Oral Corrective Feedback in the EFL Classroom. An Empirical Investigation of Forms- and Message-Focused Phases of EFL Lessons

Table of Contents

1 Introduction

2 Theoretical Background
2.1 Oral Linguistic Errors and Mistakes and Different Error Types
2.2 Types of Oral Corrective Feedback
2.3 Factors which Influence the Treatment of OCF in the English Foreign Language Classroom
2.3.1 Phases of a Lesson
2.3.2 Oral Corrective Feedback Concerning Different Ages of the Learners
2.3.3 Teacher Preference Concerning OCF Types
2.4 Research Question and Hypotheses

3 Methodology
3.1 Informants of the Study and the Data Collection Procedure
3.2 The Employed Research Instrument
3.3 Data Handling
3.4 Methodological Strengths and Limitations

4 Results
4.1 The Extent of OCF Concerning Different Error Types in Forms- and Message- Focused Phases of a Lesson
4.2 The Usage of Different Feedback Types in Forms- and Message-Focused Phases of a Lesson
4.3 The Extent of OCF Concerning Different Ages of the Students in Forms- and Message- Focused Phases of a Lesson
4.4 The Usage of Different Feedback Types Concerning Different Ages of the Students in Forms- and Message-Focused Phases of a Lesson
4.5 The Extent of Teacher Preferences Concerning Different OCF Types in Forms­and Message- Focused Phases of a Lesson

5 Discussion
5.1 The Extent of OCF Concerning Different Error Types in Forms- and Message- Focused Phases of a Lesson
5.2 The Usage of Different Feedback Types in Forms- and Message-Focused Phases of a Lesson
5.3 The Extent of OCF Concerning Different Ages of the Students in Forms- and Message- Focused Phases of a Lesson
5.4 The Usage of Different Feedback Types Concerning Different Ages of the Students in Forms- and Message-Focused Phases of a Lesson
5.5 The Extent of Teacher Preferences Concerning Different OCF Types in Forms­and Message- Focused Phases of a Lesson

6 Conclusion

References

Appendices

Appendix 1: Error and Mistake Categories

Appendix 2: Oral Corrective Feedback Types, Examples and Their Advantages and Disadvantages

Appendix 3: Advantages and Disadvantages on OCF in General

Appendix 4: Studies Referred to in This Master Thesis

Appendix 5: Professionelle Kompetenz von Lehrkräften

Appendix 6: Observation Sheet of the Pre-test and Its Modifications

Appendix 7: Observation Sheet

Appendix 8: Implicit OCF Strategies in Forms- and Message-Focused Phases

Appendix 9: Explicit OCF Strategies in Forms- and Message-Focused Phases

Appendix 10: Reformulations in Forms- and Message- Focused Phases

Appendix 11: Prompts in Forms- and Message- Focused Phases

Appendix 12: Contrast of the Distribution of Feedback Types Concerning Different Error Types in Phases that Focus on Forms and Message

Appendix 13: Absolut Numbers Concerning Age Differences and Feedback Types in Forms- and Message-Focused Phases

Appendix 14: Absolut and Relative Numbers Concerning Age Differences and Explicit and Implicit OCF Strategies in Forms- and Message-Focused Phases

Appendix 15: Absolut and Relative Numbers Concerning Age Differences and Prompts and Reformulations in Forms- and Message-Focused Phases

Appendix 16: Teacher Preferences Concerning Different Feedback Types in Forms­and Message-Focused Phases

Appendix 17: Absolut and Relative Numbers Concerning Teacher Preferences in Regard to Explicit and Implicit OCF Strategies in Forms-Focused Phases

Appendix 18: Absolut and Relative Numbers Concerning Teacher Preferences in Regard to Explicit and Implicit OCF Strategies in Message- Focused Phases

Appendix 19: Absolut and Relative Numbers Concerning Teacher Preferences in Regard to Prompts and Reformulations in Forms-Focused Phases

Appendix 20: Absolut and Relative Numbers Concerning Teacher Preferences in Regard to Prompts and Reformulations in Message-Focused Phases

List of Tables

Tab. 1 Classification of OCF Strategies According to Ranta & Lyster (2008, p. 152)

Tab. 2 Taxonomy of OCF Strategies According to Ellis (2009, p. 8) and Sheen and Ellis (2011, p. 594)

Tab. 3 Information on the Observed Classes of this Study

List of Figures

Fig. 1 Learner Errors and Types of Oral Corrective Feedback (adapted from Lyster & Ranta 1997, p. 44; Argüelles et al., 2019, p. 110)

Fig. 2 Codification of Feedback Types

Fig. 3 Distribution of All Errors Types (n = 592)

Fig. 4 Distribution of all Feedback Types (n = 592)

Fig. 5 Correction of Errors Concerning the Phase of the Lesson

Fig. 6 Distribution of All Error Types and Their Correction in Phases That Focus on Forms (n=347)

Fig. 7 Distribution of All Error Types and Their Correction in Phases That Focus on Message (n=241)

Fig. 8 Contrast of the Distribution of Feedback Types in Forms Focused and Message-Focuses Phases

Fig. 9 Correction of Errors Concerning the Grade and the Phase of the Lesson

Fig. 10 Age differences Concerning Different Feedback Types in Forms-Focused Phases

Fig. 11 Age differences Concerning Different Feedback Types in Message-Focused Phases

Fig. 12 Teacher Preferences Concerning Different Feedback Types in Forms- Focused Phases

Fig. 13 Teacher Preferences Concerning Different Feedback Types in Message-Focused Phases

