF strategies should be adapted to the situation. Therefore, teachers must have knowledge about different methods of correction, and be able to start out with as implicit feedback as possible, and if necessary, provide more explicit feedback and be prepared to correct the same error many times (Ellis, 2009, p. 14).

Individual differences must also be taken into consideration, according to Ellis (2009, p. 14), as different learners require different kinds of feedback. Some may not like to be corrected explicitly in front of the whole class, while others may not recognize implicit feedback. CF may cause anxiety in some learners and teachers must watch them carefully and adapt CF strategies to reduce the level of anxiety.

This is the perspective, developed by Ellis (2009), that the research questions and the interview guide of this study was inspired by, and the process of analysing the data will be guided by and compared to this theoretical perspective.

1.3 Definitions

Before any previous research can be presented some key concepts must be defined, as they contain information which is important to the context. Uptake refers to learner's reactions following feedback which indicates that the feedback has been noticed (Ellis et al. 2006, p. 366). The learners can repeat what the teacher has said (the correct form provided to them), or respond by saying “Yes”, for example, to let the teacher know that they have noticed the feedback.

Various corrective feedback strategies are mentioned throughout the text, and the different types are explained next. Strategies 1-6 are defined by Ellis (2009, p. 9), and strategy number 7 “Metalinguistic Information” is defined by Lyster and Ranta (1997, p. 47).

Recast is a strategy used by the teacher (T) to change the incorrect content to the correct form by paraphrasing, and still maintain the meaning of the learner's (L) utterance

Example 1. L: I go there two times.

T: You've been there twice? Did you like it there? The corrector can choose to repeat the learner's utterance and emphasise the incorrect form, which is known as repetition

Example 2. L: I go there two times. T: You GO there?

L: I have been there to times.

By making a clarification request the corrector shows that the learner's utterance was not understood

Example 3. L: What you go on weekend? T: What?

An explicit correction is a clear indication of an error including the correct form being provided to the learner

Example 4. L: I go there two times.

T: Not go, have been. We say, “I have been there.” An elicitation is a less clear indication of an error. Here the corrector repeats only the correct part of the learner's utterance and uses rising intonation before stopping to let the learner finish the utterance

Example 5. L: Yesterday I go there. T: Yesterday I..?

(7)

7

Sometimes the corrector does not have to say anything at all, but uses some form of paralinguistic signal, such as a gesture or facial expression, to indicate an error

Example 6. L: Yesterday I go there.

T: (waves hands towards their back to indicate past tense)

When metalinguistic information is given the corrector provides a clue about the appropriate language structure but do not include the correct form

Example 7. L: Yesterday I go there.

T: Yesterday? Then you need past tense, right?

2. Literature review

As early as the 60's Corder (1967) recognized that some teachers did not benefit from the research aimed at error finding and explaining their origin, which was a common target of research at that time. The teacher he referred to was “more concerned with how [italics in original] to deal with these areas of difficulty than with the simple identification of them” (p. 162). More recent research has focused on this exact issue of how to handle non-native output.

Further, Corder emphasized the importance of distinguishing between errors and mistakes, the latter referring to slips of the tongue or pen due to stress, concentration difficulties, and other performance influencing factors. Mistakes will not influence the learning process, whereas errors are likely to do so as they are systematic and reflect faulty structures that have already been acquired (1967, p. 167). Corder was however aware of the problem facing teachers and researchers when they are to decide what is a temporary mistake and what is a reoccurring error (ibid), which Ellis (2009) agrees with and argues that the difference between errors and mistakes is not always obvious, and it will only become somewhat noticeable by the teacher after a longer time teaching a group of learners. Even if distinguishing between errors and mistakes is possible, the problem remains when it comes to the correction itself: teachers will have to decide on the spot when an oral error/mistake occurs and may not have time to analyse the situation, which is why sometimes mistakes are treated as errors (p. 6).

Lightbown and Spada (2006) gave similar advice on some of the issues as Ellis, though on some points the opinions differ. According to them, learners' attention needs to be drawn towards the erroneous structures by providing explicit feedback (p. 190), while Ellis says that choice of feedback strategy should depend on the type of error and situation (2009, p. 14). Both do however agree that the timing of CF must be taken into consideration when choosing a strategy to apply, as it may potentially affect learners emotionally, causing anxiety and embarrassment if the learner is sensitive about immediate correction in front of others (Lightbown & Spada 2006, p. 191, Ellis 2009, p. 14).

Harmer (2007), like Ellis (2009), suggests that teachers should talk to their students about CF and explain the purpose of it, how it will be implemented in the classroom, and why. Students should also have an opportunity to share their feelings towards CF (Harmer 2007, p. 144). He also agrees with the idea that feedback should start out implicitly by hinting that something is wrong or asking students to repeat their utterance, but if the students cannot correct themselves or do not even notice the CF, the strategies need to become more explicitly corrective, e.g. metalinguistic information (p. 144). But according to Harmer (2007) the procedure should be different depending on the focus of the work, i.e. accuracy/fluency work. The target

(8)

8

for accuracy work should be to identify mistakes, make the learner aware of it and eventually get them to correct themselves, while the purpose of fluency work is more focused on content and meaning and feedback should therefore target content-based problems. During fluency work the feedback can be “gentle” (p. 145-146), using strategies that simply provide the correct form, without insisting on learner repair at that moment. After the task has ended more feedback can be given based on what the teacher has observed throughout the lesson. This feedback can be given to the whole class at once, having the students correct the errors that are brought up, or given individually to each student (p. 146-147).

