The Effects of Direct and Indirect Written Corrective Feedback on EFL Students' Accuracy


Research Paper (postgraduate), 2018

60 Pages


Excerpt


Table of Contents

Dedication

Acknowledgements

Abstract

List of Tables

List of Figures

List of Abbreviations

General Introduction

Chapter One: The Review of Literature
Introduction
1. Definitions of Key Terms
2. Effectiveness of Written Corrective Feedback (WCF)
2.1 Early Studies on the Effectiveness of Written CF
2.2 Early Studies Design Flaws
2.3 Recent Studies on Effectiveness of Written CF
3. Effectiveness of Different Approaches and Types of Written CF
3.1 Approaches of Providing Written CF
3.2 Effectiveness of Different Types of Written CF
4. Types of Errors to Be Targeted with Written CF
5. Appropriate Time to Provide Written CF
Conclusion

Chapter Two: Methodology
Introduction
1. Research Approach
2. Instruments
2.1 Controlled Practice Exercises
2.2 Interviews
3. Participants
4. Target Structure
5. Data Collection Procedure
6. Data Analysis Procedure
7. Research Questions and Hypotheses
Conclusion

Chapter Three: Results and Analysis
Introduction
1. Overall Effectiveness of Written CF
2. Effectiveness of the Two Written CF Types
3. Results of the Interview

Chapter Four: Discussion and Recommendations
Introduction
1. Discussion of the Results
1.1 Effectiveness of written CF
1.2 Effectiveness of the Two CF Types
2. Recommendations
Conclusion

General Conclusion

References

Appendices
1. Appendix A: A Sample of Controlled Practice Exercise for the Pre-test
2. Appendix B: A Sample of controlled Practice Exercise For the Post-test
3. Appendix C: A Sample of Direct CF with Written Meta-Linguistic Explanation
4. Appendix D: A Sample of Indirect Coded CF
5. Appendix E: The Interview

Dedication

I dedicate this work to my beloved Parents and my grandmother. Indeed, without their unconditioned emotional and financial support, this work would not come to existence.

I also dedicate this work to my little brothers, Hassan Ezzahouani and Mohammad Ezzahouani wishing them a good luck in their lives.

Acknowledgements

All praise and thanks to Allah who granted me with will, strength, and means to complete this monograph. Without his grace and mercy, this work would never come to be accomplished.

I would like to express my heartfelt gratitude and sincere thanks to my supervisor Dr. Lalla Meriem OUAHIDI for her continuous support and guidance throughout the process of conducting this research paper. Indeed, without her instructive feedback, this work would not come to completion.

Special thanks to my friends and classmates: Rachid BAHHAMI, El-mehdi EL-HAKIMY, Ikram ANZAR and Merieme SIDKI for their valuable and concrete help during the preparation of this research paper.

My deepest appreciation is extended to S4 students Group 1, the participants of the the study in this paper, for their cooperation and patience.

I owe my warm and sincere thanks to my parents for their unconditioned love, support, affection and advice they have provided me with during all the phases of my life.

Last, but not least, I wish to thank everyone who contributed directly or indirectly to the fulfillment of this work.

Abstract

The study in this research paper longitudinally examines the effectiveness of direct written Corrective Feedback (CF), with written meta-linguistic explanation, and indirect coded written CF in helping EFL students to improve their writing accuracy regarding the use of past simple and present perfect tenses. 38 intermediate EFL students at Sultan Moulay Slimane University in Beni Mellal, Morocco, participated in the study. They underwent a pre-test and an immediate post-test (the treatment) over a period of two weeks. During this period, the participants were divided into three groups; the first group received direct CF, with written meta-linguistic explanation; the second group received indirect coded CF while the third group was a control group; therefore, it did not receive any feedback. The first question of the study found that written CF was effective since the treatment groups outperformed the control group in the post-test with a mean difference of (MD= 2,01%). The second question of the study, in turn, revealed that indirect coded CF was more effective than direct CF accompanied with written meta-linguistic explanation. The indirect CF group exceeded the one received direct CF in the post test with a mean difference of (MD= 3,9%). Due to time constraints, it was not possible to administer a delayed post-test. Therefore, an interview was held with four participants from the experimental groups to support the findings of the second research question. The interview revealed that indirect coded CF provokes students' problem-solving skills. Therefore, the participants believe that this type of feedback is likely to be retained for future situations. Based on these findings, some recommendations were suggested at the end of the fourth chapter.

List of Tables

Table 1: Studies on the Effectiveness of Written CF

Table 2: Descriptive Statistics of the Experimental Groups and the Control Groups in the Pre-test and the Post-test

Table 3: One Way Anova analysis for the Pre-test and the Post-test

Table 4: Descriptive Statistics of the DCF Group and the IDCF Group in the Pre-test and the Post-test

Table 5: Independent Sample T-test for the Pre-test and the Post-test

List of Figures

Figure 1: Student 3 Responding to the First Question of the Interview

Figure 2: Student 2 Responding to the Second Question of the Interview

Figure 3: Student 3 Responding to the Second Question of the Interview

List of Abbreviations

Abbildung in dieser Leseprobe nicht enthalten

General Introduction

Improving students L2 writing accuracy has long been the primary goal of second language instruction since the 19th century with the emergence of the Grammar Translation Method (Freeman, D. Anderson, M., 2000). The absolute idea, then, was that teachers can improve their students’ accuracy by correcting their mistakes by and large. In fact, Bitchener, J. Ferris, D. (2012) noted that for composition teachers, then, correcting students’ papers was not only a part of their job, but it was assumed to be their primary job (P. 29). However, not until Truscott, J. (1996) published his article entitled "The Case Against Grammar Correction", did this idea become doubtful. Truscott (1996) and many other early researchers like Semke (1984) and Kepener (1991) argued that error correction or corrective feedback should be abandoned for its negative if not harmful effects on students' writing. This argument instigated many researchers later on (ferris 1999, 2004, Bitchener. et al. 2005, 2015, Goksoy Nazli 2016, Sia cheung 2017) to investigate the effectiveness of written CF. So far, recent research has refuted the claim of Truscott and proved written CF to be effective. Yet, whether teacher should provide this feedback directly or indirectly is still an open area in the field. In fact, Bitchener (2007) reported that up to 2007 only five studies have been published comparing the effectiveness of direct and indirect CF.

The present study, therefore, comes as a contribution to the body of research which attempts to fill in this gap, trying as much as possible to overcome the design shortages of previous studies . Subsequently, this study aims at (1) investigating the overall effectiveness of written CF by comparing the performance of the control group and treatment groups in the pre-test as well as in the post-test, and (2) investigating the effectiveness of direct CF accompanied with written meta-linguistic explanation and indirect coded CF by comparing the performance of the two experimental groups in the aforementioned tests.

To meet the aims mentioned above, this research paper is organised in terms of three chapters. Whereas the first chapter reports the findings from studies in the field, the second chapter represents the methodological design of the study. The third chapter, in turn, reports the findings of the study in relation to the research questions and, finally, the fourth chapter discusses these findings with previous literature in the field suggesting some recommendations for language writing instructors.

Chapter One: The Review of Literature

Introduction

The field of written Corrective Feedback (CF) has been a fertile ground for a considerable amount of research for more than two decades. Hendrickson. J. (1978) proposed five main questions that any study in this field should aim to answer; these questions are (1) should learners’ errors be corrected? If so, (2) which errors should be corrected? (3) How should they be corrected? (4) When should they be corrected? And finally (5) who should correct them? (Hendrickson 1978, P. 389). Within this framework, many studies have been conducted addressing some or all of these questions and, yet, introducing controversial findings. In this review, we will focus on the findings of some of these studies in terms of whether written CF should be provided or not? If so, are some types of written CF more effective than others? Which errors should be targeted? And finally what is the appropriate occasion to provide learners with written CF?

