The effects of teacher direct and indirect corrective feedback on grammatical errors in the writings of 10th grade students at kim anh high school

ABSTRACT This study investigated the effects of teacher written direct and indirect corrective feedback on 10th grade students‟ three types of grammatical errors, namely verb tenses and

VIETNAM NATIONAL UNIVERSITY, HANOI UNIVERSITY OF FOREIGN LANGUAGES AND INTERNATIONAL STUDIES

FACULTY OF POST-GRADUATE STUDIES

PHÙNG THỊ BÍCH NGUYỆN

THE EFFECTS OF TEACHER DIRECT AND INDIRECT CORRECTIVE FEEDBACK ON GRAMMATICAL ERRORS IN THE WRITINGS

Ảnh hưởng của hình thức phản hồi trực tiếp và gián tiếp

của giáo viên đối với việc sửa các lỗi ngữ pháp trong bài viết

của học sinh lớp 10 trường THPT Kim Anh

ACKNOWLEDGEMENTS

Firstly, I would like to express my gratitude to my supervisor, Ms Hoang Thi Xuan Hoa, PhD for her helpful guidance, critical comments and enthusiastic support

I would not have been able to complete this thesis without her great encouragement

I am indebted to my colleagues for their valuable assistance for the pursuit and fulfillment of this course

My heartfelt thanks go to the students who willingly spent their time participating in the study

Last but not least, I owe a great debt of gratitude to my beloved family members and close friends, who always stand by me no matter what happens

ABSTRACT

This study investigated the effects of teacher written direct and indirect corrective feedback on 10th grade students‟ three types of grammatical errors, namely verb tenses and forms, attitudinal adjectives and articles, in their writings Forty five students were randomly assigned into three groups and received direct corrective feedback, indirect corrective feedback and no feedback respectively Students in each group produced three narrative paragraphs (a pretest, a revision, a post-test) The results of data analysis showed that both direct and indirect corrective feedback helped participants reduce errors related to verbs and attitudinal adjectives from the pretest to the post-test However, only direct feedback contributed to a significant reduction of article errors from the pretest to the post-test Between group comparisons of the post-test results revealed that direct corrective feedback appears to be slightly more effective than indirect corrective feedback in helping participants reduce errors related

to verbs and attitudinal adjectives In addition, the effect brought about by direct and indirect feedback on reducing article errors compared to the control group was not statistically significant The findings are discussed in the context of the related

literature and areas of future research are suggested

TABLE OF CONTENTS

Acknowledgements i

Abstract ii

List of Tables vi

List of Figures vii

List of Abbreviations vii

CHAPTER 1: INTRODUCTION 1.1 Statement of the problem and rationale of the study 1

1.2 Aims and objectives of the study 3

1.3 Significance of the study 3

1.4 Scope of the study 4

1.5 Organization of the paper 4

CHAPTER 2: LITERATURE REVIEW 2.1 Approaches to the teaching of L2 writing 5

2.1.1 Product approach 5

2.1.2 Process approach 6

2.2 Grammatical errors in L2 writing 7

2.3 Feedback in L2 Writing 8

2.3.1 Definition 8

2.3.2 Types of feedback 9

2.3.2.1 Direct Corrective Feedback 9

2.3.2.2 Indirect Corrective Feedback 10

2.3.3 Review on studies about feedback issues 10

2.3.3.1 The effectiveness of corrective feedback 10

2.3.3.2 The impacts of direct and indirect corrective feedback 12

CHAPTER 3: METHODOLOGY 3.1 Setting 15

3.2 Research method 17

3.3 Participants 15

3.4 Data collection 16

3.4.1 Data collection instruments 16

3.4.2 Data collection procedure 17

3.5 Data analysis 20

CHAPTER 4: RESULTS AND DISCUSSION 4.1 Pretest results 22

4.1.1 Verb errors in the pretest 22

4.1.2 Errors of attitudinal adjectives in the pretest 23

4.1.3 Errors of articles in the pretest 24

4.2 Post-test results 25

4.2.1 Answer to the first research question 25

4.2.1.1 Verb tenses and forms 25

4.2.1.2 Attitudinal adjectives 27

4.2.1.3 Articles 29

4.2.1.4 Summary of the results and discussion 31

4.2.2 Answer to the second research question 32

4.2.2.1 Verb tenses and forms 33

4.2.2.2 Attitudinal adjectives 34

4.2.2.3 Articles 36

4.2.2.4 Summary of the results and discussion 37

CHAPTER 5: CONCLUSION 5.1 Summary of findings 38

5.2 Limitations of the study and suggestions for further studies 39

5.3 Implications 39

REFERENCES

APPENDICES:

Appendix 1: Placement Test

Appendix 2: Pre-test

Appendix 3: Post-test

Appendix 4: Pretest result

Appendix 5: Post-test result

Appendix 6: Students‟ writing samples

LIST OF TABLES

Table 3.1- Example of direct corrective feedback and indirect corrective feedbackTable 3.2.: Summary of data collection procedure

Table 4.1- Descriptives (Test 1 - Verb)

Table 4.2 - Test of Homogeneity of Variances (Test 1 - Verb)

Table 4.3 - ANOVA (Test 1 - Verb)

Table 4.4 - Descriptives (Test 1 – Adj)

Table 4.5 - Test of Homogeneity of Variances (Test1 - Adj)

Table 4.6 - ANOVA (Test1 - Adj)

Table 4.7 - Descriptives (Test1 - Article)

Table 4.8 - Test of Homogeneity of Variances (Test1 - Article)

Table 4.9 - ANOVA (Test1 - Article)

Table 4.10 - Paired Samples Statistics (Pair 1 – DF - Verb)

Table 4.11 - Paired Samples Test (Pair 1 – DF - Verb)

Table 4.12 - Paired Samples Statistics (Pair 2- IF - Verb)

Table 4.13 - Paired Samples Test (Pair 2 – IF - Verb )