Ellis' guidelines will now be examined further by reviewing a few other studies that have investigated some of the above-mentioned concepts in more detail. Among them is another study by Ellis along with Loewen and Erlam (2006) in which they found that metalinguistic feedback can be more beneficial than recasts because metalinguistic information is more salient, i.e. learners are more likely to notice the feedback (p. 363). However, recasts did also lead to learner repair, but the example given from such feedback episodes shows that the learner after recasts would only repeat the teacher's correct form, while metalinguistic information forced the learner to think about language structure and search their own memory for the correct form (Ellis et al. 2006, p. 363). Their study also managed to show that the benefits of metalinguistic feedback would appear to a larger extent in the delayed post-tests, which indicates that some long-term acquisition had taken place. It is suggested that the results could have been different if the learners had not already had some previous knowledge about the structures that they were tested for, and that it is not possible to tell, based on their study, whether feedback will affect the acquisition of new structures (Ellis et al. 2006, p. 364).

Lyster and Saito (2010) compared the effects of recasts, explicit correction and prompts, the latter being a collection of different CF strategies, which according to them include elicitation, metalinguistic information, clarification requests, and repetition. They explain that “although prompts include a range of CF types, they have one important feature in common that differentiates them from reformulation moves: they withhold correct forms and instead provide clues to prompt students to retrieve these correct forms from their existing knowledge.” (p. 268). Prompts were shown to be more effective because the groups that were exposed to that kind of feedback made more progress than those who received recasts. Younger learners are described to be more sensitive to CF, and appropriate CF given to younger learners is suggested to have a positive effect on their language development. Appropriate feedback according to Lyster and Saito (2010) seems to be a mixture of recasts and metalinguistic information (p. 294-295).

Panova and Lyster (2002) looked at what feedback strategies were used and the effectiveness of them, among learners in adult education. They found seven different types of CF used by the teacher: recasts, translation, clarification requests, metalinguistic feedback, elicitation, explicit correction, and repetition (p. 582). Recasts together with translation were used in 77% of the cases, clarification requests in 11%, metalinguistic feedback in 5%, elicitation in 4%, explicit correction in 2%, and repetition in 1% of the cases (Panova & Lyster 2002, p. 586). Perhaps the most significant finding in this study was the low rate of learner repair: only 8% of all the errors were repaired after teacher feedback (ibid.), indicating that the most commonly used strategies may not be the most effective. CF strategies that resulted in most learner repair were repetition and elicitation, followed by metalinguistic feedback and clarification requests (p. 587). Teachers, as the numbers show,

(9)

9

preferred to use recasts rather than strategies that prompt learners to repair (p. 588), even though the repair rate for recasts was low, which could be explained by assuming that teachers were not aware of the effects of different CF strategies. Another explanation is provided by Panova and Lyster (2002), who instead of saying that teachers lack knowledge, claim that teachers make choices very consciously as the low level of proficiency among the learners prevent teachers from using other strategies, i.e. if learners have “limited linguistic resources” (p. 588) they might not be able to understand metalinguistic information, for example. Instead, recasts are seen as opportunities for teachers to provide target language input and improve communication, which has value in itself, even if it is not always noticed as CF by learners (p. 591).

Another study focused on adult learners was conducted by Sheen (2004) in Korean EFL classrooms, where the learners had mixed levels of English proficiency but their overall level of education (other than English) was high. The results of previous studies were compared to see if CF use across different instructional settings would vary (p. 271). The Korean students were mostly exposed to CF in the form of recast – 83% of all the CF episodes – and the total amount of recasts recorded in all the studies combined wereover 50% (p. 283). The effects of recasts did vary across the different settings: the repair rate was around 69% for the Korean learners, which was compared to the use with Canadian learners where the repair rate was 34% (p. 285).

An explanation of the preference for recasts in the Korean setting is said to be the focus on meaning in fluency activities, and the teacher's attempts to maintain the flow of communication (Sheen 2004, p. 291). In other words, recasts are used because the teacher thinks that they interfere the least. Because the use of recasts was relatively high in all settings (although highest in Korea), but the repair rate following recasts varied a lot between the settings, the researcher gives some suggestions to why recasts were so successful in Korean EFL compared to the Canadian ESL (English as a Second Language). Learners' previous education as well as age can influence their ability to respond to recasts, Sheen suggests, and because of the high levels of previous education in Korea compared to the low levels in Canada and the age difference (Korean learners were older than the Canadian) the Korean learners were much more receptive to recasts (p. 291). Some examples of Korean recasts show that the teacher chose to focus on one form at a time by reformulating only the problematic structure/word, and by adding stress, rising intonation, or emphasis the recasts became more salient. The teacher also gave learners theopportunity to notice and repair before moving on with the conversation (p. 292). Additionally, class size is mentioned as a factor for recast salience, i.e. it may be easier to notice recasts as CF in smaller groups (p. 293).