1. Definitions of Key Terms

The term “feedback” generally refers to any information or any sign that is given to learners on their performance of a learning task aiming at improving their performance (Ur. P 1996, P. 242). This term is used in Second Language (SL) literature to mean both positive and negative feedback. Positive feedback (also known as positive evidence) affirms that learners’ output is correct and possible in the target language. Negative feedback (referred to as corrective feedback), on the other hand, indicates that learners’ output is, somewhat, linguistically deviant and not possible in the target language (Ellis. R. 2009 and Nassaji. H. Fotos. S. 2011). In other words, corrective feedback refers to the feedback that SL learners receive on the linguistic errors they make in their oral and written output of the target language (Ellis. R Sheen. Y 2006) (as cited in Hinkel. E 2011, P. 593). In this paper, the term “corrective feedback” will be used to mean the teachers’ response to students’ erroneous written production.

The term “accuracy” refers to the person’s ability to use correct forms without errors that may affect the phonological, syntactic and discourse features of a language (Bryne 1988) (as cited in Sheen. Y 2013 P. 819). In other words, accuracy refers to the degree of deviancy from the norms of the target language; this degree is measured by the number of errors that occur in the learner’s spoken or written output (Housen. A. Kuiken. F. 2009 and Bitchener. J. Ferris. D. 2012). Accuracy is contrasted with fluency which means students’ ability to produce organised and eloquent production that will help them to maintain communication in both speech and writing of the target language. (Housen. A. Kuiken. F. 2009 and Sheen. Y. 2013). The term accuracy, in this paper, is used to mean students’ ability to avoid grammatical errors in their pieces of writing of the target language.

2. Effectiveness of Written Corrective Feedback (WCF)

The question of whether or not written CF helps students improve their writing accuracy is of paramount importance in the field of Language Teaching in general because, as Bitchener. J. Ferris. D. (2012) stated, if the answer is negative, there will be no need for any further research in this field. Many Language Teaching Methods have provided different and, yet, controversial answers. The pioneers of the Audio-lingual Method, for example, argue that positive feedback is desirable in the process of learning since it works as a reinforcement for the learner while negative or corrective feedback is useless and should be avoided since it works as a “punishment” and is likely to regress the learning process (Ur. P. 1996, P. 244). This method recommends that teachers should teach the correct forms from the first time leaving no room for students to make errors (Freeman, et al. 2000, P. 47). In contrast to the Audio-lingual approach, the communicative language leaching approach points out that error correction should be ignored during fluency-based activities (i.e.CF is unlikely to be effective when the focus is on students’ fluency); the pioneers of this method state that the teacher can note the errors and return back to provide CF on them during accuracy-based activities (Freeman, et al. 2000, P. 132). In support of this approach is the Inter-Language Hypothesis which states that learners’ errors form an important part of the learning process, and providing feedback on them is a way of bringing the learners’ performance closer to the Target Language (Slinker 1772, 1992) (as cited in Ur. P. ibid).

To test the claims of these approaches on empirical grounds, many studies have been conducted in real classroom settings to investigate the effectiveness of written CF as shown in Table 1 below:

Table1: Studies on the Effectiveness of Written CF

Abbildung in dieser Leseprobe nicht enthalten

In order to be able to criticise and provide an overview about each of the above studies, it is worthwhile to divide them into early studies (mid 1980s to 2003) and recent studies (from 2005 onwards) as proposed by Storch, N. (2010).

2.1 Early Studies on the Effectiveness of Written CF

The belief that teachers should correct their students’ errors was mostly acceptable and not negotiated, until Truscott (1996) published his article, entitled “The Case Against Grammar Correction”, in which he argued theoretically and practically that written CF is ineffective and may even be harmful. From a theoretical point of view, he points out that language learning is a gradual and a “developmental” process; therefore, students produce correct utterances only when they reach the suitable cognitive stage no matter how much error correction they receive (P. 358). On practical grounds, Truscott (1996) stated that many studies conducted in this regard (e.g. Cohen and Convaleanti 1990) revealed that, in most cases, teachers fail to notice their students’ errors and even when they notice them they provide only a limited explanation compared to the actual use of the grammatical rules that govern those errors (P. 350). In addition, studies conducted by Brannon (1981), Cohen (1987), Maxely (1989) and Leki (1990) revealed that over-correction of students’ errors downgrades them and affects their achievement negatively (Truscott, 1996, P. 351).

Based on these arguments, Truscott (1996) concluded that grammar correction is harmful and should be abandoned. He points out that teachers in classrooms can do anything “except grammar correction” (P. 360)

However, Truscott (1996) was severely criticised by Ferris (1999) and considered to be premature, overly strong and even dangerous to the field because: (1) Truscott neglected the fact that there are different types of written CF. Therefore, the ineffectiveness of one type does not necessarily include the other types (Ferris, 1999, P. 4). (2) Truscott has cited only the studies that support his claim and forgot about the ones which contradict him. He even has either under- or over-stated the findings of the studies in his review to fit his arguments (Ferris, ibid, P. 5).

Many of other early studies (Semke 1984, Kepner 1991, Polio, Fleck Leder 1998) have agreed with Truscott’s claim that written CF should be desisted. Semke (1984) found that error correction is ineffective in terms of accuracy and may even be harmful in terms of fluency (as cited in Chandler, 2003, P. 268). In the same context, Kepner (1991) reported that there was no significant error reduction among students who received surface-level feedback and those who received content related feedback (as cited in Bitchener, 2008. P, 105). Similarly, Polio et al (1998) found that both the experimental and the control group improved in accuracy with no significant difference between them (as cited in Chandler, 2003. P. 269). However, the findings of these studies cannot be heavily relied upon, nor can they be taken as an answer to the question of written CF effectiveness since they suffer from many design flaws (Ferris 2004, Ferris Bitchener 2012, Bitchener 2008 and Storch, N. 2010).

2.2 EarlyStudies Design Flaws

In order for any study to be reliable, it must fulfil many design requirements, which were, according to Ferris (2004), Ferris Bitchener (2012), Bitchener (2008) and Storch (2010), prominently absent in many of the early studies. The findings of a certain study are reliable only if the study includes:

2.2.1 A Control Group

In order to make sure that the improvement in accuracy was the result of written CF alone, there must be a comparison between the experimental group (i.e. the group of learners that received written CF) and a control group (i.e. a group of learners that did not receive any treatment) (Ferris Bitchener, 2012, P. 51). Many of the early studies that were mentioned earlier (e.g. Kepner 1991 and Chandler 2003) did not include a control group. Therefore, it is difficult to say whether the output was the result of the feedback alone or of other classroom instructions (Ferris Bitchener, 2012. P 68). Ferris (1999, 2004) and Truscott (1996, 2004) also agreed that studies which did not include a control group, hence, failed to compare the effects of corrective feedback and non-corrective feedback cannot be taken as an evidence of the effectiveness of corrective feedback (as cited in Storch, 2010, P. 32).

2.2.2 A New Piece of Writing

After measuring students’ current level of accuracy by means of a pre-test of writing, it is necessary to administer an immediate post-test on new pieces of writing, following the provision of written CF so as to avoid the intervention of other variables between the treatment and the new text (Bitchener, 2008 and Ferris Bitchener, 2012). Bitchener (2008) insisted that “both the pre-test and the post-test should be included in the design to allow comparability between them” (P. 108).

Many of the early studies assessed students’ accuracy by considering only revised texts. However, text revision, according to Guenett (2007) and Truscott (1999, 2004, 2007), does not provide an evidence that written CF has long-term effects beyond the revision stage (as cited in Storch, 2010, P. 32); in other words, when students are asked only to revise their previous texts, it becomes difficult to know whether or not they can apply the knowledge they gained from the CF treatment in new contexts. Thus, the need for a post test on new texts is necessary.

2.2.3 Appropriate Writing Conditions

A number of early studies (Smeke 1984, kepenr 1991 and Polio et al 1998) provided feedback on students’ journals. However, according to Ferris (2003), journals are unlikely to motivate students’ attention to grammar accuracy (as cited in Storch, 2010, P. 33). In addition, even in studies that used more appropriate tasks than journals, students were asked to do their writing at home. This makes it difficult to know if the student wrote his essay by himself or he resorted to an additional assistance (Storch, 2010, P. 33).

2.3 Recent Studies on Effectiveness of Written CF

After Ferris, D. (2004) had investigated the design shortcomings of the existing research on written CF, she made an urgent call for more “longitudinal, carefully designed and replicable studies” (P, 60) to answer the question of written CF effectiveness. Henceforth, many recent studies (Bitchener, J., Young, S. Cameron, D 2005, Bitchener 2008, Bitchener, J. Knoch, U 2008, Bitchener, J. Rummel, S. 2015, Kang, E. Hang, Z. 2015, Wanger, J. 2016) have sought to overcome the early studies design flaws in an attempt to provide an adequate evidence for the effectiveness of written CF.