Table 4.14 - Paired Samples Statistics (Pair 3 – NF - Verb)

Table 4.15 - Paired Samples Test (Pair 3 – NF – Verb )

Table 4.16 - Paired Samples Statistics (Pair 4 – DF – Adj)

Table 4.17 - Paired Samples Test (Pair 4 – DF – Adj)

Table 4.18 - Paired Samples Statistics (Pair 5 – IF – Adj)

Table 4.19 - Paired Samples Test (Pair 5 – IF – Adj)

Table 4.20 - Paired Samples Statistics (Pair 6- NF – Adj)

Table 4.21 - Paired Samples Test (Pair 6 - NF – Adj)

Table 4.21 - Paired Samples Test (Pair 6 - NF – Adj)

Table 4.22 - Paired Samples Statistics (Pair 7- DF – Article)

Table 4.23 - Paired Samples Test (Pair 7- DF – Article)

Table 4.24 - Paired Samples Statistics (Pair 8 - IF – Article)

Table 4.25 - Paired Samples Test (Pair 8 – IF – Article)

Table 4.26 - Paired Samples Statistics (Pair 9 – NF – Article)

Table 4.27 - Paired Samples Test (Pair 9 – NF – Article)

Table 4.28 - Descriptives (Test 2 – Verb)

Table 4.29 - Test of Homogeneity of Variances (Test 2 – Verb) Table 4.30- ANOVA (Test 2 – Verb)

Table 4.31 - Multiple Comparisons (Test 2 – Verb)

Table 4.32 - Descriptives (Test 2 - Adj)

Table 4.33 - Test of Homogeneity of Variances (Test 2 - Adj) Table 4.34 - ANOVA (Test 2 - Adj)

Table 4.35 - Multiple Comparisons (Test 2 – Adj)

Table 4.36 - Descriptives (Test 2 – Article)

Table 4.37 - Test of Homogeneity of Variances (Test 2 – Article) Table 4.38 - ANOVA (Test 2 – Article)

Table 4.39 - Multiple Comparisons (Test 2 – Article)

LIST OF FIGURES

Figure 3.1: Process of setting up groups

LIST OF ABBREVIATIONS

EFL: English as Foreign Language

ESL: English as Second Language

CHAPTER 1: INTRODUCTION 1.1 Statement of the problem and rationale of the study

Obviously, writing plays an important role in communication because apart from speaking writing enables people to express their feelings and thoughts They use

it to define themselves, clarify their knowledge and their ideas, to understand the problems that may face them as well as to find solutions for such problems Being so, writing is a means for self-actualization What people learn about themselves and develop within themselves through writing can help them to realize their individual potential and to achieve potential goals (Hughey, et al, 1983)

In the field of second and foreign language learning and teaching, writing has been seen as essential integral part of any language syllabus as learning to write effectively seems to be of prime importance Writing also enhances language acquisition because learners experiment with words, sentences, and large chunks of writing to communicate their ideas effectively and to reinforce grammar and vocabulary they are learning in class (Bello, 1997) It helps to consolidate learning to render it available for use in other areas such as listening, speaking and reading (Mohamed, 2000)

In spite of the importance of writing, it is one of the most difficult language skills to master According to Richards and Renandya (2002), the difficulty lies not only in generating and organizing ideas, but also in translating these ideas into readable text They argue that the skills involved in writing are highly complex, and learners have to pay attention to higher level skills of planning and organizing as well as lower level skills of spelling, punctuation, word choice, and so on The difficulty becomes even more pronounced if student writers are at a low level of language proficiency

In Vietnam, of four skills that students learn in the subject of English, much attention has been paid to developing writing skill In 2005, the Ministry of Education and Training decided to reform the structure of the English test in Vietnam National High School Graduation Examination 2015 with more emphasis on the writing skill Accordingly, apart from multiple choice grammar questions, students were asked to write a short paragraph However, the result of the exam indicates that students‟

writing performance was generally poor as a large number of students could hardly do the task Regarding students‟ attitude towards learning writing, according to Le (2008) (as cited in Pham (2009), only 6.9% of students wanted to learn writing When asked about the reality of English writing at a specialized high school in the Mekong Delta, teachers reported that most of the students had problems with their writing such as (1) students have few opportunities to respond to teacher feedback thoughtfully and critically, (2) their written texts contain a lot of grammatical inaccuracies and (3) they have negative attitudes toward writing in English It can be seen that the teaching and learning writing English in Vietnam seems to face considerable difficulties

Given the importance of writing in communication and in the field of EFL learning and teaching, the difficulties facing learners in developing their writing ability over time in EFL contexts, a lot of research has been conducted in order to explore the effects of various techniques on helping ELF learners become better writers, especially the impact teacher feedback on students‟ writings However, the controversy over the effectiveness of corrective feedback remains unresolved According to the most extreme views, such as Truscott (2007), corrective feedback is seen as not only ineffective but also potentially harmful In contrast, a number of L2 researchers and practitioners (e.g Bitchener and Knoch, 2008; Chandler, 2003; Ferris, 2002; Sheen, 2007) claim that corrective feedback is of value in promoting greater grammatical accuracy Also, there have been studies that examined the relative effectiveness of varying feedback types, and direct and indirect corrective feedback has received due attention of researchers To date, studies examining the effects of direct and indirect feedback on L2 learners‟ writing have yielded mixed results Furthermore, in Vietnam, there is little research addressing the issue Especially, most previous studies have participants who are university students, and hardly any studies whose participants are high school students can be found

The above-mentioned reasons are the incentive for the researcher to conduct a

study on “The effects of teacher direct and indirect corrective feedback on

grammatical errors in the writings of 10 th grade students at Kim Anh High School” The results of the study are expected to make a contribution to the ongoing

debate about the issue and to help the researcher to use corrective feedback more effectively in the teaching of writing

1.2 Aims and objectives of the study

The current study aims at investigating the effects of teacher direct and indirect corrective feedback on grammatical errors in the writings of 10th grade students at Kim Anh High School

In order to achieve the aims, the study seeks to find the answer to the following research questions:

1 Do teacher written direct corrective feedback and indirect corrective feedback help 10th grade students at Kim Anh High School to reduce grammatical errors in writing from an initial task to a new task?