Some say that CF in fluency-focused activities should be different from that in accuracy-focused activities (see for example Harmer 2007, p. 145-147), so Milla and Pilar García Mayo (2014) set out to examine if teachers would use different CF strategies depending on the focus of the class, and also to compare the effects of different CF strategies and learners' reactions to CF (p. 5). The observations took place in an EFL classroom and a CLIL (Content and Language Integrated Learning) classroom and showed that the EFL teacher used more explicit feedback than the CLIL teacher, who preferred more implicit methods (p. 9). It is assumed that the reason behind the preferences of CF methods is connected to the different focus of the two settings. Not only did the EFL teacher provide more explicit feedback than the CLIL teacher, but a lot of feedback was given in many ways (Milla & Pilar

(10)

10

García Mayo, 2014, p. 10), as seen in one occasion where the teacher responds to an error by using seven different types of CF and two of them more than once. The EFL teacher is said to use all of these strategies to focus learners' attention towards the error and make sure they notice the correct form, while the CLIL teacher is careful not to interrupt the learner too much by reformulating the incorrectly pronounced word (Milla & Pilar García Mayo, 2014, p. 10).

When looking at learner uptake following CF Milla and Pilar García Mayo (2014) found that the overall uptake was much greater in EFL than in CLIL: 82% compared to 52% (p. 11). The CF strategies led to different amounts of learner uptake, both in EFL and in CLIL. Only recasts and elicitation led to uptake in CLIL, but in EFL all CF types (recasts, clarification requests, repetition, elicitation, metalinguistic clues) except explicit correction led to uptake (p. 12). Recasts were used in both settings but only led to about 40% uptake. The low amount of uptake could be explained, according to Milla and Pilar García Mayo (2014), by salience, i.e. recasts are not always noticed as CF (p. 12). Another reason for the low uptake rate is suggested to be caused by the teachers themselves: the teachers did not always give learners an opportunity to reflect upon the issue, because the conversation moved on shortly after CF (p. 13). It is suggested that when focus was on meaning the learner did not pay any attention towards the recasts and moved on with the conversation, without showing any sign of having noticed the recast, directly following the feedback (p. 13).

In the CLIL class elicitations were effective, although rarely used, as seen in one example where the learner immediately changes the utterance from the L1 to the TL following the teacher's request (Milla & Pilar García Mayo, 2014, p. 14). The teacher of the EFL class, on the other hand, had to use many other strategies following elicitations for the learner to understand what was wrong, and how to get it right (p. 14).

2.1 Conclusion

The general advice given to teachers and teacher students by Ellis (2009), Lightbown and Spada (2006), and Harmer (2007) is that corrective feedback in some form should be included in educational contexts. Many studies have investigated the effects of CF, like the studies reviewed above, which clearly shows that CF is used by teachers in many countries around the world to some extent, whether they are aware of researchers’ recommendations or not.

The opinions of the four authors also correspond with the findings of the other studies when it comes to context and variation. Ellis (2009) and Harmer (2007) mention the importance of adapting CF strategies to the context, i.e. individual learner, situation, and focus of the lesson, while Lightbown and Spada (2006) only mentionexplicit feedback, which includes many types of strategies, but they do not give any examples of what types they recommend. The other studies show that teachers tend to vary their CF, and sometimes they use multiple strategies to make sure the feedback is noticed. The EFL teacher in Milla and Pilar García Mayo's study (2014) used explicit feedback first and added other strategies when uptake did not occur, which is the opposite of what Ellis suggested, i.e. that implicit feedback should be tried first, and explicit feedback added if needed.

When summarizing the main findings of the reviewed studies it is clear that recast was the most frequently occurring type of CF, but that in most cases explicit feedback, such as elicitation, metalinguistic information, and clarification requests, were more effective. Recasts were effective for older learners with a high level of

(11)

11

previous education, and/or when teachers gave enough opportunities for uptake and repair. Younger learners were found to benefit from both explicit and implicit feedback. The teachers participating in these studies tended to change their CF strategies depending on the context and were more likely to use explicit feedback during accuracy focused activities, and implicit feedback during fluency focused activities.

3. Method

This study investigated four Swedish upper secondary teachers' reasoning about their own corrective feedback strategies in oral EFL learning environments, and how their reasoning correlates to previous research in the subject. Teachers were asked to take part in a recorded interview together with one other participant. A table with the seven different CF strategies was given to the participants and all CF strategies were explained and defined before the interview began. Teachers were also informed that the questions concerned oral CF only, i.e. feedback on learners' spoken production, and that they must keep this in mind when answering the questions. The questions were based on Ellis' guidelines (2009, p. 14) on CF use. See interview guide in Appendix 1.

3.1 Method discussion

A qualitative research method was chosen as this study aims to provide a deeper understanding of the teachers’ actions. Starrin and Svensson say that a qualitative research method aims to “…identify yet unknown, or provide more information about known phenomena, characteristics and meanings, with focus on variation, structures and processes” (my translation) (1994, p. 23), which corresponds with the aim of this study.

The choice to conduct the interviews in pairs was made after an evaluation of the positive and negative aspects of different interview methods. A group interview was considered to be a good choice because of the possibility of getting access to many different opinions in a short amount of time. Another positive aspect of this is that it enables discussion between the participants, which in turn can help the participants investigate and evaluate their own strategies - something that can be difficult to do unless you are reminded of it by hearing about someone else’s strategies. However, by interviewing more than one person at a time there is a risk of focusing too much on a few participants, the ones most prone to share opinions, whilst other participants remain in the background. To eliminate this problem as much as possible but still getting the positive aspects of shared experiences, the decision was made to interview two teachers at a time.

