Khóa luận tốt nghiệp: Using ChatGPT in Academic English writing: English-majored students' perceptions and practices at An Giang University

VIETNAM NATIONAL UNIVERSITY HO CHI MINH CITYAN GIANG UNIVERSITY FACULTY OF FOREIGN LANGUAGES \ sam VIET NAM NATIONALUNIVERSITY HO CHI MINH CITY UNDERGRADUATE THESIS USING CHATGPT IN ACAD

VIETNAM NATIONAL UNIVERSITY HO CHI MINH CITY

AN GIANG UNIVERSITY

FACULTY OF FOREIGN LANGUAGES

\ sam VIET NAM

NATIONALUNIVERSITY

HO CHI MINH CITY

UNDERGRADUATE THESIS

USING CHATGPT IN ACADEMIC ENGLISH

WRITING: ENGLISH-MAJORED STUDENTS'

PERCEPTIONS AND PRACTICES AT AN GIANG

UNIVERSITY

HUYNH THI NGOC CHAU

AN GIANG, MAY 2024

VIETNAM NATIONAL UNIVERSITY HO CHI MINH CITY

AN GIANG UNIVERSITY

FACULTY OF FOREIGN LANGUAGES

UNDERGRADUATE THESIS

USING CHATGPT IN ACADEMIC ENGLISH

WRITING: ENGLISH-MAJORED STUDENTS'

PERCEPTIONS AND PRACTICES AT AN GIANG

UNIVERSITY

HUYNH THỊ NGỌC CHAU

ID NO DTA207533

ADVISOR: PHAN THI THANH HUYEN, PhD.

AN GIANG, MAY 2024

Thesis Title: USING CHATGPT IN ACADEMIC ENGLISH WRITING:

ENGLISH-MAJORED STUDENTS' PERCEPTIONS AND PRACTICES AT

AN GIANG UNIVERSITY

By: HUYNH THI NGOC CHAU

Field of Study: EDUCATION

Advisor: PHAN THỊ THANH HUYEN, PHD.

Accepted by the Faculty of Foreign Languages, An Giang

University on May 2024

Examiner 1 Examiner 2

2 C0 000000000000000000000000000000000000000000600 cê000000600000006060000060606060000000600000e0eeeeeeeeeee°e

Advisor

PHAN THI THANH HUYEN, PHD.

Dean of the Faculty of Foreign Languages

I would like to express my deep gratitude and appreciation to all those who have

supported me throughout the entire process of carrying out and writing up thisthesis

First and foremost, I am immensely grateful to my advisor, Dr Phan Thi Thanh Huyen, for her valuable and dedicated guidance It is through her profound feedback and words of encouragement that I found the motivation to

successfully complete this graduation thesis Furthermore, her expertise played

a crucial role in shaping and driving the success of this thesis I consider myselffortunate to have had the opportunity to work under her guidance

I would like to extend my sincere gratitude to the faculty members of the Faculty

of Foreign Languages at AGU, especially Dr Nguyen Thi Dung and Dr Bui Le

Diem Trang, for their valuable contributions and detailed comments on theinitial chapters Their insights and valuable feedback greatly assisted me in the

process of shaping ideas Furthermore, I would like to express my sincerest

gratitude to Ms Lu Thi Thu Nguyen, M.A and Dr Nguyen Tien Ngo for theirinvaluable support during my data collection process Without their assistance,

it would have been challenging for me to successfully complete this research I would also like to extend my thanks to Ms Nguyen Thi Nhat Minh, M.A for

her encouraging words that helped me overcome psychological pressure

Moreover, I would like to extend my heartfelt appreciation to Ms Nguyen Thi

Hai Yen for her invaluable support

I would not forget to extend my heartfelt appreciation to my family and friends

such as Huynh Nhat Hao, Nguyen Thi Nguyen Kim, Vo Thi Hong Nghi, and

Pham Thi Tuyet Nhu, who have been there consistently, supporting and readilyassisting me throughout the entire process Their words of encouragementprovided strength during the most challenging times Additionally, I would like

to express my deepest gratitude to my classmates from DH21TA1 andDH21TA2 for providing valuable insights and expanding my understanding of

the research topic.