2 If so, which kind of feedback (direct corrective feedback or indirect corrective feedback) is more effective?

This study proposes the following hypotheses:

1 Teacher direct corrective feedback and indirect corrective feedback have significant effect on helping 10th grade students at Kim Anh High School to reduce grammatical errors in writing from an initial task to a subsequent task

2 Teacher direct corrective feedback is more effective

1.3 Significance of the study

Findings of this study will contribute to the discussion about whether or not corrective feedback is effective and helpful in helping students reduce grammatical errors in their subsequent writings

Unlike some previous studies which did not have a control group, this study will involve a treatment group receiving both teacher written indirect and direct corrective feedback and a control group receiving no corrective feedback

This study will also focus on three grammatical errors frequently made by EFL students, which is different from several previous studies that were either too broad by addressing too many areas of error or too specific by focusing only on one error

As far as the teaching of writing is concerned, the study will also contribute to the practice of teaching writing in EFL context Findings of this study might be of interest to foreign language writing teachers and researchers, especially those at secondary and high schools They may have better understanding of the effects of teacher direct and indirect corrective feedback on students‟ writings; thence have more suitable feedback strategies to help their students

1.4 Scope of the study

The study is restricted to explore the effects of two strategies of feedback, direct corrective feedback and indirect corrective feedback on grammatical errors made by

10th grade students In addition, only three targeted grammatical errors were in this research They are the common types of grammatical errors and the ones that occurred most frequently during the first writing task, including verb tenses and forms, attitudinal adjectives and article usage

Regarding the sample of the research, the participants were only selected from

10th grade students at Kim Anh High School in the school year of 2014-2015

1.5 Organization of the paper

The study includes five chapters as follows:

Chapter 1 (Introduction) presents the problem and the rationale, the aims and objectives, the scope and the significance of the study

Chapter 2 (Literature review) provides the theoretical background of the study and reviews related studies

Chapter 3 (Methodology) describes the participants, the instruments and the procedures

to be employed to conduct the study

Chapter 4 (Results and Discussion) answers the research questions with data presentation, data analysis and the comparison among the finding themselves and the assumptions discussed in the Literature Review

Chapter 5 (Conclusion) summarizes the main issues discussed in the study, the

limitations of the research and some suggestions for further study

CHAPTER 2: LITERATURE REVIEW

This chapter sheds lights on the literature review of the study, specifically some approaches in teaching wiring, grammatical errors and teacher corrective feedback Firstly, an overview of approaches in teaching writing will be provided Secondly, grammatical errors will be discussed Finally, a closer look will be cast into teacher corrective feedback by outlining the main types of feedback, briefly reviewing main previous research on the effectiveness of feedback and summarizing the key findings

of main studies

2.1 Approaches to the teaching of L2 writing

There have been a number of ways of teaching writing However, the two most commonly and popularly known at the moment are the product-based approach and the process-oriented approach (Nunan, 1995)

2.1.1 The product approach

The product approach emerged as a combination of structural linguistics and behaviorist learning theory, which was popular in the 1960s (Silva, 1990) Product approach is the traditional approach to teaching writing which focuses at the result of the act of composition The writing teachers who subscribe to the product approach are more concerned to see what a final piece of writing will be like and measure it against criteria of vocabulary use, grammar use, and medical considerations such as spelling and punctuations, as well as content and organization (Brown, 1994) Students in the classes adopting the product approach typically are given writing models to imitate before they are evaluated by their teachers (Ferris and Hedgcock, 2005) and writing is taught through four stages, which are familiarization, controlled writing, guided writing and free writing

Since the focus in the product approach is on form, it is easy to use with large classes It is also easier to mark compositions because the teacher can easily direct his

or her attention to the form while correcting This approach is useful for situations where the emphasis on form is important or where the focus on structure is the main

target It has been widely used and teachers are quite familiar with it It might also be suitable for lower level learners because it helps them correct and eliminate their errors (Tribble, 1996)

However, product approach has a number of limitations First, as Ferris and Hedgcock (2005) explain, this approach focuses on the writing structure and use of vocabulary as the main indicators of writing development, neglecting the writing processes that students go through in writing (pre-writing, drafting, revising and editing) Second, it may create problems as it restricts the teaching of writing to syntactic and grammatical accuracy, thus limiting students‟ understanding of good writing (Hyland, 2003) Third, the focus on grammar skills has not proved to be effective in improving writing ability and many researchers reject the emphasis on grammar, arguing that it has little to do with the act of writing (Zamel, 1976) While Hinkel and Fotos (2002) believe that grammar teaching can be helpful and productive

in ESL and EFL writing classrooms, other studies take the opposite view Hudson (2001) argues that to prove that teaching grammar improves writing, further research needs to be carried out Another weakness in the approach lies in the assumption that good writing can be achieved by applying certain functional rules In fact, writing is much more than that (O‟Hare, 1973) Fourth, it restricts students‟ creativity as it relies

on imitation (Hyland, 2003) Fifth, the use of language in this approach is restricted to fixed patterns that are learned by imitating other models (Pincas, 1962)

2.1.2 The process approach

The process approach came into existence in reaction to product approach pedagogies (Miller, 1991) This approach leads students to the phase of a finished text publication as it goes beyond linguistic knowledge to focus on linguistic skills and involves identifiable stages (Merriwether, 1997) Reid (1993) describes writing as a multi-stage process Goldstein and Carr (1996) refer to the process of writing as a range of strategies that include prewriting, planning, drafting and revising (Hedge,