List of Abbreviations

Abbildung in dieser Leseprobe nicht enthalten

Introduction

“The greatest mistake you can make in life is to be continually fearing you will make one” (Hubbard, 1927, p. 170). This statement does not only refer to the behaviour of many people in their daily lives, it does also especially refer to students in English as a foreign language (EFL) classrooms who rather do not participate during lessons than risk making an error or a mistake as they are afraid of losing their face (cf. Decke-Cornill & Küster, 2015, p. 155). However, errors and mistakes are of high importance from a diagnostic point of view as they hint at the current level of students, their course of the learning process, their success or failure of their learning strategies as well as what type of support they might need (cf. Kieweg, 2007, p. 3; Vetter, 2007, p. 36; Haß, 2017, p. 355). And even further, students are also in need of feedback to check their own ability in the foreign language and not to consolidate erroneous grammatical structures, vocabs or an incorrect pronunciation (cf. Timm, 2009, p. 203). Hence, whenever an oral error or mistake occurs in the EFL classroom, teachers have to deal with the situation by deciding whether the specific error or mistake should be corrected or not and if so, how it should be corrected to support the learning process of the student without inhibiting him or her on an emotional level. Therefore, the topic of oral corrective feedback (OCF) is relevant for all EFL teachers and their daily practice in class. Studies concerning this topic as part of applied linguistics have their origins in 1977 when Craig Chaudron first examined reactions of the target language speaker to the second language learner's errors. Ever since, there has been a growing interest in the dealing with errors in the context of second or foreign language acquisition (cf. Havernek, 2002, p. 36). Although many studies exist in this field, there are only a few studies conducted in German schools. As these German studies rather investigated OCF in the Italian and Spanish foreign language classroom than in the EFL classroom (cf. Klepping & Königs, 1991; Lochtmann, 2002) or were rather concerned with students' feelings in the context of OCF than the actual correction by the teacher (cf. Havernek, 2002), further insights into the topic of OCF in German EFL classrooms were needed. Therefore, an empirical project has already been conducted by me during the practical semester in three classes of grade seven. As 30 hours of observation have not been enough to gain representative data and further factors that influence the way of giving OCF in forms- and message-focused phases of a lesson were not included in the examination, additional investigations in this field were required.

Therefore, this master thesis builds on the results of that study and aims again at answering the following guiding question by widening the sample size and investigating additional factors:

- In how far is oral corrective feedback given in forms- and message-focused phases in the EFL classroom?

To form a basis for the research, the theoretical background on the topic of OCF is presented in the second chapter which contains information on the distinction of oral linguistic mistakes and errors, on types of OCF as well as on situations in which it takes place, on OCF concerning different ages of the learners and on teacher preferences regarding OCF types. Based on that chapter, the research question and its sub-questions as well as relating hypotheses are presented in chapter 2.4. Further, the third chapter is concerned with the methodology used for this study containing information on the observed students and teachers, the data collection procedure, the employed research instrument, the data handling as well as methodological strengths and limitations. Afterwards, the results of this study are listed and described in chapter four and discussed in chapter five. The main outcomes of the whole study are summarised in a short conclusion, implications for the classroom are given as well as prospects for further investigations in this field based on limitations of this study are stated in chapter six.

2 Theoretical Background

This chapter contains information on oral linguistic errors and mistakes, different error and feedback types as well as factors which influence the dealing with OCF in the classroom.

2.1 Oral Linguistic Errors and Mistakes and Different Error Types

In the context of OCF on inaccurate spoken output of learners of a foreign language, several foreign language researchers and EFL academics differentiate between the terms ‘error’ and ‘mistake’. As there is no generally accepted definition for each of these terms, the most common definitions are presented in this chapter. According to Lewandowski (1990, II, p. 297 cited in Klippel & Doff, 2015, p. 198), both, ‘errors’ as well as ‘mistakes’, are deviations from the existing linguistic norm, contraventions of linguistic accuracy, linguistic rules or linguistic appropriacy as well as forms that can lead to misunderstandings or difficulties in communication. Based on that, Corder (1967, p. 167) already distinguished between ‘errors' and ‘mistakes' in 1967. ‘Errors' are faults that take place as a result of a lack of knowledge or competence. They occur systematically and reveal the actual learners' level at a specific stage. This definition is still valid and used by EFL academics such as Macht (1998, p. 353), Timm (2009, p. 205) and Grimm et al. (2015, p. 285). These academics add that errors are not recognised by the learner, on account of which they are likely to occur repeatedly.

In contrast, ‘mistakes' are defined by Corder (1967, p. 167) as a result of processing failures which occur unsystematically. He mentions that ‘mistakes' are of no significance to the process of language learning and further EFL academics add that this type of faults might only be performed once. As ‘mistakes' are recognised by the learner in many cases, they are often corrected by them. ‘Slips of the tongue' and ‘lapses' can belong to the category of ‘mistakes' as they are based on a weakness of performance (cf. Corder, 1967, p. 167; Macht, 1998, p. 353; Timm, 2009, p. 205; Grimm et al., 2015, p. 287). Haß (2017, p. 356) preponderantly agrees with these definitions but distinguishes ‘slips' from ‘mistakes'. He states that only mistakes point out a lack of performance. They might occur when a student has already understood the linguistic concept in general but needs more practice for a secure handling of it. ‘Slips' occur in case that students who have already understood a specific concept are inattentive while speaking (cf. ibid.). Kieweg (2007, p. 4) provides different definitions for both terms. He states that only ‘errors' represent deviations from the existing linguistic system, as well as contraventions of linguistic accuracy, rules and appropriacy, while ‘mistakes' are deviations from existing linguistic conventions within speech situations in terms of style. ‘Mistakes' refer, for instance, to the usage of inappropriate items, idioms or discourse organisation (see appendix (app.) 1).