When interviewing more than one participant at a time there is a risk of not getting completely reliable data, as the participants might withhold information that they are reluctant to share with their colleagues, which would potentially affect the result of the research. One way of eliminating this problem as much as possible is to carefully design the questions in a way that would cause the least amount of discomfort for the participants. Before deciding on including a particular question in the interview different types of possible answers to that question had to be considered and evaluated by answering the question “would I be comfortable sharing this information with my own colleagues?”. Of course, this risk cannot be completely eliminated using this method only, which is why an additional evaluation of the results has to be made after completing the interviews.

(12)

12

There are other possible risks of asking the “wrong” questions: either that the questions are designed in a way that when answered it will not result in any relevant information, or that the questions are misinterpreted by the participants. Asking the right sort of questions is an important part in the process of preparing for an interview, but equally important is the ability to be flexible during the interview

(Ely, 1991/1993, p. 72-77), since no answer can be fully anticipated in advance and the interview might change direction depending on such things as the participants’ interpretation of the questions. For this reason, a semi-structured interview was chosen, and an interview guide was designed to lead the participants in the right direction, without leading them towards an anticipated or desired answer. This method of interviewing could also help the situation if participants seem to be holding back information, as follow up questions can be added, and participants can be asked to explain matters in different ways.

3.2 Selection of participants and ethical considerations

The participants in this study consist of four upper secondary teachers of English as a foreign language in a Swedish school who all volunteered to take part in the interviews. The invitation was sent to all English teachers at the same school, which was a deliberate choice as it would narrow down the research area. It was also seen as an opportunity to expose colleagues to differences in their practice, which in turn could lead to further discussion between them and also might increase chances of a more equal learning environment for their students. The teachers who wanted to participate answered the invitation, which meant that everyone had the same opportunity to participate. No attention was paid to variables such as age, gender, years of experience etc. in the selection of participants.

The number of teachers volunteering to participate in the interview eventually became four, since no other teachers at that particular school had enough time to participate. In this situation a decision had to be made whether to contact another school in order to get more participants, or to narrow it down to only include teachers from one school. The latter was chosen, partly because of time constraints, but also as the plan was to conduct the interviews in pairs and it was considered more convenient for the participants to be interviewed at their own work place and with someone already familiar to them.

The participants’ names have been replaced with a letter (A, B, C and D.) in the text, in order for them to remain anonymous, and any details that could potentially reveal their identity is not included in the presentation of collected data. The participants were informed about how the interviews would be carried out and that they were free to change their minds at any time - before, during, or after the interview - and thereby cancel their participance. As the interviews were recorded the participants were also informed about the processing of the data well in advance of the interview via an information letter (see Appendix 2).

3.3 Analysis

Data collected via the recorded interviews was transcribed and analysed in three steps (see Image 1). First, the transcribed interview was colour coded to expose any material in some way related to the research questions (see section 1.1). The next step in the analysing process was to categorize further, by dividing the participants’ answers into a positive and a negative category, in relation to the research questions. For example, one answer to the question “Which of the strategies do you think are the most effective when it comes to learner repair?” was “Definitely metalinguistic

(13)

13

information because that leads further somehow”. This was a shorter piece of text in a longer paragraph which was coloured yellow, as it contained information directly connected to one the research question concerning teacher awareness. The phrase along with other parts of the answer was then coloured green, as it indicated that this teacher was aware of the strategies’ effects to some extent, i.e. a positive indication. Lastly the colour coded material was compiled and compared, with the aim of finding some common denominators, or to see where the teachers’ opinions and/or practice differ. The findings were then used to discuss and answer the research questions.

Image 1. The process of analysis.

4. Result

In section 4.1 the interview questions, as well as a selection of the participants’ answers, will be presented as a part of the analysing process. If a question is answered similarly by all four participants only one answer, or summaries of all answers, will be discussed. However, any answer that contradicts a preceding one is always included or summarized. Section 4.2 will discuss and explain the material further, before the research questions can be answered in section 5.

4.1 Interview

When the interviews began, the participants were first asked to describe their previous knowledge and opinions about corrective feedback (CF for short) before they decided to take part in the interview.

Teacher A:

I guess that, in theory, it must be part of my teacher training but that's a long time ago. But then again in practice that's something else, because it's something we do pretty much every day when teaching I would say.

Participant’s answer related to the research questions (coloured yellow) Positive answer in relation to the research questions (coloured green) 2. Negative answer in relation to the research questions (coloured red) Comparison of results Conclusion

(14)

14

Teacher B:

I didn't recognize it as a field [before the interview] that we discuss […]. I think it's practice, it’s what we do.

Teacher C:

Corrective feedback strategies means […] how we're making corrections, how we correct mistakes the students are doing during conversation, as I have understood. […] Sometimes we don't correct at all because when we are listening to students, for example at the national test, we are listening carefully, we make notes, but we don't make any corrections, because it depends on how much we can understand of the content. That we can understand the message. […] corrective feedback is the issue that we should think about from different angles, different aspects of correction.

Teacher D:

I was not familiar with the term corrective feedback until now. However, I studied some related to feedback in general, primarily concerning formative feedback. […] my knowledge was, as I began teaching, quite limited. Usually you would like to stop a student during their incorrect speaking patterns to help them become aware of “OK this is something that is not correct according to standard English grammar” but just like you [other participant] mentioned there's also a concept of politeness, because if the student is continuously stopped while making mistakes it could become quite an awkward situation. […] so, I think that the idea of how to […] give feedback in the most positive way is something that is very hard, and I feel that I haven't studied much about it when I started the teachers’ program. However, […] these days, if you look at that concept of formative feedback, how you want to make the students aware of patterns, I think I've become quite more familiar with this, because I've become more savvy as a teacher, but also because I've looked up those topics on my own. […] if the students are never made aware of what mistakes they commit, there's no way for them to learn.