2005, as cited in Mohamed, 2013 ) explains that the process of composing a text goes through different stages of revision, editing and generating such as being motivated to

write, getting ideas together, planning and outlining, making notes, making a first draft, revising, editing and getting ready for publication (Hedge, 2005, as cited in Mohamed,

2013 )

Freeman and Freeman (2004) identify a number of advantages in the process approach First, it motivates students to deliver their own messages and become creative Second, it involves teachers and students in responses to texts through peer feedback and discussions Third, it deals with mistakes in writing skills such as spelling and grammar through teacher-student conferencing Fourth, it naturally moves writing from invention to convention (i.e writing becomes a practice of a set of cognitive process instead of a demonstration of linguistic knowledge) However, the process approach has some limitations First, it is time-consuming, especially with large classes Second, teacher-student conferences could be difficult to schedule due to time pressure Third, it requires a great deal of marking Fourth, it might discourage students who are not familiar with the process writing as they may consider revision as failure (Corpuz, 2011 )

2.2 Grammatical errors in L2 writing

Making errors is the most natural thing in the world and it is evidently attached

to the human beings (Beuningen, 2010) A number of experts in linguistics have presented various definitions of error Among them, Norrish (1983) considers error as a systematic deviation, when a learner has not learnt something and consistently gets it wrong Ellis (1994) also defines error as a deviation from the norms of the target language To support this point, Ellis (1994) further clarifies that the standard written dialect which is widely spread and used to teach non-native speakers is considered as the norm particularly in language classroom settings In short, the norms of the target language are defined as the standard written dialect, and if so, error refers to the deviation of that standard Tsui (1995) considers an error in the classroom as (1) something that is rejected by the teacher because it is wrong or inappropriate, (2) something that the teacher does not want or (3) something that does not conform to the rules which the teacher lays down

Richard (1992) and Ellis (1994) have a quite similar view on distinguishing between a mistake and an error They explain that errors occur when learners do not know the correct usage reflect gaps in learner knowledge whereas mistakes occur because, in particular instance, learners are unable to perform what they know reflect occasional failures in performance In this sense, error refers to a deviation which occurs when language learners have not yet acquired correct usage of the target language

Norrish‟s (1983) classified errors into three types, namely errors, mistakes and lapses An error is a systematic deviation which is made because the learner has not learnt the correct form After the learner is taught the language form, he may be able to use it correctly or may not sometimes The inconsistent deviation is termed mistake Another type of wrong usage, which is neither an error nor a mistake, is a lapse A lapse occurred due to lack of concentration, shortness of memory, fatigue, etc

Although there are different definitions and types of errors, in this paper the term “errors” is used to refer to both “errors” made when the learners try to do something with the language they do not know and “mistakes” which occur when the students have learned something but have forgotten it or are careless in their writings

2.3 Feedback in L2 Writing

This section provides an overview of a variety of issues related to feedback in L2 writing Firstly, the definitions and importance of feedback are presented Next, different types of feedback are briefly introduced Finally, key issues in feedback and research studies are tackled, with emphasis on the effectiveness of different types of corrective feedback on students‟ writing

2.3.1 Definition

Feedback is defined as teacher's input to a writer's composition in the form of information to be used for revision (Keh, 1990) Nicol and Macfarlane (2004) consider feedback as information provided by teachers to help students trouble-shoot their performance The writer of this paper would define it as teacher's response to students'

writing in the form of written comments that aim to help students improve their writing performance

Providing feedback is an essential function of teaching and learning (Beuningen,

2010) The importance of feedback is also stressed by Cole and Chan (1994) They stated that feedback may serve not only to let learners know how well they have performed but also to increase motivation and build a supportive classroom climate Providing the right kind of feedback to the students can make a significant difference

in their achievement

2.3.2 Types of feedback

Although feedback is classified into some different types, this part only presents key points about teacher direct corrective feedback and teacher indirect corrective feedback, which are the focus of this study

2.3.2.1 Direct Corrective Feedback

Direct corrective feedback is defined as a type of correction that draws students' attention to the error and provides a solution to it (Bitchener and Ferris, 2012) In other words, the teacher shows students where their errors are and corrects these errors by providing the correct form This type of correction takes a variety of forms such as a) cross-outs: when the teacher omits any wrong addition from students‟ original texts, b) rewrites: when the teacher rewrites a word, phrase or a sentence, providing the correct spelling, structure or form on students‟ original texts and

c) additions: when the teacher adds any missing items on students‟ original texts (e.g

prefix, suffix, article, preposition, word, etc)

Direct corrective feedback aims to help students edit their writing and improve their performance in future tasks (Bitchener and Ferris, 2012) Ferris (2002) argues that

it is useful in treating errors of prepositions and other issues of idiomatic lexis She also claims that it is useful in the final stages of the writing process to help students focus

on the remaining errors in their texts and refer to them in future tasks Students' linguistic proficiency is important to determine the amount of direct corrective feedback they receive as advanced learners are more likely to benefit from it

2.3.2.2 Indirect Corrective Feedback

Indirect corrective feedback refers to situations when the teacher marks that errors have been made but does not supply the correct forms, requiring the learners to diagnose and correct their errors (Lee, 2005) When giving indirect corrective feedback the teacher underlines, circles or highlights errors on students' original texts

to indicate the location of these errors without correcting them Students are asked to study their errors and correct them (Ferris, 2002) In other words, indirect corrective feedback emphasizes the role of students in understanding and correcting their errors rather than being provided with the corrections

Indirect feedback is applied by underlining students' writing errors so that students understand that there is a problem that should be 'fixed.' Teachers may use lines, circles or highlighting to indicate the location of errors They also need to decide how explicit indirect feedback should be based on the goals they want to achieve by providing feedback

2.3.3 Review on studies about feedback issues

This section briefly reviews main studies about corrective feedback According

to their research issues, studies are divided into two major groups namely the effectiveness of corrective feedback and the impacts of direct and indirect feedback