By referring to linguistic correctness, the field of grammar, lexis and pronunciation can be distinguished in which linguistic errors¹ are likely to occur in the EFL classroom. The use of a wrong tense as well as the misuse of auxiliaries, of the active and passive, of relative clauses or of conditional clauses can be, for instance, categorised as grammatical errors while the misuse of prepositions and conjunctions as well as the usage of non-existing words or the wrong word in a specific context can be classified as lexical errors. Errors in pronunciation are made whenever wrong phonemes are uttered for a specific word (cf. Ministerium für Schule, 2007, pp. 32-33). A fourth error category ‘the unsolicited use of the first language (L1)' can be added according to Lyster and Ranta (1997, p. 45). This error can be recorded in German EFL classrooms whenever students use a German word within their utterance as an indication that they do not know or that they have currently forgotten the correct English translation (cf. Aranguiz & Espinoza, 2016, p. 108).

2.2 Types of Oral Corrective Feedback

OCF in the EFL classroom takes the form of a response of the teacher or students to an erroneous utterance of one student and it contains a correction or guidance for self­correction (cf. Ellis, 2009, p. 3; Saville-Troike, 2012, p. 115; Lyster et al., 2013, p. 1; Klippel & Doff, 2015, p. 199). To respond to the before mentioned learner error types, there are different OCF strategies, which were first identified through observational studies such as the ones conducted by Lyster and Ranta (1997, pp. 46-49) or Ellis (2009, p. 9) and are defined by several researchers as investigations in the field of OCF steadily grow. To avoid definitional fuzziness, the most recent and representative feedback strategies are presented in figure (fig.) 1 under the caption “teacher strategies”.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1 Learner Errors and Types of Oral Corrective Feedback (adapted from Lyster & Ranta 1997, p. 44; Argüelles et al., 2019, p. 110)

Explicit correction (1) is a direct and explicit reference to the committed error. This might take place by mentioning that in general an error has been committed, by hinting at the type of error that has been committed or by directly correcting the erroneous word or utterance (cf. Lyster & Ranta, 1997, p. 46; Timm, 2009, p. 214; Saville-Troike, 2012, p. 115; Pawlak, 2014, p. 136). Recasts (2) which are also often referred to as ‘implicit reformulation’ display the opposite of the previous mentioned OCF strategy as the committed error is not made overt to the student and the class. The teacher rather embeds a correction in his utterance by reformulating a part or the whole utterance of the student and by maintaining the original meaning of the student’s utterance (cf. Timm, 2009, p. 220; Pawlak, 2014, p. 136; Thaler, 2014, p. 318; Aranguiz & Espinoza, 2016, p. 109). Error repetition or echoing (3) as a third OCF strategy takes place whenever a student’s error or the whole erroneous utterance is repeated by the teacher. The error might be highlighted by specific intonation (cf. Pawlak, 2014, p. 137; Thaler, 2014, p. 318; Aranguiz & Espinoza, 2016, p. 109). Elicitation (4) refers to three techniques aiming at eliciting the correct form from the student. This might take place by repeating the student’s utterance up to the last correct word with raising intonation, by using a question to elicit correct forms or by asking the student to reformulate the utterance (cf. Pawlak, 2014, p. 129; Aranguiz & Espinoza, 2016, p. 109; Pedrazzini, 2017, p. 104). Metalinguistic feedback (5) includes specialist terminology used by the teacher to direct the attention of the learner to a specific rule of the foreign language. This OCF strategy does not provide the correct form but it is often connected with a form of explicit correction (cf. Pawlak, 2014, pp. 136-137; Pedrazzini, 2017, p. 104). A clarification request (6) as a sixth strategy of OCF indicates to the students either that the utterance was not well-formed or that the utterance has been misunderstood (cf. Grimm et al., 2015, p. 188; Maolida, 2017, p. 182; Pedrazzini, 2017, p. 104). Paralinguistic feedback or non­verbal ways (8) of OCF can be labelled as another feedback strategy (cf. Ellis, 2009, p. 9). Hereby, an error is indicated by using a gesture or mimicry (cf. Maolida, 2017, p. 182). Further, it is possible to use sounds (9) to point out that an error occured (cf. Macht, 1998, p. 362; Timm, 2009, p. 223; Scrivener, 2011, p. 289).²

According to Ranta and Lyster (2008, p. 152), OCF strategies can be categorised into reformulations or prompts (see table (tab.) 1). Reformulations provide learners with the correct reformulation of their erroneous utterance while prompts withhold correct forms and instead comprise a variety of signals, other than reformulations, that request some kind of self-repair from the learner (cf. Ranta & Lyster, 2008, p. 152).

Table 1 Classification of OCF Strategies According to Ranta & Lyster (2008, p. 152)

Abbildung in dieser Leseprobe nicht enthalten

Another taxonomy of OCF strategies, which is presented in tab. 2, is based on the distinction between (a) explicit vs. implicit OCF strategies and (b) input-providing vs. output-prompting OCF according to Ellis (2009, p. 8) and Sheen and Ellis (2011, p. 594). Explicit OCF strategies refer to feedback moves that overtly point out that an error took place while implicit OCF strategies are given whenever the OCF move occurs without an overt indicator of an error occurrence during the course of interaction (cf. Aranguiz & Espinoza, 2016, p. 108). Analogous with Ranta's and Lyster's (2008, p. 152) distinction between reformulations and prompts, input-providing OCF is considered as OCF that provides learners with the correct form, whereas output-prompting takes place when the correction of the form is elicited from the learner (cf. Ellis, 2009, p. 8; Pedrazzini, 2017, p. 104). Therefore, input-providing error correction does not necessarily require the undertaking of a repair of the incorrect utterance, while this is the case when output­prompting OCF strategies are used. Output-prompting OCF strategies are also often referred to as the ‘negotiation of form' (cf. Ranta & Lyster, 2008, p. 153; Pawlak, 2014, p. 141).