Next, the participants were asked to estimate how many times they provide oral corrective feedback of any kind in each English lesson, as an attempt to learn more about their experiences of CF.

Teacher A:

Some sort of average I suppose, because with some classes you don't really have to do that a lot. With some, well you know, you don't want to because you want to encourage them to speak as much as possible, but at the same time help them improve and progress.

Teacher B:

Yeah, I would agree. Sometimes […] maybe you have touched upon certain things you see in grammar or whatever... pronunciation... before, and you can get back to them by referring to that situation “you don't remember when we did that or when I talked about this?”. So, I guess overall as we progress and have more common ground - stuff that we've done together maybe. It's more [CF] if we work on specific preparation for speaking exercises or speeches or that type of thing it might be that we tried to polish certain things. In those situations it can be more, but it's difficult to give an estimate.

(15)

15

[…] overall in a lesson that is specified concerning grammar and the structures of language, maybe 20 times in regular class sizes between 15 to 30 students, so somewhere there.

All participants agree with each other and say that it is difficult to give an estimate as the amount of feedback depends on situation, students, course level, etc.

The interview moved on to investigating the participants use of the different CF strategies by asking them to look at a table with examples and explanations of the strategies, and then tell which of them they think they use regularly in their own teaching. The strategies that they reportedly use are: recast, metalinguistic information, explicit correction, paralinguistic information, and elicitation. Repetition and clarification request were either not mentioned or were said to be avoided strategies by the participants.

Recast, metalinguistic information, and explicit correction were said to be the most common ones, but explicit correction mostly followed by metalinguistic information. The reasons behind their answers will be discussed in section 4.2.

To the question “Do you think the strategies you use are actually helping the learners to increase their proficiency?”, all teachers mostly believed that they were, or as participant B put it:

That is the number one goal, to get them to become better writers and better speakers. So definitely I think what I'm doing is helping. I feel I have a very clear strategy, I know what I want, and I know what I'm after, and I try to I think I make that work when I get the chance […]. I think my overall strategy is to somehow teach them to minimize mistakes […]

Participant C generally sees “a positive development in terms of grammar. However, there are students who still have the same mistakes, the same patterns, and they continue”. The teacher also says that when the students “talk cognitively about their learning they say, ‘I still need to get better at grammar’ but nothing actually happens”.

The next question was designed to make the participants consider the term learner repair, i.e. the learner understands what was incorrect with their speech and then corrects themselves. “Which of the strategies do you think are the most effective when it comes to learner repair?”

Teacher A:

Definitely metalinguistic information because that leads further somehow. If you do it that way, not just the student that has said something, but all of them, can take part in the information process and get what they can from that discussion and improve. I think with a couple of the others [strategies] they simply don't work very well because in speech situations when you say something in front of the group or a lot of times this situation is kind of tense, and someone might be nervous. In a situation where you're supposed to deliver something and say something in front of the group and you’re under the risk of making mistakes and you're aware of the fact that, maybe you don't say the right thing, or that others know more English, or better. It's a situation that's very tense for a student and in those situations you don't learn very much anyway, so even if I were to recast... yeah, they might pick up on it, but if you use some of the others that just makes the student feel uncomfortable, and I don't think they will remember anything from it, and I don't think the others might remember much either. Maybe someone has a really bad experience, maybe others

(16)

16

will see that “I won't make that mistake again”, but I don't think it's a very good strategy to use for learning. So definitely metalinguistic information, for me at least.

Teacher B:

I agree with you there. I'm also thinking of the elicitation one because sometimes, I mean, if you say, “yesterday I go there” and I said, “yesterday I...” it could be picked up as if I didn't quite hear them but they do get a chance to think quickly and see, you know “maybe I need to rephrase this”. Most don't take it that way and can't go in and get it right. Sometimes they persist and then you have to sort of, you know, address it. Ask if there are any other possible ways to phrase it. But maybe with paralinguistic signal... I make this face so to signal, sort of... I'm not sure what it is called, but if you do that face slightly, usually they get it “Aha! OK maybe that wasn't the thing to say”. They can try again before I have the time to use any of these strategies, to make them try to make the it more correct.

Teacher D:

Elicitation because as the corrector repeats part of the learners’ utterance, using rising intonation or a break, or a signal, the learner should complete it. I tell my students when we make correction of the mistakes, try to repeat it twice or three times, and suddenly a student does it correctly himself, but sometimes mirroring [recast] is a good way. Do it correctly and the students repeat it correctly.

Teacher C:

I’m thinking a lot about this and I would argue that number 7, metalinguistic information... I listened to my colleague here who talked about building awareness, and I think that's one of the foundational ideas: we have to make the student aware. But then I also looked at this one particularly when it says “provides clues about language structures” because I'm thinking that we need to make them aware of the structure as to how the language is built, but then also to give them the support to continue, to progress within their proficiency training, and I think that number 7, metalinguistic, is the one that allows us to do that in a certain way. Because it gives them the awareness of the structure “OK this is how it works” but then also give them part of the structure hinting towards the right direction. The students then take those final steps themselves. That's what I think.