In their study, Bitchener et al. (2005) found that the experimental groups that received written CF generally outperformed the control group with a slight improvement of the experimental groups that received feedback on past tense and article usage over the group that received it on preposition usage. Similarly, studies by Bitchener (2008), Bitchener Knoch (2008) and Bitchener Rummel (2015) revealed that students who received written CF improved in their accuracy than the control group so far as the English article system is concerned. On the other side, Wanger, J. (2016) reported that grammatical complexity (i.e. the number of the correct usages a rule may have) determines the effectiveness of written CF. He states that “if the binary construction rules idiosyncratically, written CF can still be effective. If the construction has more than a binary option for correct usage, written CF will not help in the acquisition process” (P. 173). However, in a meta-analysis study conducted by Kang, E. Hang, Z. (2015), they concluded that many variables including learner’s proficiency level, the setting and the genre of the writing task can affect the efficiency of written CF from moderate to large effects.

So far, the debate about the effectiveness of written CF seems to be a vehement and an on-going one. However, the claim supporting written CF effectiveness seems to hold sway since it is based on empirical studies with a very few and insignificant design defects. This has shifted the interests of many recent studies to inquire into whether some types and approaches of written or CF are more effective than others. This will be our concern in the following section.

3. Effectiveness of Different Approaches and Types of Written CF

3.1 Approaches of Providing Written CF

It is an unquestionable fact that achieving a long-term effectiveness is the aim of any feedback provider. Yet, the approach to be followed in achieving this goal is still in question. Studies conducted on whether teachers should comprehensively correct all the students’ errors (unfocused approach) or focus on a selected type of errors (focused approach) have shown varying results.

3.1.1 Unfocused Approach

A wide number of early studies (Semke 1984, Robb et al 1986, Kepner 1991, Chandler 2003) that provided students with a comprehensive feedback on a wide range of error categories reported that there was no difference between the “control group” and the group that received written CF (as cited in Ferris Bitchener 2012, P. 54). However, as mentioned earlier and as Ferris Bitchener (2012) argued, due to many design defects, the findings of these studies cannot be reliable; in other words, no one of these studies contained a real control group. The groups that they claimed to be a control group also received a content-related feedback which may also refer “to issues of linguistic errors and accuracy” (ferris Bitchener, 2012, P. 56). Indeed, the absence of control group in these studies makes it difficult to know whether the improvements were the results of the treatment alone or there is the integration of other classroom instructions (Bitchener, 2008). Another reason why these studies found written CF to be useless is that the feedback was unfocused (i.e. was given on a very wide number of linguistic forms). This, according to Ferris Bitchener (2012), gives a rise to the issue of whether those relatively low proficiency learners can absorb that “heavy” cognitive load and, therefore, benefit from the feedback provided. In fact, the learner “is unlikely to be able to reflect too much on each error” (Ellis. R. 2008. P, 102). Such debatable findings led many researchers to conduct further research with a focused approach to account for the effectiveness of written CF.

3.1.2. Focused Approach

Considering the shortcomings of the unfocused approach mentioned above, many recent studies ( Bitchnere 2008, Bitchener Knoch 2008, Bitchener Rummel. S. 2015, Sheen. Y. 2007 and Ellis 2008) have tried to provide learners with a selective feedback (i.e. to focus on a very limited number of error categories) and each study found that the treatment group outperformed the control group. Bitchener (2008), Bitchenr et al. (2008), Sheen (2007) and Ellis (2008) reported that the experimental group has indeed improved in accuracy than the control group so far as the use of English Articles is concerned. Similarly, Bitchener Rummel (2015) found that written CF helped students to improve their accuracy concerning the past simple tense and the use of Articles. It is noticeable that many studies have focused on the same error category- English Article System. Answering why all this focus lied on the same category, Ferris Bitchener (2012) put “in order to test the theoretical claims made against written CF, it is important that the claims be examined when the feedback is given in the same linguistic domain rather than a range of domains that may introduce a variable effect” (P.58).

Out of this controversial debate, taking into account the shortcomings of the unfocused approach mentioned above, it can be concluded that focused written CF may be proven more effective, at least for the time being, since the learner’s attention is on a small and ,yet, an intensive amount of knowledge. In this regard, Ellis (2008) states that “the more the attention, the more likely the correction is to lead to learning.” (P. 102). However, as many researchers (Bitchener 2008, Ellis 2008, Ferris Bitchener 2012) insisted, this area is still in dire need for further research.

3.2 Effectiveness of Different Types of Written CF

Although it can be seen from the previous body of research that there is a strong support for the effectiveness of written CF, the question of whether or not some types of written CF are more effective than others still insist in the literature of written CF. Studies that attempted to answer this question have examined and overly discussed two major types of written CF: Direct and indirect written corrective feedback and each of these two types encompasses several options.

Direct corrective feedback has been commonly defined as the explicit provision of the correct linguistic form or structure above or near the error (bitchener 2008, Bitchener Knoch 2008, Bitchener Rummel 2015, Ferris Bitchener 2012, Ellis 2007, Storch 2010, Kang, E. Hang, Z. 2015).

It can be carried out in different ways including the crossing out of the unnecessary words or morphemes, the insertion of a missing word or morpheme, and/ or the provision of the correct form or structure (Bitchener 2008), as shown in the example below:

Abbildung in dieser Leseprobe nicht enthalten

in the river.” (cited in Ellis, 2010, P. 99).

Bitchene (2008) proposed that additional forms of direct written CF may also include written meta-linguistic explanation (i.e. the provision of the grammatical rules and examples that govern the targeted error at the end of the script with reference to the concerned error) and/ or oral meta- linguistic explanation in the form of a mini-lesson to present and discuss the rules and examples of that govern the error indicated (P. 105).

Direct written CF is said to be helpful for learners because (1) it is immediate and reduces the confusion they may encounter when they fail to understand or remember certain error code (Ferris Bitchener, 2012, P. 65). Moreover, (2) it offers explicit guidance on how to correct errors, thus, it is obviously preferable for learners with low proficiency levels who do not know what the correct form is (Ellis, 2007, P. 99. However, Ellis, R. (ibid) argued that direct written CF requires minimal effort in the part of the learner, thus, it is unlikely to contribute to long- term learning.

Indirect written CF, on the other hand, refers (Bitchener, Young Cameron 2005 Bitchener 2008, Ferris Bitchener 2012, Ellis 2009b, Storch, 2010) to the cases where the teacher indicates that an error has occurred without providing any correction. Bitchener (2008) points out that indirect CF can be provided in three ways: either coded, uncoded or marginal; Coded indirect CF refers to the instances where the location and the type of the error are shown by means of writing a symbol, predetermined in classroom, near or above the error (e.g. writing ‘PS’ to indicate wrong use of the Past Simple Tense); uncoded indirect CF, however, indicates only the location of the error by circling or underlining it without any code; in marginal indirect CF, even less explicit, the teacher only writes in the margin the number of errors occurred in a specific line (P. 105).

Indirect written CF has the advantage that it drives learners to engage in guided learning and problem solving and help them reflect about the linguistic forms (Ferris Bitchener, 2012, P. 65). Therefore, this type is said to be more likely to lead to long-term learning (Elliss, 2009b, P. 99). However, Ellis, R. Et al (2008) argued that indirect written CF can work only if the learners already know the structures and forms of the language so as to be capable of self-correcting them. Therefore, this type can only increase students’ control of the structures that have already been internalised and may not lead to the learning of new linguistic forms (as cited in Storch, 2010, P. 40).

Ferris Bitchener (2012) suggest that the effectiveness of both types may depend on (1) the context of providing written CF (i.e. language learning classes or compositions classes) and (2) the student’s proficiency level (P. 65).

Many studies (Ferris Robb 2001, Bitchener, Young Cameron 2005, Bitchener 2008, Bitchener knoch 2008, Bitchener Rummel 2015, Hosseiny, M. 2014, Goksay, A Nazli, O. 2016) were carried out to investigate the merits of each of these types and they exhibited mixed results.