In particular, I would like to express my gratitude to the research participantsbecause their cooperation and willingness to share their experiences played asignificant role in completing this research

Above all, I want to acknowledge the support and resources provided by thestaff at the Library of An Giang University It is through their dedicatedassistance that I was able to access relevant research materials and resources

I hereby declare that the thesis entitled “Using ChatGPT in academic English

writing: English-majored students’ perceptions and practices at An GiangUniversity” represents my own work and that has not been previously submitted

to any other university or institute in application for admission to degrees orother qualifications

May 2024

Huynh Thi Ngoc Chau

ChatGPT based on artificial intelligence (AI) and developed by OpenAI is currently becoming popular in education after being widely accepted in many fields Students can use this technology to hone their academic writing skills.

The benefits of ChatGPT include translating text, providing vocabulary lists,improving sentence structures, and editing articles In other words, personalizedlearning opportunities are developed While academic English writing is one ofthe mandatory subjects in the English curriculum at the Faculty of ForeignLanguages, An Giang University (AGU) and their students are familiar with

ChatGPT, they may still confront difficulties creating academic writing in

English for different reasons Responding to this gap, this present research aims

to understand how English-majored students at AGU perceive and practiceChatGPT in academic English writing Deploying a mixed-method approach,

this study collected data from 176 EFL students at AGU Subsequently, six

students were chosen from the participants for the second stage of the research,

the individual interview The findings suggest that English-majored students

have positive perceptions and practices of using ChatGPT in academic Englishwriting English-majored students find ChatGPT to be a valuable tool in theiracademic English writing Despite facing difficulties such as limited

vocabulary, constructing sentence structures, time constraints, and fear of making mistakes in Two-part question Essays and Argumentative Essays, they have embraced ChatGPT as a supportive learning aid alongside teacher

feedback and reference materials They have recognized that ChatGPT canassist them in generating ideas, checking vocabulary and grammar, developingstructure, and expanding their vocabulary Consequently, English-majoredstudents have utilized ChatGPT to offer feedback, utilize templates, and receive

suggestions for essay structure and vocabulary This study contributes to

exploring the importance of using ChatGPT in education by studying the

common difficulties in academic English writing, as well as the perceptions and

practices among English-majored students regarding the use of ChatGPT foracademic English writing purposes

Key Words: AI, ChatGPT, academic English writing, perceptions, practices,EFL students

CHAPTER 1: INTRODUCTION oc eeeeeceeeseeseeeeeseeeeeeseesesseeseeseeseeaeeaeeaeeaes 1

1.1 Background of the SŠtudyy - c1 1v vn vn rưn 11.2 The rationale of the researCH . + + + k*3E*EEeEeekkeesrkrrerreree 21.3 Research Objectives ceeccesccescecesecssceceeceseeeseeceseceseeceaeceneeesaeceeessaeeeaes 31.4 Research Questions 1n 31.5 The significance of the research 0 ecceecceeeeseeseeeeceseeececeeeceseeneeeseeeees 4

1.6 Structure of the Thesis - 11 9199 91 11 31 ng ng nệc 4

CHAPTER 2: LITERATURE REVIEW SÁT ri, 6

2.1 Definitions 1 6

"ng 6

"A o c0 72.1.3 Writing in Ø€TIGTAÌ G1111 91H ng ng 82.1.4 Academic English WTIfITIE - -5 s6 3 1 231191123111 9 ren 8

2.2 Characteristics of good academic English writing - 9

2.3 Difficulties in writing academic English «+55 «<< +ssss+ 11

PIN No vui n /-ä 11

“tru nan 112.3.3 MOTIVAfIOH Gv HT HT TH hà 12

2.3.5 Paragraph Organization 000 122.3.6 ôu ca 12

2.4 The benefits of ChatGPT in supporting academic English writing 13

2.5 Related SfUÌ1€S - Gv ng gu ng ng nà 15

2.5.1 ¡ái 0/001 15

2.6 ConcÏUSIOTI - 2220203011111 111 1111111 999355 011k k k5 cà 17CHAPTER 3: RESEARCH METHODOLOGY 5 +5 << ++<+x<e2 18

3.1 Research n9 18

3.2 Population, Sampling and PartICIpanS - - 5+5 +++ss+sx+sssssss 19

3.2.1 Questionnaire participants - c5 3 13233 eesereserres 203.2.2 Semi-structured interview parfICIDATẨS 5c ssscs+ s2 213.3 Data collection instruments 5 vn ng, 22