The teachers didn’t seem to fully understand the concept of learner repair, as they were more prone to discuss learner acquisition. Some attempts were made to clarify the concept and then the interview moved on to discuss long term knowledge, in hope that the participants would notice the difference. What strategies actually make the learner learn?

Teacher B:

[…] metalinguistic information […]. It has to work for more than the actual situation, and with that approach you can teach how to think to make it correct the next time, and then onwards in the future. […] Some of the others tend to work in that situation too. To give the correct example, maybe it could be useful so that everyone can get the correct version down in their notes, but for the learner to become more self-aware of what strategies to use, what to look for, before speaking.

Teacher A:

I agree with you [Teacher B]. I think that would work with most groups, but if you have a group in English 5 that's not very strong overall in English, I think first of all you have to make sure that they know the terminology of past tense, or present tense,

(17)

17

or progressive... and usually that’s a lot of hard work to teach that to the extent were it really sticks. In my experience […] I think it would be hard to use, you know, “in this context you have to use the past tense to make it work”. I think with those groups maybe recast would work better, but then with other classes and more advanced courses of course, yes metalinguistic probably, because that would give a good explanation for why [rule explanation].

Teacher C:

Number 5 elicitation and number 7 metalinguistic [information]. Once you've given students the support of the structures that they need you also want them to become aware of how to perform the correct grammar and so on, but to do it by their own volition. Therefore number 5 where the example says, “yesterday I...” you indicate to the student “do you recall what we talked about last time?”. The same thing also happens in number 7, you say “yesterday, then you need to...” and you kind of repeat what you intended them to know and also provide the same kind of scaffolds.

The teachers were all in agreement that providing metalinguistic information is most likely helpful when it comes to long term knowledge, but teacher A also sees potential difficulties using that strategy with less advanced groups, since it requires some previous knowledge about language terminology and grammatical structures.

Whether focus is on fluency (i.e. communicational tasks) or accuracy (e.g. grammar and pronunciation) in a particular lesson makes a clear difference for how the teachers use CF. All agreed that they are more likely to use CF of any kind during accuracy focused lessons, since in fluency focused lessons it is more important to allow the learners to speak without anyone interrupting them. Participant B compares it to sports, where accuracy focus is the training, and fluency focus is a match, because that is usually where the students are being assessed.

[…] you could see it as a training and as a match […]. If I have people, you know, doing exercises - speaking situations in pairs or groups - I might be there too, not to interrupt them but to guide them and in that situation, I want to have a more informal setting so that they feel that I am there as a coach [who they can ask for help]

The importance of not interrupting the students was emphasized further when the participants were discussing CF impact on the flow of communication. Teacher C says

[…] you want to build confidence in students, in their language proficiency, and if we are practicing oral communication - supposed to talk about something naturally - it's important to make the students aware of their mistakes, but if the only thing we do is kind of let the hammer of judgement smash everything they say over and over again, we might discourage them of speaking. [It is] Very common to see students who don't want to talk because they've been told repeatedly how poor their English is. So, I would rather improve their fluency with accuracy lessons.

This particular subject was discussed later on in the interview, but first the focus turned to different types of errors and the teachers were asked to tell whether they use different CF strategies depending on the nature of the error.

Teacher C:

[…] if I recognize a pronunciation mistake I would probably use explicit [correction] “no we don't say…” and then something incorrectly, […] whereas if there is a

(18)

18

grammatical one, which is almost like a mathematical explanation “you have to use this kind of grammar at this morphological situation” […] then I probably would use metalinguistic [information].

Teacher B:

I know that I don't want to repeat something that sounds really wrong. I don't want to write it on the board and I don't want to say it too many times. And then there are also delicate things that you don't get the chance, that often, to say certain things. Sometimes you have to take the chance to correct something if it's the type of error that occurs very rarely. Someone might have use for it. It depends on if it's a class situation too, or if you have fewer students or one to one... but yeah, I use different strategies depending on the error, but it's difficult to give examples.

Teacher A:

[…] most likely I do, but I can't think of any examples for sure. I think I'd rather use different strategies depending on the learners.

Teacher A continues to discuss this matter with the next question, which was “Do you think that you use different strategies depending on the learner?”, and specifically refers to different proficiency levels. To this teacher it is more important for learners of lower levels of proficiency to actually speak than to speak correctly, and therefore would rather use recast and hope for learner uptake, than to use for example metalinguistic information. Teacher A also says

[…] if I have a very motivated learner, with a higher ambitions and high goals, they could probably use pretty much any of the strategies. Especially the metalinguistic because they want to know what makes one answer more accurate than the other.

All the other teachers had similar answers, emphasizing proficiency level as the most important factor, but Teacher C also talked about the importance of not differentiating between learners depending on other factors, such as their cultural background, gender, etc.

One could argue that long term knowledge cannot be acquired unless the learners are aware that they have made mistakes and are being corrected, which is why the question “Do you think that your students are aware that they are being corrected?” was included in the interview. Although the teachers for the most part seemed to be uncertain about this, they were all able to give a few examples of methods they use which made them believe that their students understand the purpose of the feedback. Teacher C explained a method in which the students are explicitly told that they are being given feedback via a recording, and all throughout the course the students are asked to evaluate themselves and their progress, using the recorded feedback as support.