Table 2 Taxonomy of OCF Strategies According to Ellis (2009, p. 8) and Sheen and Ellis (2011, p. 594)

Abbildung in dieser Leseprobe nicht enthalten

As also presented in fig. 1, in addition to these eight OCF strategies, further OCF types can be categorised concerning the strategy provider and the time of the correction. Peer feedback (7) can be initiated by the teacher by asking for instance whether anyone spotted an error or whether anyone is able to reformulate the erroneous utterance of his or her classmate. Further, it could also be offered by the students themselves in form of any of the before mentioned OCF moves. Apart from peer feedback, the student who made the erroneous utterance could also take over the part of the strategy provider by correcting himself / herself after the occurrence of the error (cf. Klippel & Doff, 2015, p. 201). Here, this is categorised as self-correction (12). Concerning the time of the correction, immediate OCF and delayed corrective feedback (10) are distinguished. While the former OCF move directly follows the student's erroneous utterance, the latter takes place at the end of a longer talk of a student, a student's poster presentation or in general at some time during the lesson without explicit reference to the specific student (cf. Timm, 2009, p. 223; Grimm et al., 2015, p. 188). It is also possible that no feedback (11) is given as might be useful in phases which focus on message (see chapter (chap.) 2.3.1). Examples of each of these types of OCF as well as their advantages and disadvantages are presented in app. 2.

2.3 Factors which Influence the Treatment of OCF in the English Foreign Language Classroom

Discussing the advantages and disadvantages of each OCF type individually, as presented in app. 2, is not sufficient to decide in which way OCF should be given in the EFL classroom. In general, opinions on the question whether OCF should be given at all are extremely divided. The most representative reasons for and against OCF are presented in app. 3. Nevertheless, as both sides offer valid arguments, teachers seem to be in a dilemma in terms of giving oral feedback. Hence, to meet students' affective and cognitive needs equally, an awareness of suitable situations for giving feedback is required. Therefore, teachers should be aware of various factors which interact with OCF effectiveness and hence interfere with the decision of giving feedback or not and if so, which type of feedback moves should be used. These factors might be the situation of the lesson, whether the focus lies on forms or on message, the age of the learners as well as teachers' preferences for certain OCF types (cf. Lyster et al., 2013, p. 30; Fan, 2019, p. 198). The way in which these factors influence the oral correction in the EFL classroom are described against the backdrop of previous research. An overview of all studies referred to is given in app. 4.

2.3.1 Phases of a Lesson

Two general foci of a phase during a lesson can be distinguished which meet different conditions in terms of OCF: phases that ‘focus on forms’ and phases that ‘focus on message’³ (cf. Timm, 2009, p. 214; Scrivener, 2011, p. 286; Klippel & Doff, 2015, p. 202).

Phases of a Lesson that Focus on Forms

Phases during a lesson which ‘focus on forms’ concentrate on deliberate grammar or vocabulary teaching which aims at the student’s understanding and accurate production of new grammatical structures or the new vocabulary in written and spoken English (cf. Graham & Parry, 2007, p. 19). This includes systematic phases of practice of these particular forms such as text-manipulation activities (cf. Pawlak, 2014, p. 248). In this context, English researchers agree that erroneous utterances that are related to the object of study should always and immediately be corrected (cf. Timm, 2009, p. 217; Scrivener, 2012, p. 286; Thaler, 2014, p. 318; Klipple & Doff, 2015, p. 202). This is based on the aspect that the student’s attention is maximally focused while and shortly after having uttered a sentence including a new form, on account of which this kind of “teachable moment” (Pawlak, 2014, p. 118) should not be missed by postponing corrective feedback (cf. ibid.). Further, several classroom-based studies conducted by for instance Carroll and Swain (1993), Ellis et al. (2006), Ammar and Spada (2006), Sheen (2007), Dilans (2010), Lyster and Saito (2010), Yang and Lyster (2010) and Yilmaz (2012) investigated the effectiveness of OCF on the acquisition of a variety of different grammatical and lexical forms and came up with the results that explicit OCF strategies as well as prompts are more effective than recasts. Further, there are only a few studies like the two studies by Saito & Lyster (2012a; 2012b) that investigated the role of OCF in the phonological second language (L2) development. These studies suggest that recasts can play a crucial role in this development but again explicit OCF strategies might be needed to ensure larger effects especially in difficult cases of the target language. Another study undertaken by Mackey et al. (2007) shows that explicit OCF strategies are in general more accurately perceived by students than implicit OCF strategies. And further studies such as the ones by Gitsaki and Althobaiti (2010), Aranguiz and Espinoza (2016) or Maolida (2017) that measured the effectiveness of different feedback strategies concerning their uptake by the students are in conformity with the before mentioned results. Based on all of these studies, Pawlak (2014, p. 127) concludes that OCF should be overt enough in forms-focused phases in order to be noticed by the learners. Nevertheless, a classroom-based study carried out by Pawlak and Tomczyk (2013), as well as computer-aided studies conducted by Loewen and Erlam (2006) and Sauro (2009) came up with the result that implicit OCF strategies like recasts might indeed not be more effective than explicit OCF strategies but at least, they can show the same effects. However, as uptake cannot guarantee in each way that a feedback strategy is effective and as most of the other studies based on pre- and post-tests concerning the intervention only focused on the acquisition of one grammatical, lexical or phonological feature, Sheen (2011, p. 165) argues that these results will not necessarily be representative of all of these linguistic features. Moreover, Thaler (2014, p. 318) states that a variation in the usage of feedback types is useful to address different learner types.