3.3.1 QueStiONMAILE 223.3.2 Semi-structured intervView cscesesecssecsseesecesesseceseceeeeseeeeeneesaeeas 253.4 Data collection procedures c5 222 3311321 31EEErsersre 26

E100) 0i 7n 26

3.4.2 Semi-structured ITIf€TVICW sọ nnHnHnnnHngnggkt 263.5 Data nh 263.6 Validity and reliability of the sfudy 5-ccS- + ssekrserseerrrrerre 273.7 Ethical Considerations 2.0.0 eeesseeseesseesecsecssesseessesseessesseesseeseesseesaeeaes 28

CHAPTER 4: FINDINGS AND DISCUSSION .c.cccccĂ2 30

4.1 Difficulties in academic English WrItIng - 5+ s-ss+s++sssessssz 30

4.1.1 Types of academic English essays that EFL students struggled with

— A 30

4.1.2 Students’ specific difficulties in academic English writing 32

4.1.3 Seeking support from VarIOUS SOUTC€S - 5 5-55 +sess+ 35

4.2 English-majored students’ perceptions of using ChatGPT in academic

English Writing .cecccesscsesecssecesseceecesceceneceeeceseceeeseaeceaeeceseceeeeeaeceeensaes 38

4.3 English-majored students’ practices of using ChatGPT in academic

English Writing eececcescescceseceeeseceseesceesecceeseceeeseceaeceeesseeecesseceaeeseenaeess 41

4.4 DiSCUSSION nraỒỮ - 44

4.4.1 Difficulties in academic English wrIting - s55 s<<<<<5+ 44

4.4.2 English-majored students’ perceptions of using ChatGPT in academic English Writing :cccccscessesseescesseeseeseessecseeeseeseeeseeseeeseenseens 45

4.4.3 English-majored students’ practices of using ChatGPT in academic

Ji134010805461171-20077777 47

CHAPTER 5: CONCLUSIƠNS -.-Ăn LH HH HH Hư, 48

5.1 Summary of the R€SuÏS ó5 2c 3311821113911 rrrk 48

5.1.2 Perceptions 17 49Bốc e- 495.2 Implication of the ResearCh - + 25 2s vs ng ng 50

ho 515.4 Recommendations ce eeessessccsseeseesecscesseecesseeseesecssesseesseceeeseeaeeaes 52

h9) 52

REFERENCES 0117 54

APPENDICES 271777 66

APPENDIX I1 - QuestionnaIre - c6 111332111133 5E 1Exxxee 66

Appendix 1.1 - Questionnaire (English version) -.«««<< s<+<s<2 66

Appendix 1.2 — Questionnaire (Vietnamese VerSiOfI) ‹ -< << 5+ 72

APPENDIX 2 — Ethics for ITẨCTVICW - Gv ng nưkt 77

Appendix 2.1 - Information sheets for student participants (English

Appendix 3.1 - Interview questions (Vietnamese VerSiO)) -‹- 81Appendix 3.2 - Interview questions (English ver:si0fi) - « 82APPENDIX 4 — A sample of one interview — Auto transcrIpt 83

Appendix 4.1 — A sample of one interview — Auto transcript (Vietnamese VEOVSION) esccesscsesccesscessccessesscecesscssceessscsseeseeceseeeseeceseceseesesecsseeeseceseeseaeenseees 83

Appendix 4.2 — A sample of one interview — Audio transcript (EnglishVEVSTON) sccecccsccessesccesecscceseesccesecesecsecesececesecsceesececeeseceseceeeseceeeeseceeeeaeeeaeens 86

LIST OE THE TABLES

Table 1 Description of the participants -5 5 55+ S3 s ssseersererre 20

Table 2 Students’ information related to the Writing course 21

Table 3 Description of the participants for the interview - ‹ - 22Table 4 Questionnaire Structure -c5 SG 3113311135511 Eerrske 23