Another frequently discussed factor throughout the interview was learner motivation, and again for this topic it was thought to matter a lot. Teacher D says

[…] some pupils are very keen on corrective feedback: they are curious, and they would like the teacher to give them feedback, to explain why they have made a mistake, and they want to understand the purpose of feedback, and they would like to do better. But some students are sometimes insulted, or they don't worry about it - they just want to hand in papers and to get rid of this. So, it depends as well whether the students are really interested to do it better, then the teacher is going to correct

(19)

19

them and to help them to understand why they have made a mistake, but some students just don't worry […]

See section 4.2 for further discussion on the subject of learner motivation.

On the second to last question the participants were asked if they need to target the same types of errors many times (with the same student/group of students), to which they all answered that they do. Some could not mention any specific error types, e.g. grammatical or phonetical, but teacher C points out subject-verb-agreement and differentiating between the v/w sounds, for example viking/wiking*, as especially difficult for Swedish learners of English, but did not specify what strategies are used when those occur.

One aspect of CF is the risk of causing anxiety in the learners. The teachers frequently talked about this throughout the interview, but on the last question they were asked to analyse this further, as well as discussing what they thought can be done to prevent learner anxiety.

Teacher C:

I think if you balance feedback between what they're great at what did you do improve I think the chance of anxiety decreases.

Teacher D:

[…] if you really care about your students here, then we don't give them any negative feedback in a group situation […] we usually give them an individual approach and especially in the Scandinavian environment. The Scandinavian students are very sensitive, are very anxious about making mistakes, and sometimes they don't dare to speak in public. This is the cultural aspect of giving corrective feedback. […] we have to think about this: in Sweden we should use totally different approach to corrective feedback, compared to other countries. Sweden is going to be a multicultural country and global country, but we should keep in mind the sensitivity and the cultural aspects of teaching. […] especially I would like to mention, in Sweden you don't use the red pencil or the red pen to correct mistakes, which in other countries is pretty common - writing a big fat F on the paper.

The answers all indicated that the teachers see a risk of causing anxiety through CF, and that when working with different types of individuals the teacher must be careful with what strategies to use, as well as giving positive feedback to draw attention to what the learners actually know and do well.

4.2 Unexpected results

All four participants were mostly in agreement on all questions during the interview, although the explanations and motivations would vary to some extent. The teachers also answered the questions very thoroughly and carefully and mostly discussed what was asked of them. However, some unpredicted subjects were also discussed quite extensively, and those subjects will be included in this section.

An especially interesting subject in the interview was the strategies the teachers said to be using on a regular basis in their classroom, because although the question was asked in a positive manner, they all tended to also answer it negatively, i.e. saying what strategies they do not use. Repetition was mentioned as an avoided strategy, as it is seen as a particularly harsh way of letting learners know that an error has been made “When you have a class of 30 [students] you want to sort of minimize any embarrassment for students, because I don't think that makes a very

(20)

20

good learning” (Teacher A). This participant says that emphasizing on the error in front of the student’s classmates by repeating the phrase, can cause embarrassment and as a result lead to anxiety. Further, teacher A, C, and D explains that clarification requests are avoided unless they did not actually hear or understand the student’s utterance. The teachers would usually not request clarification as a feedback strategy, i.e. pretend that they did not hear, because “it doesn't indicate that I want them to do something differently. If I always say ‘what’ they have no clue what I expect of them” (Teacher C). The teacher does not see this as an effective feedback strategy since it does not necessarily promote either learner uptake or learner repair: the student can say the exact same phrase without even understanding that it contained any sort of error.

The difference between learners, their personality and level of proficiency, was thoroughly discussed during the interview following the interview questions, but one factor that was not considered when designing the interview was the learners’ motivation, which was thought of as having a big influence on the effects of CF. The short explanation for this was that students who do not want to learn are less likely to notice CF and repair the incorrect utterance. Teacher B even says that if the students are satisfied with their proficiency level they are not very likely to push themselves further, meaning that CF is not effective at all.

[…] motivation is everything, so if they don't want to learn...that happens a lot in a lot of students: are they are great at English and sometimes they go for the good enough level, and sometimes when you really try to get them to the next level, minimize mistakes or whatever... some people just don't go for that next level. They are happy [...]

Although the teachers mostly agreed on all subjects, there were some answers by two participants in the same interview situation (to different questions) that contradicted each other to some extent but was not recognized or discussed by them during the interview. Teacher C, when discussing different CF strategies depending on the learner, said

I hope to be a teacher who doesn't differentiate between learners depending on their culture, gender, ethnicity, and so on. I personally would argue that in a classroom environment those aren't the most important things: they should be treated equally.

This teacher obviously refers to the right to equal education regardless of the students’ background etc. which is different from saying that a teacher should not adapt their strategies to the learners’ needs. Interestingly, the subject was not discussed further when Teacher D answered the last question (about learner anxiety) and emphasized on the importance of thinking about the cultural aspects when giving corrective feedback. When teaching in a multicultural classroom, as many are in Sweden, there is a possibility of meeting students with previous experiences of teaching methods and learning environments that are different from those most commonly practiced in Sweden, which is what Teacher D was referring to. However, since no discussion took place between the participants it is unclear whether they agreed with each other on this subject, or not.

(21)

21

5 Analysis

The following sections will further explain, analyse, compare and discuss the collected data in relation to the selected previous research.