Phases of a Lesson that Focus on Message

In phases which ‘focus on message', a meaningful communication plays a significant role. In that case, language is not treated as an “object of study” (Long & Robinson, 1998, p. 18), as might be the case within the before mentioned forms-focused phases, but rather as a “medium of communication” (ibid.). This might be the case within oral presentations of students, discussions, simulations, role plays, creative writing tasks or any kind of conversation with a focus on a specific content (cf. Timm, 1998, p. 362; Pawlak, 2014, p. 118). Even though the slogan “message before accuracy” (Timm, 1998, p. 362) is generally widely accepted, which implies that OCF is reduced to a minimum, two types of errors should always be corrected in these phases: 1) errors which interfere with the comprehension of an utterance and might lead to a communicative breakdown and 2) errors which violate social conventions and norms (Wulf, 2001, p. 118). Timm (2009, p. 218), Scrivener (2011, p. 285), Thaler (2014, p. 317), Klippel and Doff (2015, p. 201) and Haß (2017, p. 355) agree that grammatical errors play a less significant part in contributing to these error types than the unsolicited use of the L1, lexical errors and errors in pronunciation as well as errors based on a lack of social and intercultural competence. Therefore, errors such as grammatical errors, slips of the tongue or errors which do only marginally affect the communication do not necessarily have to be corrected (cf. ibid.). Nevertheless, Timm (2009, pp. 217-218) and Grimm et al. (2015, p. 287) also state that some grammatical errors can be corrected if the risk of consolidation of this particular error is too high or if it occurs too frequently. Further, it is important to keep it transparent to the students when OCF is neglected to avoid confusion about expected, but not given OCF (cf. ibid).

Whenever OCF takes place in phases whose overriding focus is on meaning and communication, it is referred to as ‘focus on form' that implies “an occasional shift of attention to linguistic code features [...] triggered by perceived problems with comprehension or production” (Long & Robinson, 1998, p. 23). EFL academics' opinions on how feedback should be given in these phases are divided. Edge (1989, p. 37), Timm (1998, p. 316), Hedge (2000, p. 291), Scrivener (2011, p. 225), Pawlak (2014, p. 119) and Thaler (2014, p. 318), for instance, advocate to give delayed feedback following a communicative act of the students. Scrivener (2011, p. 225) and Thaler (2014, p. 318) advise teachers to note down common errors, to sum them up and to give compendious feedback afterwards or to create appropriate exercises for further lessons. By giving delayed feedback, the flow of communication is not interrupted, the focus on meaning remains and students who are likely being intimidated or discouraged from future participation are supported (cf. Scrivener, 2011, p. 225; Pawlak, 2014, p. 119). Further, Timm (1998, p. 316), Wulf (2001, p. 118) and Scrivener (2011, p. 227) mention that it is important to give students enough time for self-correction or to contingently use feedback types which foster student-generated repair such as prompts. Scrivener (2011, p. 227) categorises these types of feedback under the term ‘scaffolding' as they offer helpful strategies for students to create their own spoken structure. Timm (1998, p. 362) and Pawlak (2014, p. 132) also advocate recasts in these phases as they might be embedded in the conversation without obstructing the communicative flow and without devaluing the student's efforts. Moreover, using explicit correction during phases that focus on meaning is rejected by several researchers such as Seedhouse (2004, p. 153), Pawlak (2014, p. 249) and Grimm et al. (2015, p. 288). Nevertheless, Timm (1998, p. 362) states that it is possible for the teacher to whisper the correction explicitly to the student in a discreet and cooperative way.

In contrast to the wealth of studies conducted concerning the effectiveness of different OCF strategies in forms-focused phases, only little research was done in this field regarding message-focused phases. Even further, the only studies such as the ones by Ellis et al. (2001), Panova and Lyster (2002), Maolida (2017) and Pedrazzini (2017) investigated the frequency of different feedback strategies in these phases instead of their effectiveness. The results of Maolida’s (2017, p. 191) as well as of Pedrazzini’s (2017, p. 110) study show in both cases that input-providing OCF strategies are more often used in message-focused phases than in forms-focused phases. Concerning this matter, Maolida (2017, p. 191) argues that a possible reason might be the teacher’s choice to directly continue with the topic and not to interrupt the fluency of the student’s speech too much. Ellis et al. (2001, p. 308) and Panova and Lyster (2002, p. 587) correspond to these results. Nevertheless, all of them establish that recasts are significantly more often used than explicit correction. It can only be speculated about reasons for these results on the basis of the advantages of recasts described in app. 2.

In conclusion, as there is no sufficient empirical basis for the effectiveness of different OCF strategies in message-focused phases and some theoretical benefits can be found for each feedback type, no definite statement about the most useful manner to give OCF in these phases can be made at this point (cf. Pawlak, 2014, p. 189).

2.3.2 Oral Corrective Feedback Concerning Different Ages of the Learners

The specific role of age in second language acquisition is still not clearly understood as Mackey and Oliver (2002, p. 459) point out. Even further, there are only a few studies that examined age differences in the context of OCF (see e.g. Oliver, 2000; Mackey & Oliver, 2002; Lyster & Saito, 2010). According to Oliver (2000, p. 143), who investigated ten teachers working with primary school aged English as a second language (ESL) students at the age of six to twelve and with adult ESL students, younger learners are generally more sensitive to the effects of OCF than older students. He mentions that the age of child learners corresponds to a time in their cognitive and psycholinguistic development during which the provision of OCF is most beneficial for their learning process (cf. ibid.). Mackey and Oliver (2002, p. 471), who observed oral corrective feedback in ESL classrooms with 22 children at the age of eight to twelve and compared these results to studies conducted with adult ESL learners, confirm that OCF has higher effects on child learners than on adult learners. Nevertheless, they admit that there are also effects on older learners, but these are less immediate in comparison to the effects on children's L2 (cf. Mackey & Oliver, 2002, p. 475). The study conducted by Lyster and Saito (2010, p. 289), that examined fifteen studies in the context of OCF concerning different treatments in relation to age difference, also points out that younger learners are more affected by OCF than older learners. They mention that a possible reason could be that implicit learning mechanisms are engaged which are more characteristic of younger than of older learners (cf. ibid., p. 293).

Moreover, Lyster and Saito (2010, p. 282) argue that younger learners benefit from prompts and explicit correction more than from recasts, while older learners benefit from all of these OCF strategies to a similar extent. This might be explained as child learners generally need more guided support to detect linguistic information in classroom input (cf. Munoz, 2006, p.12). In contrast, older language learners already have substantial analytical abilities, which is why it is likely that they are able to notice linguistic information in an autonomous manner (cf. Lyster et al., 2013, p. 27). Instead, Mackey and Oliver (2002, p. 473) argue that especially child learners show a great sensitivity to recasts and prompts as these OCF strategies are functionally similar to L1 feedback given by their caregivers.