Bitchener et al., (2005) investigated the effectiveness of direct written CF together with oral meta- linguistic explanation and direct CF alone. Their study revealed that the group that received direct written CF with oral meta-linguistic explanation outperformed both the group that received only direct CF and the control group so far as the use of the past tense and definite articles is concerned. However, more recently Bitchener (2008) and Bitchener Knoch (2008) compared the effectiveness of direct CF accompanied with written and oral meta-linguistic explanation, targeting the same linguistic features (past tense and definite articles), and direct CF alone and found that both experimental groups outperformed the control group with no significant difference between the two treatments.

On the other hand, Ferris Robb (2001) inquired into the effectiveness of coded and uncoded indirect CF and reported that both groups exceeded the control group in accuracy but there was no difference between the two types of indirect CF.

A very limited number of studies have compared between direct and indirect written CF. Hosseiny (2014) and Bitchener Rummel (2015) compared the two types in addition to meta-linguistic feedback and pointed out that the experimental groups improved in accuracy with no difference between these types. On the contrary, Goksay Nazli (2016) revealed that indirect CF on areas of vocabulary and spelling is more effective than direct CF while the latter works best on other fields.

All in all, it is clear that no clear-cut answer can be obtained from such a limited and conflicting body of research. In fact, the need for further studies to compare both types of CF is an urgent one as Bitchener (2008) states “firm conclusions will only become available if further research, incorporating both types within the design of a single study, is carried out. (P. 107).

4. Types of Errors to Be Targeted with Written CF

Before teachers know which type of errors to target with written CF, they must first distinguish between what is an error (i.e. a gap in students’ competence) and what might be just a mistake (i.e. introduced by performance limitations e.g. memory, tiredness, boredom...). Ferris Bitchener (2012) argued that the distinction between errors and mistakes can be done in two ways: (1) if the deviant form occurs in several occasions in a single text, then it is likely to be an error. (2) if the learners can accurately use those deviant forms in new pieces of writing, after receiving indirect written CF on them, then they are likely to be just mistakes and do not need any feedback (P. 130).

A distinction has long been made in the literature of corrective feedback between “global” and “local” errors. This dichotomy of “global vs. local” errors was first introduced by Burk Kiparsky (1972) based on the extent to which an error impedes the comprehensibility of the message (as cited in Hendrickson, 1978, P. 391). Hendrickson (1978) defined global errors as those errors that interfere with the meaning of the message and make a proficient speaker of the language consider the message incomprehensible (P. 391). Local errors, on the other hand, refer to those errors that “make the structure of a sentence appear awkward but, nevertheless, cause a proficient speaker of the language no difficulty in understanding the intended meaning.” (Hendrickson, ibid). Global errors, according to Burk (1975), should receive top priority for correction. He states that correcting one global error in a sentence clarifies the intended meaning more than correcting several local errors in the same sentence and that limiting the correction on to global errors increases students’ motivation towards the target language ( as cited in Hendrickson, 1978, P. 391)

Two decades later, Ferris (1999) introduced another dichotomy of “treatable vs untreatable” errors suggesting that ‘treatable’ refer to those errors that occur in a rule-governed way and learners can solve them simply by resorting to a set of rules. Such errors include verb tense, subject-verb agreement, article usage, etc... ”. “Untreatable” errors, on the other hand, refer to those idiosyncratic deviations which are not governed by rules and, hence, require the use of the acquired knowledge of the target language to correct them. Examples of such errors include word choice errors, word missing or omitting, unidiomatic structures, etc...” (As cited in Bitchener et al. 2005, P. 194). Bitchnere et al (ibid) suggested that written CF works best with treatable errors than untreatable ones.

To conclude, Hendrickson (1978) states that there seems to be a consensus among researchers that errors which impair communication should receive an important amount of CF (P. 392). Yet, for this feedback to be effective, teachers should be aware of the appropriate time to provide it.

5. Appropriate Time to Provide Written CF

Finding the appropriate occasion to provide written CF may be easier for teachers in Language Learning Classes than those in composition classes. This follows that, teachers in composition classes argue that their focus goes beyond the provision of CF on the sentence level. They are, in fact, more concerned with developing students’ critical thinking skills, rhetorical awareness, and argumentation styles, etc Language Learning class teachers, however, concentrate more on students’ accuracy (Ferris Bitchener, 2012, P. 126).

Nevertheless, Ferris bitchener (ibid) point out that the appropriate occasion to provide written CF depends on several key factors. The first one is students’ proficiency level. Students with low proficiency require more CF on “ basic” errors while high-proficiency students need less amount of CF only on stylistic issues as these students have become proficient in the word- and sentence-level (Ferris bitchener, ibid). The second factor that determines when CF should be provided, according to Ferris Bitchener (2012), is teacher’s goals. When the teacher wants his or her students to make an accurate use of certain linguistic forms and structures, i.e. when the focus is on accuracy, it would be appropriate to provide written CF (P. 126). However, when teachers want their students to produce fluent pieces of writing for communicative purposes (i.e. when the focus is on fluency), It would not be appropriate to draw students’ attention to issues of accuracy as this may distract them. Generally, when the focus is on both fluency and accuracy, Ferris Bitchener, (2012) recommend that teachers should respond to fluency first and then to accuracy (P. 127).

Conclusion

To sum up, the sections of this review have revealed a number of issues concerning the value of written CF on students writing, the effectiveness of different types of written CF, the types of errors to be targeted with written CF and the appropriate occasion to target them. Likewise, this review has also emphasised the urgent need for further research to compare the effectiveness of both types of written CF, targeting a selected number of errors, on students’ new pieces of writing. To address this need, the present research paper is undertaken with EFL students at Sultan Moulay Slimane University to inquire into the extent to which direct and indirect written CF, on a selected number of linguistic forms, help students improve their accuracy of new pieces of writing.

Chapter Two: Methodology

Introduction

The present experimental study attempts to fill in the gap highlighted in the review of literature concerning the effectiveness of different types of written corrective feedback in targeting a focused number of linguistic structures. This chapter serves to describe the methodology used in this study to answer the research questions. First, a description of the research approach, instruments and participants is provided. Then the target linguistic structure is discussed. Finally, an overview is given about the data collection and data analysis procedures.

1. Research Approach

To meet the objectives of the study, the researcher opted for a mixed methods approach. This approach involves gathering both numeric as well as text information so that the researcher ends up with qualitative and quantitative data (J. W. Creswell, 2002, P. 21). This approach, hence, captures the best of both qualitative and quantitative approaches and allows the inquirer to have a deep understanding of the research problem by first surveying a large number of the population and then following up with few of them to know their opinion about the research topic (Creswell, ibid, P. 24). In the present study, the researcher collected quantitative data by administering a pre-test and post-test to see how the two types of feedback affect students writing accuracy and then ended up with interviewing four participants to collect qualitative data about their opinions concerning the treatment they were given.

2. Instruments

Unlike most studies on written CF that have used writing prompts as an instrument for data collection, this study used “Controlled Practice Exercises” in both the pre-test and the post-test ( see appendix A and B), in addition to personal structured interviews at the end of the study.

2.1Controlled Practice Exercises

To collect the data for this study, the researcher used “Controlled Practice Exercises” in both the pre-test and the post-test. This instrument was first used by Bitchener, J. (2008) as a secondary instrument after the post-test to test students’ ability to use the feedback in future situations. In the current study, however, the researcher used these exercises as the primary data collection instrument for two main reasons; (1) this type of exercises is more time-saving than writing prompts. (2) These exercises oblige students to stick to the use of the target structure as they are required to select the correct answer from the available choices, unlike essay writing where students can avoid the use of a certain structure.

2.2 Interviews

At the end of the immediate post-test, personal structured interviews were conducted with four participants from both feedback groups. A personal structured interview, as defined by Kothari C. R. (2004), involves a “face-to-face contact with the other person or the interviewee” using “a set of predetermined questions and highly standardised techniques of recording” (PP. 97-98). In the study at hand, the interviews were conducted with four students (2 males and 2 females) who were randomly picked up from both feedback groups. The participants were asked four predetermined questions (see appendix E) and their answers were audio-taped. The researcher conducted these interviews in order to have more insights on the problems faced by the participants while doing the correction based on the feedback they were given and also to see whether or not they believe this feedback will help them use the target structure in future situations.