2.3.3.1 The effectiveness of corrective feedback

The discussion regarding the effectiveness of feedback has been prominent in recent years Attitudes toward corrective feedback have evolved from the strict avoidance of errors and thus quick and direct error correction before the 1960s, to the condemnation of error correction as something harmful in the late 1960s, and to a more critical view of the necessity and value of error correction in the 1970s and 1980s The controversy over the topic of corrective feedback, however, remains unresolved in the 1990s (Lee, 1997, cited in Khatib & Bijani, 2012, p 103)

The heated debate on the effectiveness of corrective feedback in L2 writing was initiated in 1996 in an article by Truscott (1996), who argued that grammar correction should be avoided in L2 writing and stressed that teachers should not correct grammar

because of its potentially harmful impacts He presented three arguments against error correction The first was that the learning process was too complex to believe that students could improve through providing them with corrective feedback Secondly, giving corrective feedback to students at a time when they were not ready to learn a specific language form or structure was barely possible Thirdly, he argued that whatever knowledge students acquired as a result of correction would dissipate over a short period These arguments led to an increase of research focusing on the effectiveness of corrective feedback on students‟ writing

Some researchers (e.g., Cohen and Robbins, 1976; Kepner, 1991; Sheppard, 1992; Polio et al 1998) also claim that grammar corrections do not have a positive effect on the development of L2 writing accuracy According to the most extreme views, such as Truscott (2007), corrective feedback is seen as not only ineffective but also potentially harmful

In contrast, a number of L2 researchers and practitioners (e.g Bitchener and Knoch, 2008; Chandler, 2003; Ferris, 2002; Sheen, 2007) claim that corrective feedback is of value in promoting greater grammatical accuracy For example, Ferris (1995, 1999, 2003, 2004) have rejected the argument raised by Truscott with respect to the ineffectiveness of implementing written corrective feedback on L2 learners‟ writing Ferris (1999) argues that Truscott‟s stance against grammar correction and ineffectiveness of corrective feedback on fostering L2 learners‟ writing ability is premature She claims that corrective feedback is an integral constituent of L2 writing instruction Furthermore, it is claimed that corrective feedback provides L2 learners opportunities to notice the gaps in their L2 knowledge which in turn leads them to test inter-language hypotheses and engage in meta-linguistic reflection which results in prompting L2 writing (Van Beuningen, 2010) Ferris also argues that corrective feedback is an inseparable issue in writing pedagogy and L2 researchers and practitioners are assumed to focus on questions with respect to effective implementation of corrective feedback

Chandler (2003) also argued that students who received corrective feedback then revised their writing improved over time

In their studies, (Ashwell, 2000; Ferris, 1997; Ferris & Roberts, 2001; Sachs & Polio, 2007) also found that participants whose errors were corrected were able to make more accurate revisions than those who did not receive any corrective feedback

However, results from studies investigating the effect of corrective feedback on subsequent writing (e.g Chandler, 2003; Kepner 1991; Polio et al., 1998; Semke, 1984), are inconclusive

2.3.3.2 The impacts of direct and indirect corrective feedback

There are some studies that examined the relative effectiveness of varying feedback types, and direct and indirect corrective feedback has received due attention

of researchers To date, studies examining the effect of direct and indirect feedback on L2 learners‟ writing have yielded mixed results

It has been claimed that students would benefit more from indirect corrective feedback because they have to engage in a more profound form of language processing

as they are self-editing their output (e.g Ferris, 1995) However, this hypothesis could not yet be confirmed since results from studies exploring the relative effectiveness of direct and indirect corrective feedback (e.g Chandler, 2003; Ferris et al., 2000; Frantzen, 1995; Lalande, 1982, Robb et al., 1986) are inconclusive

A study by Lalande (1982) showed that students who received indirect corrective feedback outperformed students in a direct correction group Pham (2009) also concluded that indirect feedback was more effective than direct feedback in helping reduce students‟ grammatical errors significantly Frantzen (1995) and Robb et

al (1986) on the other hand, found that direct and indirect corrective feedback were equally effective A study by Ferris et al (2000) revealed yet another pattern; whereas indirect correction proved to be most effective in improving students‟ accuracy in subsequent writing, students who received direct feedback made the most accurate revisions Finally, as opposed to Lalande (1982) and Ferris et al (2000), Chandler (2003) found that direct corrective feedback resulted in the largest accuracy gains, not only in revisions but also in subsequent writing

Van Beuningen et al (2012) investigated the effectiveness of direct and indirect corrective feedback on students‟ overall accuracy, grammatical accuracy, non-

grammatical accuracy, grammatical complexity and lexical diversity in L2 writing The result of the study showed that direct and indirect feedback was useful in improving grammatical and non-grammatical accuracy as both experimental groups outperformed the control groups For overall accuracy, the effect of direct corrective feedback was greater than indirect corrective feedback For grammatical complexity and lexical diversity, Van Beuningen et al (2012) wanted to examine Truscott‟s (2007) claim that corrective feedback resulted in simplified writing However, no significant difference was found between all groups in structural complexity or lexical diversity

While a lot of studies make claims about the effectiveness and ineffectiveness of corrective feedback, the results of studies are controversial

Studies that did include a control group and investigated the short-term effectiveness of error correction (e.g Ashwell, 2000; Ferris, 1997; Ferris & Roberts, 2001; Sachs & Polio, 2007), found that participants whose errors were corrected were able to make more accurate revisions than those who did not receive any corrective feedback In contrast, results from studies investigating the effect of corrective feedback on subsequent writing (e.g Chandler, 2003; Kepner 1991; Polio et al., 1998; Semke, 1984), are inconclusive Methodological shortcomings might explain the contradicting findings of these studies

Semke (1984), who compared the effects of error correction to the effects of content-focused comments, found that error correction had no effect on students‟ accuracy and a negative effect on their written fluency However, in his research, groups received different treatments in terms of time Students in the content-focused condition had twice as much time to produce new material than students receiving corrective feedback Therefore, it might be the case that these results could be explained by the different amount of writing practice between the two treatment groups