Table 5 Questionnaire structure about the difficulties that English-majored

students face in academic English WTIfITE s5 5 + s+ksseeeseeessere 24

Table 6 Questionnaire structure about students’ perceptions of using ChatGPT

in academic English Writing cccceceesscecceseeeeeeceeceeseceseeseeeseceeesecneeeeeneeeaes 24

Table 7 Questionnaire structure about students’ practices of using ChatGPT

for academic English writing cece ssceseeseeseeseeeessesscesceeeeeceeeeseseeseeens 24

Table 8 Reliability Statistics of the pilot study about English-majored

students’ perceptions and practices of using ChatGPT in academic English

WING oo 25

Table 9 Question structure for semi-structured 1nterVIews - ‹ - 26

Table 10 Summary of the data collection and analyS1s -« -««+ 27

Table 11 Reliability values of Items related to English-majored students’

perceptions and practices of using ChatGPT in academic English writing 28

Table 12 Types of academic essays in English that EFL students struggled

With 1 Witting eee = 30

Table 13 Difficulties impacting the quality of English writing by EFL students

— 32

Table 14 The sources that English-maJored students sought help from to

improve their academic English writing skiÏÌS - «<+<<<<c<«+ 35

Table 15 Ways in which EFL students perceive that ChatGPT has supported

academic English WTITIT - < + 11911 ng HH ng kg 38Table 16 English-majored students use ChatGPT for its helpful features in

academic English wiiting :ccccccscssscsseesseeseceseeseeeseeseceseeeeeseeeensecnseeeeneeeats 42

LISTS OF ABBREVIATIONS

AGU An Giang University

AI Artificial intelligence

ChatGPT | Chat Generative Pre-Trained Transformer

EFL English as a foreign language

EL English Language

ELTE English Language Teacher Education

FFL The Faculty of Foreign Languages

Ll First Language

MBA Master of Business Administration

SPSS Statistical Package for the Social Sciences

CHAPTER 1: INTRODUCTION

This chapter has six parts: (1) research context; (2) theoretical basis

of the research; (3) research objectives, (4) research questions; (5) significance

of the research; and (6) organization of the thesis

1.1 Background of the Study

Since its launch by OpenAI on November 30, 2022 (Devi et al., 2023),ChatGPT has become the fastest-growing online tool, gaining over one millionusers in just five days The emergence of ChatGPT and stories related to

chatbots impacted globally (Taecharungroj, 2023), to the extent that ChatGPT

is arguably the most powerful chatbot tool In other words, it has influenced on

many areas of life, especially education (Adeshola & Adepoju, 2023) The

benefit of ChatGPT is that it accurately answers questions with simple

operations However, that is why many teachers worry that students may takeadvantage of ChatGPT to cheat in-class activities and tests Recently, one of thefaculties in France banned students from using ChatGPT and AI-based tools(von Garrel & Mayer, 2023) and warned that anyone found using ChatGPT

“without transparent reference” to create written or spoken work, could be

expelled Besides, Ethan Mollick, a professor at the University of Pennsylvania’s Wharton School of Business, believes that AI applications like

ChatGPT will ruin homework because students can copy the content thatChatGPT provides Nevertheless, he also believed that ChatGPT is a useful tool

to help teachers prepare textbooks and exercises to support teaching MBA

students at school (Rahman & Watanobe, 2023).

Many educators are concerned that their students may use ChatGPT to

do their homework (Ampe et al., 2023) Learners already use many othertechnologies to check grammar and spelling, convert language to text, and findanswers for both classwork and homework Additionally, adopting these newtechnologies could affect education by changing how learning skills areassessed; for example, schools can use it to support students with academicwriting tests during exams (Kirby, 2007)

In short, we are moving towards a world where humans and AI co-exist,

in which AI will be a useful tool, especially for research and student learning

In Vietnam, ChatGPT is being tested by many people, with many mixedcompliments and criticisms Some teachers have “tested” the capabilities of this

AI with different concerning topics ranging from social knowledge to creating

content for lectures and text translation (Gimpel et al., 2023) From initial tests,

outstanding In particular, the attribute that makes ChatGPT highly appreciated

is its ability to process, select and update information

ChatGPT can be a great assistant in office work when employees’ reports

using it take less time and the quality of the writing is also better However, many cases show that ChatGPT’s response is not accurate because the AI chatbot only processes and presents text according to the data that has been

provided (Haleem et al., 2022) Therefore, we need to review and edit generated documents Although compiling content is often complex andrequires specific in-depth knowledge, rather than relying on it entirely, people

AI-must filter it.