3. Participants

The original sample for this study consisted of 40 (20 males and 20 females) EFL intermediate students at Sultan Moulay Slimane University, Beni Mella. The majority of the participants (78, 57%) aged between 18 and 23 years old. All the participants took English courses in high school for three years. At the time of the study, the participants were studying in the fourth semester of the university programme which consists of 6 semesters, 2 semesters per year. The participants were chosen at random and randomly assigned to one of the three groups: group one received Direct CF with written meta-linguistic explanation, group two received indirect coded CF, and group three, the control group, received no feedback. However, only 38 students attended the immediate post-test (the treatment); 14 students (8 males and 6 females) from the direct feedback group, 14 students (7 males and 7 females) from the indirect feedback group and 10 students (5 males and 5 females) from the control group. Only the data related to the students who completed the study was considered in the analysis of the results.

4. Target Structure

As stressed in the review of literature, and as highlighted by Bitchener, et al (2005), written CF is more likely to be effective when targeting a focused number of treatable errors than when targeting a comprehensive number of errors. This study, therefore, focuses on a selected number of treatable linguistic structures: (1) simple past tense when used to refer to something that happened and finished at a specific time in the past. (2) present perfect tense when used to refer to an action that happened in the past and continued to the present time, or when used with subordinating conjunctions such as “since, because, after…etc.”. These two structures were selected because verb tenses are highly considered in students answers in the final exams of many subjects including Composition. In addition, during the interviews the students clearly stated that they face problems differentiating between the use of the present perfect and simple past tenses.

5. Data Collection Procedure

The data was collected for this study through two main stages during a period of two weeks, beginning with the pre-test followed by the post-test or the treatment, in addition to four personal interviews with students at the end of the study.

On the 28th of May 2019, the researcher administered the pre-test with 40 EFL intermediate students at the end of the composition session. During the test the participants were asked to fill in the gaps with the appropriate verb form. They answered the exercise in the classroom under the supervision of their teacher. Students’ papers, then, were randomly assigned to one of three groups and were accordingly given different types of feedback. Group one received direct corrective feedback with written meta-linguistic explanation; this includes the provision of the correct form above or next the error in addition to the provision of the rule and examples that govern the error (see Appendix C). Group two received indirect coded CF; this includes the underlying of the error and providing error codes above or near the error (e.g., “SP” for simple past) (sees Appendix D). Group three was the control group, hence did not receive any feedback during the study.

One week after the pre-test, students went on a one-week holiday. Immediately after the holiday, the researcher administered an immediate post-test or the treatment. Only 38 students attended this test and they were asked to carefully read the feedback given to them before they answer the exercise they were given. This also was done in the classroom under the teacher’s supervision.

At the end of the post-test, four personal structured interviews were conducted with four participants (2 males and 2 females) from both experimental groups in order to have more insights on the problems they faced when doing correction based on the feedback they were given. The answers of the interviewees were audio-taped to be analysed later on.

6. Data Analysis Procedure

The experimental study at hand consists of an independent variable and a dependent one. The independent variable is the two types of corrective feedback provided by the researcher. The dependent variable is students writing accuracy. The relationship between these two variables is identified by calculating and comparing the mean number of errors of both the experimental groups and the control group in the pre-test and the post-test to account for the overall effectiveness of written CF and also to make sure no teaching instructions interfered during the period of the study. Since there are three groups, hence three independent variables, One Way Anova Test was conducted to account for the difference between the two experimental groups and the control group. A comparison was made after the post-test between group one, which received direct corrective feedback, and group two, which received indirect coded CF, to test and account for the effectiveness of both types. As there are only two groups in this comparison, an Independent Sample T-Test was carried out to account for the difference between them.

The data collected through the interviews was qualitatively analysed to have more insights on the problems students faced while doing the correction based on the feedback they were given and whether they believe that the feedback will help them use target structure in future situations.

7. Research Questions and Hypotheses

This study is set up to answer the following questions:

1) Does Written CF help students improve their writing accuracy regarding the use of simple past and present perfect tenses?
2) Which type of written CF is more effective than the other?

While answering these questions, the following null hypotheses will be tested:

1) Written CF helps students improve their writing accuracy concerning the use of the simple past and present perfect tenses.
2) Direct CF with meta-linguistic explanation is more effective than indirect coded CF.

Conclusion

This chapter has presented an overview about the participants and the instruments of the study. The rationale behind choosing the research approach and the target structure was explained and the data collection and data analysis procedures were also discussed. The subsequent chapters will deal with the analysis and the discussion of the results obtained from the collected data.

Chapter Three: Results and Analysis

Introduction

The present study investigates the effectiveness of direct corrective feedback accompanied with written meta-linguistic explanation and indirect coded corrective feedback. In order to meet this goal and in order to answer the research questions stated in the previous chapter, a pre-test and post-test were established within a period of two weeks. After the pre-test, the participants were divided into three groups. Group one received direct error correction with meta-linguistic explanation (the symbol 'DCF' is used to represent this group here), the second group received indirect coded feedback (the symbol 'IDCF' stands for this group) and the third group was a control group, hence, did not receive any feedback (this group is represented by the symbol 'NO CF'). At the end of the study, 76 papers were collected (38 in the pre-test and 38 in the post-test) and the number of errors made by each group was calculated. In addition, face-to-face interviews were held at the end of the study to have more insights on the problems faced by the students while doing the correction based on the feedback they received. This chapter serves to represent and to analyse the results of the data collected in an attempt to answer the research questions and to test the null hypotheses stated previously.

1. Overall Effectiveness of Written CF

To answer the first research question, a comparison was made between the feedback groups as one group and the control group based on the percentage of errors made by each group. The percentage was calculated by dividing the errors made by each participant on the total number of present perfect and simple past verbs in the exercise. For example, a script contained 7 wrong usages of the target structure out of 26 obligatory usages in the exercise; the percentage of errors in this case is 35 %. Table 2 below represents the initial descriptive statistics of the comparison:

Table2: Descriptive Statistics of the Experimental Groups and the Control Groups in the Pre-test and the Post-test

Abbildung in dieser Leseprobe nicht enthalten

The descriptive statistics in Table 2 above show that 13 students from the experimental groups (7 from the DCF group and 6 from the IDCF group) and 6 students from the control group marked some improvement in the post-test. The highest scores of error reduction in the experimental groups were (17,28%), (16,37%) and (13,19%) scored by students 8, 9 and 13 respectively, making an improvement from (40%) of errors in the pre-test to (31,81%) in the post-test for student 8, from (30%) to (13,63%) for student 9, and from (45%) to (31,81%) in the post-test for student 13. The lowest score for these feedback groups was (2,73%) of error reduction gained by student 24 who improved from (30%) of errors in the pre-test to (27,27%) in the post-test.

For the control group, Table 2 shows that the highest scores of error reduction for this group were (20%), (12,78%) and (11,28%) achieved respectively by students 32, 35 and 34. For student 32, he improved from a percentage of errors of (70%) in the pre-test to (50%) in the post-test. Student 35 moved from (35%) of errors in the pre-test to (22,22%) in the post-test. Student 34 made an improvement from (30%) to (18,18%) in the post-test. The lowest error reduction rate for this group is scored by student 36 who progressed from (25%) in the pre-test to (22,22%) in the post-test, making an error reduction of (2,78%).

This descriptive comparison revealed that the participants in the experimental groups slightly improved in accuracy in the post-test than those in the control group. This implies that 13 participants in the feedback groups have read the feedback given to them, understood it and were able to apply it on the target structures during the post-test. That is, they benefited from the feedback.

However, 15 participants in the feedback groups did not show any improvement in the post-test, which means that they did not benefit from the feedback given to them. This might be due to the absence of the target structures, especially present perfect, in their mother tongue, Arabic. Hence, they could not assimilate the feedback. These participants may belong to the IDCF group and were not able to understand the symbols given to them. In addition, the absence of improvement from these participants may also be attributed to their inability to assimilate the feedback they received in such a short period of time; two weeks. Thus, they still need more time and capacity to understand the rules underlying the simple past and present perfect in English.