Polio et al.‟s (1998) study showed that both students who received corrective feedback and students who did not were able to improve their accuracy over time However, students in the experimental condition only produced half as many journal entries as the control group did, because of the editing activities they had to perform

Thus, it can be seen that the control group have more practicing opportunities and as a result, they might have better performance Perhaps, that is why the beneficial effects

of corrective feedback could not be significant

Kepner (1991) did not find any significant differences in error-counts between a group that received error corrections and another group that received message-related comments However, the participants in his research were not required to do anything with the corrective feedback they received Hence, it remains unclear whether the students processed the feedback that they had received or not

In Semke (1984) and Polio et al.‟s (1998) studies the fact that less time was allocated to writing practice in the error correction condition than in the control condition could explain why no positive effects of corrective feedback were found

In another experiment in which college learners who were required to write five essays that were collected every second week, Chandler (2003) found that getting the teacher to correct or to underline for learner self-correction resulted in a significant improvement in both accuracy and fluency in subsequent writing of the same type Nevertheless, direct correction led to the greatest accuracy in comparison with indirect corrections including underlining, description and underlining with description Yet, as learners were receiving different types of feedback in rotation, learning would have occurred between treatments Learners who received the direct correction treatment last might perform very differently from those who had received it first For example, having experimented with the three other feedback types, learners alerted to the fact that they should continue paying attention to their errors In other words, the direct correction type might have proven so effective because it was combined with other feedback treatments This study should be conducted with four different groups, working with the same teacher but without alternating treatment in order to see the effectiveness of direct correction

In short, although there has been a great deal of arguments for and against the efficacy of implementing corrective feedback on enhancing L2 writing, yet there is little available to reach to a firm conclusion Therefore, more studies should be conduct

to investigate different aspects of corrective feedback

CHAPTER 3: METHODOLOGY 3.1 Setting

The study is conducted in 10th grade of Kim Anh High School which is located

in a rural area of Soc Son District, a developing suburb of Hanoi In order to get the admission to the school, students have to take a high school entrance exam delivered annually by Department of Education and Training of Hanoi However, the examination consists of only two tests, including Mathematics and Literature test It means that all students are not required to take any tests of English in order to enter Kim Anh High School That is why a number of students do not pay adequate attention

to learning English at secondary school, and consequently their English level is lower than what is expected for 10th grade students

At Kim Anh High School, students have a 45-minute period of writing after each two weeks The writing lessons generally follow activities given in the textbook

“Tieng Anh 10” provided by the Ministry of Education and Training

3.2 Method

To investigate the effects of teacher direct and indirect corrective feedback on participants‟ grammatical accuracy in writing, a quasi-experimental study with one independent variable and one dependent variable was employed

The independent variable in this study was the teacher written corrective feedback strategies This variable comprised three levels namely direct feedback (DF), indirect feedback (IF) and no feedback (NF) The dependent variable was participants‟ writing grammatical accuracy as measured by the percentage of errors in the writing task 1 (pretest) and writing task 2 (post-test)

3.3 Participants

The participants of this study were 45 students from ten 10th grade classes In order to choose participants who are representatives of the desired population, these students were selected from 10 classes based on their English scores More

specifically, 63 students having English score at 10th grade of about from 7.0 to 7.5 coming from 10 classes were selected Because their English scores were based on different tests designed by different teachers, there might be a difference in the English proficiency of these participants Therefore, a placement test for these 63 students was necessary to select students with quite similar level of English proficiency As a rule, students of Kim Anh High School have to take a placement test of English, Math, Literature, Physics, Chemistry after 10th grade so that they can be placed into suitable groups corresponding to their levels Therefore, the researcher used the results of this test to select 45 participants from these 63 students More specifically, only 59 out of

63 students selected sat for the English test In addition, the 14 test takers whose scores were significantly higher or and lower than the others‟ were not invited to participate in the study As a result, 45 students were selected

All participants have been learning English as a compulsory subject at school for 7 years During the time of the study, participants were having summer vacation and were not taking any English classes at school

3.4 Data collection

3.4.1 Data collection instruments

Three data collection instruments, including an initial writing test (pretest), a revision of the initial task and a subsequent writing task (post-test), were employed in the study

Two writing tasks were used in the experiment Both initial task (pretest) and a subsequent writing task (post-test) were of a similar type – writing a narrative about an embarrassing situation on the basis of a series prompts given Students were instructed

to use about 100-120 words for each writing task These writing tasks were also designed in such a way that the content was unproblematic for all students, since the aim of this study was to elicit the effect of the direct and indirect corrective feedback

on grammatical accuracy The topic of the two writing tasks was similar because a learner‟s language proficiency may not be the only factor influencing students‟ performance on a particular writing task Research on writing assessment showed that

other factors, such as a task‟s topic, might also contribute to a writer‟s score (Schoonen, 2005) Therefore, participants were given two different writing tasks of a same topic so that topic influence factor can be controlled

Before having been used officially, the tests were piloted among 6 participants who were of similar backgrounds as those in the official study

3.4.2 Data collection procedure

The data collection procedure consists of 6 main steps presented as follow:

Step 1: Selecting participants

The researcher informed the teachers of the aims of the research, what the participants were expected to do to select students who had the English score at grade

10th of 7.0 -7.5

Step 2: Administering placement test and setting up groups

To enhance the equivalence of student starting performance level of writing among the groups at the beginning of this study, all participants took an English test which assessed participants‟ overall language proficiency After this test, 45 students out of 59 students whose test score were not significantly different were selected

Based on the performance level of the placement test, the students were classified into low to moderate and moderate to high score groups Students were then randomly assigned in three different groups, including (Direct corrective feedback (group 1) and indirect corrective feedback (group 2) and control group with no feedback (group 3), using a stratified random sampling technique After this stratified random assignment, each treatment group had 15 students consisting of students drawn from the low to moderate English proficiency level group and students from the moderate to high proficiency level group These steps are outlined in Figure 3.1.In this way, it was expected that each group would have similar compositions of students with similar levels of English