Finally, Mayo (2003, p. 106) concludes from her general study on the quality of performance with regard to the length of exposure that the longer the exposure to the L2 is, the better the performance becomes. Therefore, older students who started to acquire a foreign language at a young age are likely to make less errors in comparison to mistakes⁴ that are made by them. Based on that, Mayo's (2003, p. 106) study shows that older learners tend to correct themselves more often than younger learners.

From these studies, it can be concluded that both, child learners and older learners, need and benefit from OCF, but child learners need it more often. While all OCF types seem to effect older learners in similar ways, the research on appropriate OCF strategies for younger learners has still not been sufficiently investigated, on account of which no conclusion can be made at this point.

2.3.3 Teacher Preference Concerning OCF Types

EFL academics such as Kleppin (2009, p. 68) and Timm (2009, p. 226) state that teachers often follow their own corrective rituals during their lessons. Further, Timm claims that in many cases, the correction does not only depend on the before mentioned factors such as the type of the error, the phase of the lesson or the age of the students, but likewise on individual preferences for specific feedback types. Nevertheless, with reference to the model of Baumert et al. (2011, p. 23), it can be claimed that these preferences are influenced by cognitive decisions regarding the before mentioned factors based on the teacher's acquired teaching skills. Even further, they are also influenced by non-learnable attitudes and motivations (see app. 5).

The study conducted by Blex (2001) investigated nine teachers regarding their self­assessments with a view on their preferences for specific OCF types, reasons for these, as well as their actual corrective behaviour in class. It turns out that four of these teachers show different preferences ranging from implicit to explicit OCF strategies, as well as from prompts to input providing OCF moves, while the other five do not have a preference for any of the OCF types (cf. Blex, 2001, pp. 117 - 134). The latter mention that their way of correction is highly influenced by the specific student who needs the correction, by the context of the lesson as well as by the frequency of the occurrence of a specific error. With one exception, all of the teachers' self-assessments correspond with their actual behaviour in class (cf. ibid.).

Further studies conducted by Saeb (2017), Khatik and Vaezi (2018) and Argüelles et al. (2019), which also investigated teachers' preferences based on their attitudes, show that all of their interviewed teachers prefer implicit over explicit OCF strategies. These teachers mention that their preferences are highly influenced by concerns about learners' feelings and general advantages of implicit OCF strategies (cf. Saeb, 2017, p. 40; Khatik & Vaezi, 2018, p. 485; Argüelles et al., 2019, p. 112). Further, teachers in Argüelles et al.'s study argue that they tend to use the strategies they read about during their studies or that they experienced as learners during their own school time (cf. Argüelles et al., 2019, p. 112).

Even though the latter mentioned studies seem to agree on teachers' preferences being in favour of implicit OCF, the first study clearly points out that preferences vary widely. Based on the mentioned model of Baumert et al. (2011, p. 23), no other statement can be made than that teachers' preferences might all be different, based on several personal as well as educational reasons.

2.4 Research Question and Hypotheses

Based on the theoretical background, five sub-questions are formulated concerning the guiding research question:

- In how far is oral corrective feedback given in forms- and message-focused phases in the EFL classroom?

1. To what extent does oral corrective feedback take place in forms- and message-focused phases?
2. What types of OCF are used in forms- and message-focused phases?
3. To what extent does OCF take place in classes of younger EFL learners and in classes of older EFL learners in forms- and message-focused phases?
4. What types of OCF are used concerning students of different ages in forms- and message- focused phases?
5. To what extent do teacher preferences concerning different OCF types exist in forms- and message-focused phases?

Against the theoretical backdrop described in chap. 2.3, ten hypotheses are formulated concerning these sub-questions. By either confirming or rejecting these hypotheses, it is possible to develop an answer to the main question. H1 to H4 refer to the first sub­question, while H5 and H6 can be seen in reference to the second sub-question. H7 refers to the third sub-question and H8 as well as H9 to the fourth sub-question. As the research on appropriate OCF strategies for younger learners has still not been sufficiently researched, H8 is based on Lyster and Saito’s (2010) study and its results as it is based on a larger sample size than the one of Mackey and Oliver (2002) (see app. 4). H10 refers to the fifth sub-question.

(H1) In phases that focus on forms, oral corrective feedback is generally more often given than in phases that focus on message.

(H2) In phases that focus on forms, all error types are always corrected.

(H3) In phases that focus on message, grammatical errors are less often corrected than lexical ones and errors in pronunciation.

(H4) The unsolicited use of L1 is always corrected in message-focused phases.

(H5) In phases that focus on forms, explicit OCF strategies are more often used than implicit OCF strategies and delayed feedback in sum.

(H6) In phases that focus on message, recasts, prompts, no feedback and delayed feedback in sum take more often place than explicit correction.

(H7) In classes of younger EFL learners, oral corrective feedback is generally more often given than in classes of older EFL learners.

(H8) In younger EFL classes, explicit correction and prompts are more often used than recasts. (H9) In older EFL classes, students correct themselves more often than in younger EFL learners.

(H10) Teacher preferences exist in the context of OCF in forms- and message-focused phases.

3 Methodology

Within the framework of this work, an observational sheet is chosen as research instrument for the collection of quantitative data. Detailed information on the methodology of this study is given in the following sub-chapters.

3.1 Informants of the Study and the Data Collection Procedure

The informants of this study are three classes of year five, three classes of year seven and three classes of year ten as well as their teachers of a German secondary school in Bonn.

Detailed information on these classes is presented in tab. 3.