In order to provide the mean percentage of errors and standard deviation for each group, and in order to see whether or not the difference between the three groups is statistically significant, a one way analysis of variance test "One Way ANOVA'' was conducted as shown in Table 3 below:

Table3: One Way Anova analysis for the Pre-test and the Post-test

Abbildung in dieser Leseprobe nicht enthalten

As Table 3 displays, the control group scored a mean percentage of errors (35%) with a standard deviation of (13, 54 %) in the pre-test higher than the experimental groups which scored a mean percentage of errors of (33,75%) with a standard deviation of (8,98%). The mean difference between the two groups at this stage is (1,75%). However, the P value (Sig. = 0,54) shows that the difference is not statistically significant since P is higher than the standard significance level which is (P= 0,05). This indicates that the proficiency level of the three groups at the time of the pre-test was almost the same. This is an important factor for the comparison in the post-test to be reliable because if the experimental groups outperform the control group in the post-test, it will be only due to the feedback provided.

During the post-test, however, the trend reversed and the experimental groups outperformed the control group. The experimental groups gained a mean percentage of errors of (34,73%) with a standard deviation of (10,44%) while the control group decreased to a mean error percentage of (32,72%) with a standard deviation of (10,45%). The mean difference between the two treatment groups and the control group in the post-test is (2,01%), however, the P value (Sig.= 0,54) still shows that the difference is statistically not significant.

Since the experimental groups gained some improvement in the post-test though it was not statistically significant, it can be assumed that the participants in these groups benefited from the feedback given to them. Therefore, the answer to the first research question is that written corrective feedback helps students improve their writing accuracy regarding the use of the target structure. Subsequently, the first research hypothesis is accepted. Yet, this does not show which of the feedback types is more effective. Thus, the need to answer the second research question becomes inevitable.

2. Effectiveness of the Two Written CF Types

To answer this question, a comparison was made between the DCF group and the IDCF group based on the error percentage made by each group in the pre-test and in the immediate post-test (the treatment). Table 4 shows the initial descriptive statistics for this comparison:

Table4: Descriptive Statistics of the DCF Group and the IDCF Group in the Pre-test and the Post-test

Abbildung in dieser Leseprobe nicht enthalten

As Table 4 shows, 7 students out of 14 in the DCF group and 6 students out of 14 in the IDCF group showed improvement in the post-test. For the DCF group, the highest score of improvement was (17,28%) of error reduction gained by student 8 who improved from (40%) of errors in the pre-test to only (22,72%) of errors in the post-test. The lowest score in this group was gained by student 13 who improved from (40%) in the pre-test to (36,64%) in the post-test, scoring an error reduction of (3,64%). On the other hand, the highest score of error reduction in the IDCF group was (8,64%) achieved by student 25 who improved from an error percentage of (45%) in the pre-test to (36,36%) in the post-test. Student 24 gained the lowest rate of error reduction in this group (2,73%). He improved from an error percentage of (30%) in the pre-test to (27,27%) in the immediate post-test.

However, Table 4 also shows that 7 students from the DCF group and 8 students from the IDCF group did not record any improvement in the immediate post-test. Considering this initial comparison between the highest and the lowest scores of each group, it seems that the students of the DFC group scored higher, hence, outperformed those in the IDCF group. This may be due to the fact that the first group received a direct correction with rules and examples, thus, it was easy for them to assimilate the feedback and apply it in the post-test. The IDCF group was given indirect correction with codes. Therefore, they might have found it difficult to understand those symbols and, hence, were not able to apply them appropriately. Yet, it is significant to note that the number of students who marked improvement in the post-test from the IDCF group outnumbers those of the DCF group, and that this descriptive comparison concerns only those who scored the highest and the lowest gains in both groups.

In order to account for the mean percentage and standard deviation of each group as a whole, and in order to provide the statistical significance of the difference between the two groups, an Independent Sample T-Test was carried out as shown in Table 5 below:

Table5: Independent Sample T-test for the Pre-test and the Post-test

Abbildung in dieser Leseprobe nicht enthalten

From Table 5 above, it can be seen that IDCF group scored (35,35%) as a mean percentage of errors in the pre-test with a standard deviation of (8,65%) while the DCF gained a mean percentage of errors of (32,14%) with a standard deviation of (9,34%). The mean difference between these groups, then, is (3,21%). Yet, P value (sig. (2-tailed)) is 0,35 which means that the difference between the two groups is not statistically significant. This indicates that the two groups were at the same level when the pre-test was administered, which forms a strong basis for the comparison in the post-test.

The T-test in Table 5 above revealed that the IDCF group achieved a mean percentage of errors of (36,68%) in the post-test with a standard deviation of (8,08%), making a progress of (1,33%) from the pre-test. The DCF group, on the other hand, scored a mean percentage of (32,78%) with a standard deviation of (12,37%), recording, thus, a slight progress of (0,64%) from the pre-test. The mean difference between the two groups in the post-test is (3,9%) but still the P value (Sig.= 0,33) indicates that the difference is not statistically significant.

Based on the description of the T-test results above, it is obvious that the IDCF group outperformed the DCF group with a mean difference of (3,9%). This implies that , unlike what was stated in the initial descriptive statistics, the participants in the IDCF had gone through the feedback given to them, understood it and were able to apply it in the immediate post-test. However, the DCF students recorded a very slight benefit from the treatment in the immediate post-test. This may be because explicit feedback does not provoke the learners' problem-solving skills, hence, it becomes ephemeral. Therefore, the learners could not retain it during the post-test.

All in all, this comparison has, indeed, provided an answer to the second research question which is that indirect coded corrective feedback seems to be more effective than the direct corrective feedback accompanied with meta-linguistic explanations. Subsequently, the second research hypothesis is refuted.

3. Results of the Interview

In order to reinforce the findings of the second research question and in order to composite for the delayed post-test, personal structured interviews were conducted with four students who were picked up randomly from both groups. The numbers "1" and "2" are used here to represent the two students from the DCF group (1 male and 1 female) respectively while the numbers "3" and "4" are used to represent the students from the IDCF group (1male and 1 female) respectively. The interviews contained two questions as follows:

1) What problems did you face while doing correction based on the feedback you were given?
2) Do you think the feedback you received will help you use the target structure in future situations?

Each of the students was asked the two questions and required to answer in the language they prefer, they all answered in English and their answers were audiotaped and analysed below.

1. What problems did you face while doing correction based on the feedback you were given?

Based on the data collected, students 1 and 2 stated that they easily understood the feedback given to them with no difficulty to mention. Students 3 and 4, however, faced some problems in understanding those codes they received. Student 3, for example, said that he spent more time wondering about the meaning of those symbols. His answer is extracted in figure 1 below:

Figure1: Student 3 Responding to the First Question of the Interview

Interviewer: What problems did you face while doing correction based on the feedback you were given?

Student 3: Well, when I saw my answers underlined I understood that I made some mistakes, but I spent more time to understand those symbols to know the nature of the mistakes.

From the response above, it can be assumed that students faced problems in understanding the indirect coded feedback and that they have to activate their problem-solving skills in order to assimilate the correction unlike those who received direct correction with examples, hence, were able to understand it easily. This reinforces the assumption made in the descriptive results of the second question that activating students problem-solving skills may help in retaining the feedback.

2. Do you think the feedback you received will help you use the target structures appropriately in future situations?

The four participants stated that the feedback they received will help them in future occasions. Students 1 and 2 stated that the repeated application of the explicit feedback they were given will help them retain it.

Figure 2 below excerpts the answer of student 2:

Figure2: Student 2 Responding to the Second Question of the Interview

Interviewer: Do you think the feedback you received will help you use the target structures appropriately in future situations?

Student 2: Yes. I think the frequent practice of this feedback will help me maintain it the future.

The answer in Figure 2 above indicates that, though students easily understand direct CF, they are subjected to forget it if there is no frequent practice. This answer seems to agree with the assumption induced from the descriptive results of research question two that, without practice, direct feedback is likely to fleet easily.

Students 3 and 4, however, stated that since the feedback stimulated their problem-solving skills, it will be easy for them to maintain it in future situations. This is clear from the answer of student 3 in figure 3 below:

Figure3: Student 3 Responding to the Second Question of the Interview

Interviewer: Do you think the feedback you received will help you use the target structures appropriately in future situations?