One of the biggest concerns involves - the issue I mentioned earlier

regarding students who may deceive teachers by asking AI to write their essays(Pham et al., 2023) Many people advertise online about ChatGPT writingtheses and dissertations for students for a fee, thus threatening the essence ofstudent life by discouraging them from actively researching in favor of relying

on overusing this AI tool for dissertations Besides, ChatGPT’s

language-switching ability can demotivate learning a language or writing documents

because everything becomes so easy at face value with its help Therefore, the

most important issue to consider before students use this tool is that they mustclearly define their purposes of using ChatGPT to remain independent of it andthus prevent losing the ability to think scientifically and critically, and bemotivated to learn

In summary, I have discussed how AI can be a tool to improve studentlearning through being able to answer questions accurately and diversely Fromthere, students can expand their knowledge through ChatGPT’s answers

Currently, ChatGPT is considered to benefit students widely at universities

globally, including Vietnam Therefore, raising awareness about using ChatGPTfor education is necessary and deserves special attention

1.2 The rationale of the research

As a final-year student majoring in English Linguistics and Literature atAGU, I have noticed that academic English writing skill is one of the fouressential skills for English language learners (Bridgeman & Carlson, 1983)

However, English-majored students often face challenges in academic writing

in English The contributing factors include limited vocabulary and grammar,

difficulties in essay structure, and the ability to distinguish between different

writing styles used in British English versus American English or othercountries where English is the official language (Bailey, 2014) In addition, theterm ChatGPT has become popular among English-majored students at AGU

However, the use of ChatGPT as a learning supporting tool remains limited.

There are two main reasons Firstly, students feel that using ChatGPT for

academic English writing purposes is unethical (Huallpa, 2023) The

Vietnamese tradition of self-reliance and diligence in traditional learning

methods makes students feel that using technology to generate ideas or find outrelated topics is not appropriate These actions are considered "brain drain" or

"plagiarism" (Tran, 2013) Secondly, students do not have enough knowledge

about the benefits and limitations of using ChatGPT This is because

English-majored students at AGU have not been provided with any workshops or sessions on using ChatGPT Therefore, the benefits and limitations of ChatGPT

in supporting academic English writing need to be widely and professionally

disseminated This may allow students to gain perspective and practice using

ChatGPT as an effective learning supporting tool for academic English writing(Lin et al., 2023)

Up to the present time, there has been no official research conducted toinvestigate the benefits and limitations of using ChatGPT for academic writing

in English among students at the Faculty of Foreign Languages, AGU Thisstudy aims to address this gap The objective of the research is to explore the

perceptions and practices of English-majored students regarding the application

of ChatGPT in their academic English writing

1.3 Research Objectives

Overall objective: To explore English-majored students’ perceptions

and practices of using ChatGPT in academic English writing at AGU

Specific objective 1: To identify the major difficulties encountered by majored students in academic English writing at AGU.

English-Specific objective 2: To examine the perceptions of English-majored students regarding the benefits and drawbacks of using ChatGPT in academic English

writing at AGU

Specific objective 3: To describe the various ways in which English-majoredstudents utilize ChatGPT as a tool in academic English writing at AGU

1.4 Research Questions

The study aims to answer the following questions:

Research question 1: What specific challenges do English-majored students face

when it comes to academic English writing at AGU?

Research question 2: How do English-majored students perceive the advantages

and disadvantages of employing ChatGPT in their academic English writing at

Research question 3: In what specific ways do English-majored students

integrate ChatGPT into their academic English writing process at AGU?