Table 3 Information on the Observed Classes of this Study

Abbildung in dieser Leseprobe nicht enthalten

The teacher of class 5r also teaches English in class Eph a, just as the teacher of 5t also teaches in class 7y. All of the eight teachers are female. Their mother tongue is German, just as it is the case for all of the students except three exchange students from China and one exchange student from France.

As this master thesis builds on the results of the empirical project within the practical semester, the first timespan of observation concerning classes 7x, 7y and 7z, during the mentioned semester, took place from 26 November 2018 to 14 January 2019. A pre-test was arranged in each of the three classes on 12 November 2018 and 13 November 2018 (see electronic app. 1). Resulting modifications following these pre-tests are depicted in app. 6. The follow-up observations for this master thesis started on 28 March 2019 and finished on 31 May 2019. Each of the classes was observed over ten school lessons of 45 minutes each. As some classes were parallelly scheduled, not all of the observed lessons took place one after another. In each case the teachers were asked for permission for being observed concerning the topic of OCF beforehand.

3.2 The Employed Research Instrument

As the aim of this master thesis is to get further insights into how OCF is given in the EFL classroom of a German secondary school in Bonn, the research method of a standardised, open, not-participating observation in form of a field observation was chosen since it is focused on describing events and how they are manifested without the intervention of the researcher (cf. Roos & Leutwyler, 2017, pp. 213-214). Therefore, the same observation sheet which was created for the empirical project as part of the practical semester was used. This sheet consists of a table and boxes where information on basic conditions such as the class, the date of the observation and the lesson, as well as on further notes regarding the situation of the erroneous utterance can be filled in. The columns of the table comprise five variables: the student ID, the sentence and its correction, the error type, the feedback type and the focus of the phase of the lesson (see app. 7). As a top-down approach was used, all of the feedback and error types were determined before analysing the corpus. Hence, for the last three variables, code plans are included within the first row. Error types are categorised as presented in chap. 2.1. One modification was made in terms of adding the error type “unsolicited use of the first language” as the frequency of this error increased during the follow-up study (see app. 7). The codification of the feedback types is deduced from fig. 1 (see chap. 2.2) and presented in fig. 2. The codification of the focus of the phase of a lesson includes the distinction between the two foci mentioned in chap. 2.3.1.

Abbildung in dieser Leseprobe nicht enthalten

Figure 2 Codification of Feedback Types

3.3 Data Handling

First, the collected raw data (see electronic app. 2) which arose out of the observation sheets was organised in the statistics program SPSS. In the ‘variable view', eleven variables are defined,⁵ as described by Meier Kruker and Rauh (2005, p. 113). These variables display the number of the situation, the above-mentioned basic conditions as well as the five variables shown in each column of the table and the variable ‘ second feedback type'. Concerning further data preparation, the codification printed on the observation sheet was used in terms of creating a code plan concerning the variables ‘error type', ‘feedback type', ‘second feedback type' as well as ‘focus'. The whole data set containing the code plan within the variable view as well as the data view is attached to electronic app. 3. Counted frequencies and created cross tabulations are attached to electronic app. 4. The data set was also exported into Excel to set up different spreadsheets for further calculations (see. electronic app. 5).

3.4 Methodological Strengths and Limitations

The strength of the chosen research method lies in the fact that the collected data cannot be deliberately altered by the examined person as might be possible within oral or written interviews and questionnaires. In contrast to that, the chosen field observation provides authentic data concerning the occurrence of different error types and their handling by teachers in two different phases of a lesson. A further distinction is that aspects of which the examined persons might not be aware, can also be explored (cf. Roos & Leutwyler, 2017, p. 211). Nevertheless, as the investigated teachers were informed about the study beforehand, the observer's paradox⁶ cannot be overcome. Only the fact that each teacher was observed over ten lessons might lead to a more normal behaviour as the investigated persons get used to being observed. Further, as Schnell et al. (2005, p. 394) state, the human's capacity of perception is limited. In regard to this research, it might be possible that some errors, especially those that were not corrected by the teacher, were also missed by the researcher (cf. Pawlak, 2014, p. 114). This might lead to restraints in terms of the investigation's procedural objectivity. In contradistinction to this, there are less restrictions concerning the objective scoring and the objective interpretation of the collected raw data. This is due to the fact that all steps are based on the already described theoretical background and are kept transparent within this chapter. In terms of reliability, it is to be noted that these observations can neither be repeated nor parallelly tested to achieve the same results, as Roos and Leutwyler (2017, p. 251) claim, as the occurrence of the same oral errors cannot be planned. As the objectivity of the interpretation is given, it can be argued that the interrater reliability is given, so that this study seems to be reliable to a large extent. The validity of this study is also given with smaller restrictions regarding construct and content validity. The focus of this study only lies on four error types and the most relative feedback types due to length limitations of this master thesis. Error types such as a lack within the social and intercultural competence are not observed, on account of which it is arguable whether the construct validity is given. Moreover, as the observer is not familiar with personal reasons of the teacher for using specific feedback types for one student or for the class as a whole, social competence might be tested as well. This leads to restraints concerning the content validity. Again, the fact that this master thesis is restricted in length leads to the problem that further interviews that would be needed to get an insight into, for example, teacher-student-relationships or teacher preferences could not be realised (cf. Roos & Leutwyler, 2017, p. 211).

4 Results

In general, 592 errors can be counted of which 511 errors (86 %) were corrected. Among these errors, 29 % are grammatical errors, 20 % are lexical errors, 43 % are errors in pronunciation and eight % belong to the category of the unsolicited use of the L1 (see fig. 3). Fig. 4 shows that these errors were corrected by using explicit correction in 50 % and by using recasts in nine %. Delayed feedback and peer feedback were used in seven % and no feedback was given in 14 %. Further types were used in zero point five to four %.