Figure 3.1: Process of setting up groups

Group 1 (15 students) Direct feedback

- 8 students (Group A)

- 7 students (Group B)

GROUP A Low to moderate score (23 students)

Group 2 (15 students) Indirect feedback

- 8 students (Group A)

- 7 students (Group B)

GROUP B Moderate to high score

Classifying participants according

to their test score

Step 3: Delivering the initial test

Before administering the first writing task, the researcher introduced the task to ensure all participants had a comparable background knowledge on the topic Then all participants were required to write a narrative of about 100-120 words about an embarrassing situation, using the prompts given They had 35 minutes to complete the task

Step 4: Giving written corrective feedback

A teacher of English who had been trained in feedback strategies provided feedback on the paragraphs written by students The first experimental group received direct corrective feedback, the second experimental group received indirect feedback, and the control group was given no corrective feedback However, for ethical purposes, students in the control group were provided with general comments about their writing, such as “please keep up the good work.”

While direct corrective feedback took the form of identifying both the error and the target form (c.f example 1), indirect corrective feedback only consisted of an indication of the error and its category (c.f example 2)

Table 3.1.: Example of direct corrective feedback and indirect corrective feedback Example 1: Direct corrective feedback He play football yesterday

played Example 2: Indirect corrective feedback (Art) I read a interesting book last week

( Art = article error)

Step 5: Asking participants to revise their writings

In the second session, participants received direct, indirect or no feedback respectively, depending on the group students were assigned to The first part of the second session was spent on separately instructing each group on what was expected from them The direct feedback and indirect feedback group were asked to copy their text revising all errors the teacher gave feedback on Students in the Indirect group were furthermore instructed on the meaning and used of the error codes in their text if

necessary to make sure they all understand the codes The students in control group were instructed to read over their writing carefully and search for elements in need of revision and then revise them All treatment and control groups were given the same amount of time (35 minutes) to carry out their assignment

Step 6: Administering the subsequent test (post-test)

In the third session, students of all groups were asked to write a paragraph of the same type and topic as the initial task but about a different event The time allowed for

the subsequent task was similar to that for the initial test and revision, 35 minutes

Summary of data collection procedure:

The design of the experiment and division of feedback treatment are outlined below Table 3.2: Summary of data collection procedure

Stage

Group Experimental group Control

group Direct

feedback

Indirect feedback Pre-treatment stage

Administering of pretest to both control and

experimental groups

Treatment stage

Providing direct feedback, indirect feedback and

no feedback for experimental and control group

respectively

Direct feedback

In direct feedback

No feedback

Post-treatment stage

Asking students of experimental and control

group to revise their writing basing on the

feedback they received

Asking students of experimental and control

group to write a new task

3.5 Data analysis

Upon the completion of data collection, quantitative analysis was conducted First, all errors were counted and entered into the comparison tables prepared by the researcher These tables recorded all targeted grammatical errors on student‟s writings

in all the three stages and all three groups The measurement of grammatical accuracy

in the writing tasks was the error rate that was calculated as the percentage of incorrect usage of each targeted grammatical errors (verb tenses and forms, attitudinal adjectives and article usage) out of the obligatory instances For example, five incorrect uses of verb tenses and forms from ten obligatory occasions gave the error rate of 50% The increase or decrease of error rate reflects the decrease or increase of grammatical accuracy respectively

The second step in data analysis was to conduct statistical tests with SPSS statistical software

CHAPTER 4: RESULTS AND DISCUSSION

This chapter concentrates on the presentation, interpretation and discussion of the data In the first part of the chapter, data that help answer each research question will be presented In the next part of the chapter, the findings of the research questions will be gathered and discussed

As explained earlier (see Chapter 3), the experiment had three groups, including two experimental groups, Group 1 (Group DF) receiving direct corrective feedback and Group 2 (Group IF) receiving indirect corrective feedback, and one control group – Group 3 (Group NF) which received no feedback Each group consists

of 15 participants Error rate was measured by calculating the percentage of incorrect usage of each targeted grammatical errors out of the obligatory instances Two types of SPSS analysis were carried out to analyze the data, one-way ANOVA and paired samples t-test at the level of 0.05 significance

4.1 Pretest results

At the beginning of the step of data analysis, a one-way ANOVA was used to test for any initial between-group differences in error rate

4.1.1 Verb errors in the pretest

Table 4.2 - Test of Homogeneity of Variances (Test 1 – Verb)

Levene Statistic df1 df2 Sig

Table 4.1 – Descriptives (Test 1 – Verb)

Upper Bound

DF 15 36.000 15.588 4.025 27.367 44.633 18.000 68.000

IF 15 36.667 13.978 3.609 28.926 44.407 14.000 64.000

NF 15 38.667 15.305 3.952 30.191 47.142 9.000 68.000 Total 45 37.111 14.674 2.187 32.703 41.520 9.000 68.000

Table 4.3 – ANOVA (Test 1 – Verb)

Sum of Squares df Mean Square F Sig

The table 4.3 showed that there was no statistically significant difference

between groups in the pretest because the Sig value was greater than 0.05, ANOVA [F

(2,42) =0.129, p=0.879]

4.1.2 Errors of attitudinal adjectives in the pretest

Table 4.4 – Descriptives (Test 1 – Adj)

Minimum Maximum

Lower Bound

Upper Bound

DF 15 33.267 25.325 6.539 19.242 47.291 00 67.00

IF 15 48.867 24.982 6.450 35.032 62.701 00 100.00

NF 15 44.467 24.345 6.286 30.985 57.949 00 67.00 Total 45 42.200 25.206 3.757 34.627 49.773 00 100.00

Table 4.5 - Test of Homogeneity of Variances (Test 1 – Adj)