1.5 The significance of the research

The findings of this study provide insights into the benefits and

limitations of using ChatGPT for academic English writing among both students and lecturers at the Faculty of Foreign Languages at AGU.

For students, the findings from this study are beneficial to Englishlanguage learning of EFL students Using ChatGPT for academic writing

purposes will be one of the best ways to save time searching and selecting

information This means that students can use ChatGPT to access the correct

answer to questions about vocabulary, grammar, and sentence structure within

articles from related topics From there, ChatGPT helps improve the quality andefficiency of learning academic English writing and subjects related to English-majored students’ writing Understanding ChatGPT will benefit students if they

would like to work at organizations related to technology and education.

Currently, the support of AI applications within education, especially regardingforeign languages, is becoming popular

For lecturers, this research creates the foundation for modern teaching

methods to develop and expand within AGU The reason is that learning methods at the university differ greatly from those at high school At university, since teachers do not check students’ work daily, they need to create the habit

of self-study and proactively search for information This habit helps students

to collect and absorb knowledge actively, leaving teachers to spend extra time

to share more important knowledge Therefore, classroom learning time wouldbecome more efficient as instructors could no longer need to presentinformation that their students can easily find

1.6 Structure of the Thesis

This graduate thesis includes five chapters

Chapter 1 provides information about the research context, including

background, rationale, objectives, research questions, research significance and

organization of this thesis

Chapter 2 includes an introduction to the theoretical and empiricalstudies related to the study as well as a review of some previous literature onthe same topic

Chapter 3 explains the methodology used in this study Research design,

number of participants, research tools, data collection, and data analysis

procedures

Chapter 4 presents the research results along with some discussions.

Chapter 5 summarizes the research results, some limitations, andrecommendations for future research

CHAPTER 2: LITERATURE REVIEW

This chapter includes four parts comprising, 1) definitions of artificialintelligence (AI) - typically ChatGPT, writing in general, and academic Englishwriting; 2) descriptions of the characteristics of good academic English

writing; 3) an outline of the difficulties of academic English writing; 4) a discussion about the benefits of ChatGPT in supporting academic English

writing for EFL students; 5) Related Studies; and 6) Conclusion

2.1 Definitions

2.1.1 AI

AI or Artificial Intelligence is an interdisciplinary field of computer

science that focuses on creating systems that can perform tasks that humans value and need, especially to learn and solve problems (Sarker, 2022).

With the 21st century being “the era of digital technology” (Xu, 2015),

AI techniques have resurged following concurrent advances in computingpower, thus becoming an essential part of the technology industry in solvingproblems Whereas many challenging problems face machine learning, softwareengineering, and operations research, AI is being applied in almost all areas of

education, work, and daily life The encroachment of digital technology, in addition to such challenges, also opens up many new opportunities for people.

During the summer of 1952 McCarthy worked at The Bell TelephoneLaboratories and requested Claude Shannon to collaborate with him to edit a

collection of papers on machine intelligence (Rajaraman, 2014) The collection was a potpourri of papers on diverse topics such as mathematical logic, neural

nets and game playing It was published in 1956 by the Princeton UniversityPress (Goodstein, 1958) and is probably the first publication in the fledglingfield of artificial intelligence (the term was coined later) Therefore, theseresearchers were the first to explore the idea of creating machines that can think

and learn like humans In 1965, at Dartmouth College, McCarthy conducted research, development, and popularization of "machine intelligence", later referred to as "Artificial Intelligence" (McCorduck & Cfe, 2004) and thus

marked the beginning of its golden age (Jiang et al., 2022)

In recent years, while AI has developed and included different aspects,

it maintains a close relationship with information provision systems, such as

machine learning, deep learning, natural language processing, and robotics The

notable development of deep learning algorithms is considered a breakthrough

in AI and is applied to many applications supporting image and speechrecognition, and natural language processing These algorithms increased the

authenticity and reliability of applications and revolutionized the field of AI (Xu

ChatGPT is a natural language processing technology that uses AI to respond to

queries (Devi et al., 2023) As a result, it creates information more

conversationally, picks up knowledge from those talks, and can then provideever more specifically customized replies ChatGPT behaves like a person whilegiving instructions to users and providing information This technology canexecute various activities, including creating poetry, coding, answeringenquiries, writing emails and essays, and translating documents (Fitria, 2023)