Abbildung in dieser Leseprobe nicht enthalten

Figure 3 Distribution of All Errors Figure 4 Distribution of all Feedback Types (n = 592)

Types (n = 592)⁷

The results of this investigation are divided into five sub-chapters which correspond to the different researcher questions that were formulated in chapter 2.4.

4.1 The Extent of OCF Concerning Different Error Types in Forms- and Message-Focused Phases of a Lesson

In forms-focused phases, 347 errors were collected, while 241 errors were counted in message-focused phases. 98 % of the former and 70 % of the latter were corrected (see. fig. 5)⁸.

Generally, 57 % of all errors in forms-focused phases are errors in pronunciation, while this applies to 25 % in message-focused phases (see fig. 6 & 7). Further, in phases that focus on forms, twelve % of all errors are lexical errors, 28 % are grammatical errors and three % fall into the category of the unsolicited use of the first language. In contrast, in message-focused phases, 32 % of all errors are lexical errors, 30 % are grammatical errors and in 13 %, the first language was used without being solicited. Fig. 7 shows that concerning each error type occurring in phases focusing on message, except the unsolicited use of the L1, at least 31 % of the errors were left uncorrected. With reference to phases that focus on forms, no more than nine % of all lexical errors and no more than 2 % of the other errors in pronunciation remained uncorrected, while the unsolicited use of the L1 was corrected in each case (see fig. 6).

Abbildung in dieser Leseprobe nicht enthalten

4.2 The Usage of Different Feedback Types in Forms- and Message-Focused Phases of a Lesson

In forms-focused phases, the majority of OCF moves were identified as explicit correction (64 %). Eight % of these 64 % were complemented by metalinguistic feedback (see electronic app. 4). Metalinguistic feedback alone was given in five % and peer feedback, initiated by the teacher or the students themselves, made up nine %. Elicitation and sounds were each used in four % in forms-focused phases and all other feedback types were used with a proportion of one to three % (see fig. 8). Regarding the categories of feedback strategies, explicit OCF strategies made up 89 % and implicit OCF strategies made up eleven % in forms-focused phases (see app. 8 & 9). Further, reformulations were used in 82 % and prompts in 18 % in these phases (see app. 10 & 11).

Abbildung in dieser Leseprobe nicht enthalten

The second chart in fig. 8 shows that explicit correction was used in 30 % in message-focused phases. Recasts were given in 15 % and delayed correction in 14 %. As has already been mentioned in chap. 4.1, no feedback was given in 30 %. Further feedback types, except sounds, were used with a proportion of one to three %. In sum, explicit OCF strategies made up 65 % of all feedback strategies in these phases and implicit OCF strategies made up 35 % (see app.8 & 9). Reformulations were used in 93 % and prompts in 7 % (see app. 10 & 11). The distribution of all feedback types regarding the three different error types in forms-focused phases in contrast to message-focused phases are displayed in app. 12. and 23 % of ones in the latter mentioned phases were left uncorrected. In grade seven, 115 errors took place in forms-focused phases of which three % were left uncorrected and 19 % of 76 errors in message-focused phases remained uncorrected. In forms-focused phases in grade ten, 65 errors were collected, while 118 errors were counted in message- focused phases. One % of the errors that occurred in the former mentioned phases and 41 % of the errors in the latter mentioned phases received no feedback (see fig. 9).

Abbildung in dieser Leseprobe nicht enthalten

Figure 9 Correction of Errors Concerning the Grade and the Phase of the Lesson

4.4 The Usage of Different Feedback Types Concerning Different Ages of the

Students in Forms- and Message-Focused Phases of a Lesson In forms-focused phases, explicit correction was mostly used in all three grades. In grade five, it was used in 53 % of all cases, while peer feedback took place in 15 % and all other feedback types were used in two to five %, as fig. 10 shows. In grade seven, explicit correction was used in 67 %, recasts in eight %, elicitation and peer feedback in six % and all other feedback types in one to three %, except non-verbal ways and sounds which were not used. In grade ten, explicit correction was used in six-seventh (85 %) of all used feedback types. In eight %, the students of year ten corrected themselves after they had uttered an error and in one %, no feedback took place or a clarification request was used. All other feedback types were not used for this grade in forms-focused phases (see fig. 10). Regarding the feedback strategies only, explicit OCF made up 90 % in grade five, 86 % in grade seven and 98 % in grade ten, while implicit OCF strategies made up ten % in grade five, 14 % in grade seven and two % in grade ten in forms-focused phases (see app. 14). Reformulations made up 73 % in grade five, 84 % in grade seven and 93 % in grade ten, while prompts took place in 27 % in grade five, in 16 % in grade seven and in seven % in grade ten in these phases (see app. 15).

[...]

---

¹ In this thesis, the terms ‘error' and ‘mistake' are treated as synonyms. Even though the distinction between these terms might be worthy from a theoretical point of view, it is difficult to decide within an observation as a researcher, and sometimes even as a teacher, whether a student's inaccurate spoken output is based on a lack of competence or performance if one refers to Timm's (2009, p. 205) definition (see chap. 2.1).

² The understanding of the presented OCF strategies corresponds to the ones used in the empirical project on which this study is based, even though information is added, or definitions are rewritten.

³ Following definitions of the two different phases of a lesson correspond to those given in the empirical project on which this study is based, even though information is updated or added.

⁴ cf. Timm's (2009, p. 205) definition (see chap. 2.1)

⁵ The data handling as well as the methodological strengths and limitations conform, except for a few additions made in chap. 3.4, to those of the empirical project on which this study is based.

⁶ The ‘observer's paradox', a term coined by William Labov (1972, p. 209), deals with the paradox aim of an observational investigation to find out how people use language when they are not being observed.

⁷ All figures in chap. 4 are based on the data of the electronic appendices 4 and 5.

⁸ In situation 347 to 350, the teacher asked the students to focus on both the forms and the message of a student’s text that was read out loud, wherefore no clear classification concerning the phase of the lesson can be made for these four errors and they are excluded from further examinations.