Table 4.6 – ANOVA (Test 1 – Adj)

Sum of Squares df Mean Square F Sig

Between Groups 1940.800 2 970.400 1.567 221

Within Groups 26014.400 42 619.390

As far as the errors of attitudinal adjectives are concerned, it can be seen from the table 4.6 that the Sig value was larger than 0.05 Therefore, the error rate of the

groups were not significantly difference, ANOVA [F(2,42)=1.567, p=0.221]

4.1.3 Errors of articles in the pretest

Table 4.7 – Descriptives (Test 1 – Article)

DF 15 58.600 24.448 6.312 45.061 72.139 00 88.00

IF 15 58.600 21.576 5.571 46.651 70.549 25.00 88.00

NF 15 58.667 18.734 4.837 48.292 69.041 25.00 88.00 Total 45 58.622 21.212 3.162 52.249 64.995 00 88.00

The table 4.9 demonstrated no significant difference between groups in terms of

the rate of article errors in the pretest, [F(2,42)=0.000, p=1.000]

In short, results showed that there was no statistically significant difference between groups in terms of the error rate in the pretest Moreover, the task‟s topic proved not to have significant influence on students‟ writing performance Hence, any differences in error rate found later in the study cannot be ascribed to initial difference between treatment groups or task‟s topic factor

Table 4.8 - Test of Homogeneity of Variances (Test 1 – Article)

Table 4.9 – ANOVA (Test 1 – Article)

Sum of Squares

df Mean Square F Sig

Within Groups 19798.533 42 471.394

4.2 Post-test results

4.2.1 Do teacher direct corrective feedback and indirect corrective feedback help students to reduce grammatical errors in writing from an initial task to a new task?

Firstly, the researcher examined the effects of the different feedback types on each grammatical error category (verb tenses and forms, attitudinal adjectives and articles) To answer the first research question, 9 paired-samples t-tests were used These tests were used to compare the percentage of errors of each of the three groups

in the pretest (Test 1) and post-test (Test 2) Differences in the mean error rate between the two tests indicated an increase or decrease in error rate, which reflected the grammatical accuracy performance Details of the results and data analysis are discussed below

4.2.1.1 Verb tenses and forms

Table 4.10 - Paired Samples Statistics (Pair 1 – DF - Verb)

Mean N Std Deviation Std Error Mean

Pair 1 DF_Test1_Verb 36.000 15 15.588 4.025

DF_Test2_Verb 18.667 15 11.580 2.990

In the table 4.11, it can be seen that the probability value Sig.(2-tailed) was 0.000 (less than 0.05) Because this value was substantially smaller than the specified

Table 4.11 - Paired Samples Test (Pair 1 – DF –Verb)

Paired Differences t df Sig

tailed)

(2-Mean Std

Deviation

Std

Error Mean

95% Confidence Interval of the Difference Lower Upper Pair 1 DF_Test1_Verb

DF_Test2_Verb 17.333 14.734 3.804 9.174 25.493 4.556 14 .000

alpha value of 0.05, there was a significant difference in the first and second percentage of verb errors

The table 4.10 reveals that the mean of the first test was 36.000 and the mean of the second test was 18.667 Therefore, it can be concluded that there was a significant decrease in the percentage of verb errors from the pretest to the post-test

Despite the significant decrease in percentage of verb errors showed in the table above, the magnitude of the intervention‟s effect had not been revealed Therefore, the researcher continued to calculate the effect size statistics Eta squared was 0.6 According to Cohen, 1988, with the eta squared value of 0.6 it can be concluded that there was a moderate effect, with a substantial decrease in the percentage of verb errors obtained before and after the intervention

Table 4.12 - Paired Samples Statistics (Pair 2 - IF - Verb)

Mean N Std Deviation Std Error Mean

Table 4.13 - Paired Samples Test (Pair 2 – IF - Verb)

Paired Differences t df Sig

(2-tailed) Mean Std

Deviation

Std

Error Mean

95% Confidence Interval of the Difference Lower Upper Pair 2 IF_Test1_Verb

IF_Test2_Verb 9.733 14.043 3.626 1.957 17.510 2.684 14 .018

A paired-samples t-test was conducted to evaluate the impact of teacher indirect feedback on improving verb errors of students There was a statistically significant

decrease in the percentage of verb errors from Test 1 (M= 36.400, SD=14.029) to Test

2 [(M=26.6667, SD= 12.063), t(14) = 2.648, p<0.05] The eta-squared statistic (0.4)

indicated a small effect size It means that indirect corrective feedback also helped learners to reduce errors related to verb forms and types

Table 4.14 - Paired Samples Statistics (Pair 3 – NF - Verb)

Mean N Std Deviation Std Error Mean

Pair 3 NF_Test1_Verb 38.667 15 15.305 3.952

Table 4.15 - Paired Samples Test (Pair 3 – NF - Verb)

Paired Differences t df Sig

(2-tailed) Mean Std

Deviation

Std Error Mean

95%

Confidence Interval of the Difference Lower Upper Pair 3

NF_Test1_Verb

NF_Test2_Verb

6.133 19.0745 4.925 -4.430 16.696 1.245 14 233

Regarding groups receiving no feedback, although the percentage of verb errors

decrease from Test 1 (M= 38.667, SD=15.305) to Test 2 [M=32.533, SD= 11.096), t(14) = 1.245, p>0.05], the decrease was not significant because the probability value

was larger than specified alpha value of 0.05 In other words, students receiving no corrective feedback could not reduce their verb errors remarkably in the post-test

4.2.1.2 Attitudinal adjectives

For the attitudinal adjectives, as the tables below indicated, groups receiving any kinds of feedback, direct feedback, indirect feedback or no feedback, had a considerable reduction of the error rate in the post-test

Table 4.16 - Paired Samples Statistics (Pair 4 – DF – Adj)

Mean N Std Deviation Std Error Mean