Student3: When I read the feedback, I tried to use my critical thinking to understand the meaning of those symbols. This will help me retain the feedback for future.

From the response above, it is clear that students try to activate their problem-solving skills in order to figure out the meaning of the feedback, hence, it is more likely to last for future use. This claim may serve to explain the findings of the second research question that IDCF group outperformed the DCF group due to the fact that indirect coded feedback provokes students' problem-solving skills; therefore, it is likely to last for a long term.

To sum up, the findings of the interview seem to support those of the second research question. The questions of the interview revealed though students seem to face some problems understanding the codes they were provided with; they believe that indirect coded CF is more likely to retain for future occasions. This strongly reinforces the findings of the second research question in rejecting the second research hypothesis.

Conclusion

In brief, this chapter has reported the results and analyses of the research questions. The first research question revealed that written CF was effective regarding the concerned structures; therefore, the first research hypothesis was accepted. As for the second research question, it highlighted the effectiveness of indirect coded CF over direct CF with written meta-linguistic explanation; thus, the second null hypothesis was rejected. The results of the interview, in turn, seemed to support the findings of the second research question, refuting the second research hypothesis.

Chapter Four: Discussion and Recommendations

Introduction

This study aimed at investigating the effectiveness of direct written CF, with written meta-linguistic explanation, and indirect coded written CF. The study is composed of two main questions. The first question investigates the overall effectiveness of written CF while the second question compares the direct and indirect provision strategies. The previous chapter has presented the results and analyses of these two questions. This chapter, in turn, serves to discuss these results with previous literature and studies in the field. In addition, some recommendations are suggested for language teachers based on the results of the study.

1.Discussion of the Results

1.1 Effectiveness of written CF

Concerning the first question, this study found that both experimental groups, though not statistically significant, outperformed the control group in the post-test on new exercises. Therefore, written CF is found to be effective in helping students improve their writing accuracy. This finding echoes with many previous studies in the field (Ferris Robert 2001, Bitchener 2007, Sheen 2007, Ellis 2008, Choi, S.H 2013, Bitchener Rummel 2015...). For Bitchener and Rummel (2015), they focused on English articles and simple past tense as the target structure and found that written CF was effective in targeting these structures even though the difference between the experimental group and the control group was not statistically significant. Bitchener (2007), Sheen (2007), Ellis (2008) and Choi (2013), however, focused only on English definite and indefinite articles, yet, the result was the same; written CF was found to be effective.

Interestingly, the present study adds to the existing evidence which supports the effectiveness of written CF that written CF can be effective on linguistic structures other than articles, on which most of the recent studies focused. In fact, the current study found that written CF can also be effective regarding the use of simple past and present perfect tenses.

Students' benefit from the correction in the current study also suggests some support to Slinker's (1972-1992) Inter-language hypothesis which states that error correction is a way to improve students' performance (as cited in Ur, P. 1996, P. 244). This fact also reinforces Schmidt's (1995-2001) Noticing hypothesis which argues that second language acquisition takes place only when the learners pay close attention to the structures of the target language (as cited in Alkhawajah, F. 2016, P. 110).

On the other hand, this study contributes to the evidence that refutes the view held by a number of early studies (Truscott 1996, 2004 2007, Semke 1984, Kepener 1991, polio et al. 1998) that written error correction is ineffective, harmful and should be abandoned. In addition, the findings of the first research question of this study contradict with Wanger's (2016) claim that written CF is likely to work only with structures which have less than two usage options. This study, in fact, contained linguistic structures with more than binary usage options including the irregular past and past participle forms of verbs and still the correction was effective.

1.2 Effectiveness of the Two CF Types

The second research question of the study found that indirect CF group outperformed the direct CF group with a mean difference of ( 3.9%). Therefore, it can be assumed that indirect coded CF is more effective than direct CF with meta-linguistic explanation. This finding is found to corroborate with that of Alkhawajah's (2016) study. This latter was conducted with ESL learners with an Arabic native language background and revealed that indirect coded feedback was more effective concerning rule-based features (e.g., simple present) while direct correction was effective on item-based features (e.g., prepositions). The current study is also in line with Westmacott, A's (2016) study. In EFL settings in China, Westmacott found that indirect coded feedback was more useful as it prompts learners' cognitive processing.

In much the same regard, the results of this study, with regard to the findings of the interview, support and are supported by the theoretical claim held by Ellis (2008) and Ferris Bitchener (2012) that indirect corrective feedback is more likely to work with high-proficiency learners who have already internalised the target structure. The participants of the study at hand have already studied the target structure in previous stages of their study.

On the other hand, the findings of this study concerning the second question contradict some studies in the literature of the field (Lee, C. S 2014, Bitchener Rummel 2015, Goksoy Nazli 2016.). Bitchener Rummel (2015) found that there was no difference between the two feedback strategies though both groups had improved in accuracy. Lee, however, revealed that direct correction is more effective and the gains were retained in a delayed post-test. For Goksoy Nazli (2016), they found that indirect coded correction is likely to work just in the area of vocabulary and spelling while direct correction works elsewhere. However, the present study revealed that indirect coded feedback is likely to work also with treatable errors and rule-based structures such as simple past and present perfect.

Nevertheless, the discrepancy between the current study and the aforementioned studies could be attributed to methodological and design differences. Bitchener Rummel (2015), for example, conducted their study with migrant students most of whom took English courses for eight years. In addition, the study of Lee (2014) was conducted in ESL settings and the participants' native language was Mandarin. These methodological differences, including students' level of proficiency and the settings of the study, may affect the efficacy of the different types of CF (Ferris Bitchener, 2012, P. 56).

2. Recommendations

Based on the above findings and discussion, some recommendations could be suggested for language teachers in general and EFL writing instructors in particular. First of all, the findings of this study indicated that the written CF is overall effective in helping students improve their writing accuracy on a selected number of linguistic structures (present perfect and simple past). Therefore, it is suggested that teachers provide their students with corrective feedback according to their proficiency level. Yet, while providing this feedback, it is advised that teachers target a focused number of linguistic structures at a time instead of a comprehensive error correction. Secondly, the second research question revealed that indirect coded CF was more effective than direct CF with meta-linguistic explanation. Thereupon, it is recommended that language instructors provide intermediate and advanced learners with indirect coded feedback since this latter, as proposed by Ellis (2008) and Ferris Bitchener (2012), is more likely to provoke students' problem-solving skills, hence, has long-lasting effects. However, for low-proficiency students, it is suggested that they receive direct corrective feedback with meta-linguistic explanation and examples because, according to Ellis (2007), Bitchener knotch (2009) and Ferris Bitchener (2012), it provides them with clear answers and they can assimilate it easily

Conclusion

This chapter has extensively discussed the findings of the study in relation to previous literature in the field. In brief, this chapter revealed that the study contributes to the evidence in support of written CF; in addition, the study also highlighted the effectiveness of indirect coded CF over direct correction with meta-linguistic explanation. Based on these findings, some recommendations were suggested for EFL writing instructors.

General Conclusion

In conclusion, this study was set to investigate and compare the effectiveness of direct CF, with meta-linguistic explanation, and indirect coded CF. The first chapter of this paper introduced a brief review of the findings from the major studies in the field while the second chapter elucidated the methodological design of the study. The third and the fourth chapters, in turn, were devoted to report the findings of the study and to discuss them with other previous studies in the field. In addition, some recommendations were suggested at the end of the fourth chapter. Concisely, this study found that written CF is overall effective in helping students improve their writing accuracy regarding the use of present perfect and past tenses. In addition, the study also found out that indirect coded CF was more effective than direct CF with meta-linguistic explanation regarding the concerned structures.

Nevertheless, the findings of this study are to be interpreted with caution due to some limitations. First of all, the sample of this study was relatively short (N= 38). Therefore, absolute generalisations about EFL learners could not be made. Consequently, further research encompassing large samples with different ages and backgrounds is needed in order to have more generalised results. In addition, this study contained only one post-test which is the treatment with no delayed post-test. Instead, the researcher used an interview to compensate for the delayed post-test and to see whether or not students could retain the feedback for a long term. Accordingly, further research which includes a delayed post-test is required in order to account for the long-term effectiveness of these Two CF types. Finally, though it was not one of the objectives of the study, the interview revealed that students have different opinions concerning the two types of CF. This may imply that the effectiveness of different types of CF may vary according to different learners' strategies. Thus, further research is needed to investigate the effectiveness of different CF types with regard to learners’ strategies.