5.1 What CF strategies do teachers use?

The interviews clearly displayed examples of CF strategies the teachers use, and also the different situations in which they are put into practice. All teachers said that they use recast, metalinguistic information, explicit correction, paralinguistic information, and elicitation on a regular basis in their classrooms, and as can be seen throughout the interview their strategies vary to some extent when they need to consider factors such as learners, errors, short/long term knowledge etc. It seems that recast is the preferred strategy for teachers with learners of lower proficiency levels, and metalinguistic information was mentioned in connection with higher proficiency levels.

The types of errors that are being made in the learning environments proved difficult to describe, but generally it seems that explicit corrections are made if the error lies within the pronunciation, and metalinguistic information is given if the error is of a grammatical nature. This could be interpreted as a way of being less interruptive of the student’s spoken language than for example metalinguistic information, and therefore less embarrassing, but still providing the learner with a more correct pronunciation. Contradictory to this interpretation, one teacher said that metalinguistic information is given following explicit corrections, which would then lead to a longer interruption of the student’s utterance. When it comes to grammatical errors some teachers feel that it is more important to explain and teach the students about language structure. However, this does not say that the teachers are more prone to interrupt the students just because the error is of a grammatical nature, but their answers, when asked if they are more likely to give feedback during fluency versus accuracy focused lessons, all said that accuracy focused lessons will usually result in more CF. This subject will be discussed further in the next section.

5.2 Teachers’ awareness of the effects of different CF strategies

This subject proved to be more complex than anticipated. Even though the teachers shared thoughts and experiences that indicate a great understanding of the concept of corrective feedback, they were not fully able to motivate their answers without some contradiction. Effects on learner repair versus long term knowledge seemed to be the most difficult question, partly because the difference between the concepts were misunderstood, but mostly since the participants did not get a chance to prepare or collect any sort of “data” from their classrooms before the interview. This meant that they could only guess what strategies best enables learner repair. Long term knowledge on the other hand, though not either entirely reliable because of the same issue, was easier for the participants to give a qualified guess about. Their explanations to why, for example, metalinguistic information would lead to long term knowledge suggest awareness of corrective feedback effects.

All participants were fairly certain that they must target the same sort of errors repeatedly, which could be interpreted as a lack of understanding of the CF effects. It would seem logical to change strategy when noticing that the same error occurs many times with the same group/learner, and if the teachers where fully aware of the CF effects there is a possibility that the same sort of errors would not occur at

(22)

22

all. However, the teachers did not get to learn the difference between errors and mistakes (Corder, 1967, p. 167) during the interview, which could make a difference for the participants’ answers, and there are several other factors that could affect the outcome as well.

The participants do however emphasize the importance of not interrupting the students, and their answers indicate awareness of a risk of interfering with the flow of communication, which makes them choose strategies that will interfere to a lesser extent. Their answers also clearly show that their students’ wellbeing is utterly important, by choosing CF strategies to avoid embarrassing students and causing anxiety. One teacher said that it is often more important to get the learners to speak at all than to produce correct spoken language, and other participants had encountered students who would barely speak at all because of previously uncomfortable experiences in the learning environment, and therefore the teachers thought it as important to balance out the corrective feedback with that of a more positive nature. They would reportedly choose strategies which minimized the risk of causing embarrassment, such as recast.

5.3 Participant's strategies compared to previous research results

The participants’ answers both correspond with the selection of previous research on some points, but also contradict it on other points. Teachers in the previously discussed research (section 2 and 2.1) often preferred methods of correction that proved to be not very effective when it comes to learner repair and long-term acquisition. Those teachers reportedly chose those strategies consciously based on other aspects than learner repair and long-term knowledge, i.e. less interference with the flow of communication and lower risk of causing anxiety. The teachers in this study accounted for some of the same things: even though they recognized the more explicit varieties of CF, e.g. metalinguistic information, explicit correction, paralinguistic signals and elicitation, as more effective for learner repair and long-term acquisition, they often preferred to use recast. This was usually explained and motivated by viewing the implicit varieties, recast in this case, as “nicer” means of correction since it does not put focus on the error or on the student that has uttered the erroneous phrase. These teachers also emphasized on not embarrassing the students which could discourage them from speaking all together.

The negative aspects of avoiding correcting students explicitly is of course the risk of not getting through to them about the error, because the learners might not notice the error at all or will not understand how to correct it. Harmer (2007) discussed the importance of learner awareness, and that teachers must explain both the purpose of CF it and how it will be used, which could lead to a greater understanding, more uptake and hopefully repair and long-term knowledge (p. 144). In the current study all teachers believe that their students are indeed aware that they are being corrected, although since they also stated that they have to correct the same types of errors many times it is possible that the students are not as aware as the teachers think. One teacher had a strategy for informing the students about the feedback, but since the effects did not always seem to meet the expectations it is difficult to analyse the level of awareness without further research on that subject.

Other recommendations by Harmer (2007) concerned the difference between fluency focused lessons and accuracy focus. Accuracy focused lessons should target mistakes/errors and raise awareness, i.e. using explicit corrections that prompt learner repair. Fluency focused lessons, on the other hand, should include content and meaning, and errors should be treated more implicitly without the teacher

Corrective Feedback in Oral EFL Learning Environments : A Study on Swedish Teachers’ Awareness of Corrective Feedback Strategies and Effects

Examensarbete

Grundnivå

Corrective Feedback in Oral EFL Learning

Environments

A Study on Swedish Teachers’ Awareness of Corrective

Feedback Strategies and Effects

Table of contents

1. Introduction

2. Literature review

3. Method

4. Result

5 Analysis