References

Alkhawajah, I. F. (2016). The effect of direct and indirect written corrective feedback on the acquisition of rule-based and item-based linguistic features (PHD thesis). Graduate Program in Linguistic and Applied Linguistics York University Toronto, Ontario.

Bitchener, J. (2008). Evidence in support of written corrective feedback. Journal of Second Language Writing, 17, 102-118. doi:10.1016/j.jslw.2007.11.004.

Bitchener, J., Young, S. Cameron, D. (2005). The effect of different types of corrective feedback on ESL student writing . Journal of Second Language Writing,14, 191-205. doi:10.1016/j.jslw.2005.08.001

Bitchener, J., Knoch, U. (2008). The value of written corrective feedback for migrant and international students. Language Teaching Research Journal, 12(3), 409-431. http://ltr.sagepub.com

Chandler, J. (2003). The efficacy of various kinds of error feedback for improvement in the accuracy and fluency of L2 student writing. Journal of Second Language Writing, 12 (3), 267-296. doi:10.1016/S1060-3743(03)00038-9.

Choi, S. H. (2013). The effects of written corrective feedback on second language writing focused on the English article system(MA thesis). Graduate College of the University of Illinois at Urbana-Champaign.

Creswell, J. W. (2002). Research design: qualitative, quantitative, and mixed methods approaches (2nd ed).

Ellis, R. (2009a). Corrective feedback and teacher development . L2 Journal, 1(1), 3-18. doi 10.5070/l2.v1i1.9054

Ellis, R. (2009b). A typology of written corrective feedback types. ELT Journal, 63(2), 97-107. doi:10.1093/elt/ccn023

Ellis. R., Sheen. Y. (2006). Corrective feedback in language teaching. In E. Hinkel (Eds.), Handbook of Research in Second Language Teaching and Learning Vol 2 (pp. 593-610), New York: Routledge.

Ferris, D. (1999). The case for grammar correction in L2 writing classes: A response for Truscott (1996). Journal of Second Language Writing, 8, 1-10.

Ferris, D. R. (2004). The “Grammar Correction” debate in L2 writing: Where are we, and where do we go from here? (and what do we do in the meantime….?). Journal of Second Language Writing, 13, 49–62. doi:10.1016/j.jslw.2004.04.005.

Ferris, D., Roberts, B. (2001). Error feedback in L2 writing classes: How explicit does it need to be? Journal of Second Language Writing, 10, 161-184.

Ferris, D. R., Bitchener, J. (2012). Written corrective feedback in second language acquisition and writing PART I. New York, NY: Routledge.

Ferris, D. R., Bitchener, J. (2012). Written corrective feedback in second language acquisition and writing PART II. New York, NY: Routledge.

Freeman, D. L., Anderson, M. (2000 ). Techniques principles in language teaching (3rd ed). Great Clarendon Street, Oxford: Oxford University Press.

Goksoy, A. S., Nazli, O. P (2016). The effect of direct and indirect writing corrective feedback on students’ writing. Middle East Technical University English Language Teaching Undergraduate Students Conference, 16-25.

Hendrickson, J. (1978) Error correction in foreign language teaching: recent theory, research, and practice. The Modern Language Journal, 62(8), 387-398. http://links.jstor.org/sici?sici=0026

Housen, A., Kuiken, F. (2009). Complexity, accuracy and fluency in second language acquisition. Applied Linguistics, 30(4), 461-473. doi: 10.1093/applin/amp048.

Hosseiny, M. (2014). The role of direct and indirect written corrective feedback in improving iranian EFL students' writing skill. Procedia - Social and Behavioral Sciences, 98, 668 – 674. doi: 10.1016/j.sbspro.2014.03.466. http:// CC BY-NC-ND license.

Kang, E., Han, Z. (2015). The efficacy of written corrective feedback in improving L2 written accuracy: a meta-analysis. The Modern Language Journal, 99(1), 1-18 doi: 10.1111/modl.12189.

Kothari, C. R. (2004). Research methodology: methods and techniques (2ND ED). New Age International publishers.

Lee, S. C. (2014). The effects of written corrective feedback on the use of present tenses among ESL learners (PHD thesis). Faculty of Languages and Linguistics University of Malaya, Malaysia.

Nassaji, H., Fotos, S. (2011). Teaching grammar in second language classrooms: integrating form focused instruction in communicative context. New York: Routledge.

Naylor, H., Murphy, R. (2001). Essential grammar in use supplementary exercise with answers (2nd ed). The Edinburgh building, Cambridge, UK: Cambridge University Press.

Rummel, S., Bitchener, J. (2015). The effectiveness of written corrective feedback and the impact Loa learners’ beliefs have on uptake. ARAL, 38(1), 66-84 /dio 10.1075/aral.38.1.04rum,

Shen, Y. (2013). Balancing accuracy and fluency in English classroom teaching to improve Chinese non-English majors‘ oral English ability. Theory and Practice in Language Studies, 3 (5), 816-822, doi:10.4304/tpls.3.5.816-822.

Storch, N. (2010). Critical Feedback on Written Corrective Feedback Research . IJES, International Journal of English Studies,10(2), 29-46.

Truscott, J. (1996). The case against grammar correction in L2 writing classes. Language
Learning
, 46, 327-369.

Ur, P. (1996). A course in language teaching: practice and theory. The Edinburgh building, Cambridge, UK: Cambridge University Press.

Wanger, J. (2016). An explanation of the effectiveness of written corrective feedback in second language(PHD thesis). The Graduate Faculty of George Mason University, Virginia.

Westmacott, A. (2016). Direct vs. indirect written corrective feedback: student perceptions. Íkala, Revista de Lenguaje y Cultura, 22(1), 17-32. doi: 10.17533/udea.ikala.v22n01a02.

Appendices

1. Appendix A: A Sample of Controlled Practice Exercise for the Pre-test.

Controlled Practice Exercise

Please state your full name:

Please state your age: between 18-23

Abbildung in dieser Leseprobe nicht enthalten

between 23-27

Abbildung in dieser Leseprobe nicht enthalten

above

Did you study English in high school? Yes

Abbildung in dieser Leseprobe nicht enthalten

no

1) Read the letter from Chris to her friend, Jo. Chris is from San Francisco and has just had a holiday with Jo in London. Complete the sentences, using either present perfect, present continuous, simple present or simple past.

Abbildung in dieser Leseprobe nicht enthalten

2. Appendix B: A Sample of controlled Practice Exercise For the Post-test.

Abbildung in dieser Leseprobe nicht enthalten

3. Appendix C: A Sample of Direct CF with Written Meta-Linguistic Explanation

Abbildung in dieser Leseprobe nicht enthalten

4. Appendix D: A Sample of Indirect Coded CF

Abbildung in dieser Leseprobe nicht enthalten

Abbildung in dieser Leseprobe nicht enthalten

5. Appendix E: The Interview

Abbildung in dieser Leseprobe nicht enthalten

[...]

Excerpt out of 60 pages

Details

Title
The Effects of Direct and Indirect Written Corrective Feedback on EFL Students' Accuracy
College
Sultan Moulay Sliman University
Author
Year
2018
Pages
60
Catalog Number
V494049
ISBN (eBook)
9783346073815
ISBN (Book)
9783346073822
Language
English
Keywords
effects, direct, indirect, written, corrective, feedback, students, accuracy
Quote paper
Zouhaiyr Ezzahouani (Author), 2018, The Effects of Direct and Indirect Written Corrective Feedback on EFL Students' Accuracy, Munich, GRIN Verlag, https://www.grin.com/document/494049

Comments

  • No comments yet.
Look inside the ebook
Title: The Effects of Direct and Indirect Written Corrective Feedback on EFL Students' Accuracy



Upload papers

Your term paper / thesis:

- Publication as eBook and book
- High royalties for the sales
- Completely free - with ISBN
- It only takes five minutes
- Every paper finds readers

Publish now - it's free