ChatGPT differs from other chatbots in that it can respond instantly, thus producing more varied and lively discussions on almost all topics (Javaid,

2023)

ChatGPT is a tool to help improve students’ skills in ways, such as

enabling uses to ask questions accurately, so they may increase their knowledge

through interpreting answers, confidence levels and the quality of answers, and

filter information (Fitria, 2023) This technology suits various applications since

it can adjust to different conditions and situations Moreover, ChatGPT isflexible in that it may be used in various natural language processingapplications (Meyer et al., 2023) It can respond to instructions with highaccuracy and fluency, but needs a thorough understanding of the world and theability to think like a person (Huallpa, 2023)

ChatGPT is an AI-based tool for having exciting and genuine talks withpeople (Rathore, 2023) Consequently, it can have discussions that compare

better to those between people For example, ChatGPT is valuable for

businesses in the customer service and marketing ChatGPT functions as avirtual assistant that can converse with us and respond to our inquiries like anactual person (Maheshwari et al., 2023) Therefore, it uses deep learningalgorithms to generate human-machine natural language discussions

ChatGPT is designed to understand and generate human-like text based

on the input it receives (Bhattacharjee et al., 2024) It can be used for a variety

of natural language processing tasks, including answering questions, generating

text, and engaging in conversation (ChatGPT, January 5, 2024).

ChatGPT is pre-trained on a large corpus of text data, such as paper or

online articles, websites, books and uses supervised spatial techniques

(Shur-Ofry, 2023) The pre-training involves preparing a transformer-based neuralnetwork to predict the suitable expression within context

In conclusion, with the emergence and continual upgrading of AI, language models such as ChatGPT 2, 3, 3.5, and 4 of OpenAI (De Winter et al.,

2023) have attracted strong interest from researchers, especially in the field ofeducation, because they benefit academics, especially the high-volumegeneration feature (Mallik & Gangopadhyay, 2023) This occurs becausecoherent texts are indistinguishable from human literature (Van Heerden & Bas,2021)

2.1.3 Writing in general

Writing has been with humans for thousands of years, and today ismore important than ever, although the explosion of the Internet has somewhatchanged the way people share messages (Harper, 2010) Writing text is complexrequiring a series of mental activities to be coordinated (Olive, 2014) The writer

must delineate the nature, purpose, and communication function of the text, and

accurately represent the reader’s characteristics and expectations to

systematically predict what must or could be written (McKenna & Zucker,

2014) The writer has equal control over the text topic to generate or identifythe most relevant ideas that will gradually constitute the body of the text(Hayles, 2010) Additionally, they sometimes have to clarify messages,reorganize, revise, and present ideas, as well as control the coherence of theentire text (Alamargot & Chanquoy, 2001)

Writing is an effortful and complex activity and to be able to handle themany limitations in writing, the organization of cognitive activities is a basicneed for the writers (Kieft, 2006) Therefore, the students need to be supported

in handling the necessities of academic writing (Akcaoglu, 2011) However,

graduate students have few opportunities to take courses that will transition

them from writing a term paper to writing an academic paper (Austin, 2002).Moreover, these few opportunities sometimes cause concern since seemingly a

wide gap exists between what is required from graduate students and what is taught during writing classes (Restaino, 2012).

2.1.4 Academic English writing

Evidence from academic English writing has always provided themeans of assessing English language skills (Wang & Bakken, 2004) This is

because it requires the writer to demonstrate knowledge and proficiency in the

English language by combining certain skills, such as thinking, interpreting, and

To be successful in academic English writing, the student must be fully aware

of what the lecturer expects to enable her to complete that particular writing task

(Geisler, 2013) Academic English writing is a genre of writing used at higher

levels of education, typically from university and beyond, and is employed in

research and scholarly fields (Paltridge & Starfield, 2016) The purpose of usingthis academic style is to communicate and explain complex research

issues or topics (Gea-Valor et al., 2014) It is commonly found in evaluations,

essays, research papers, and theses within English majors or advanced degree

programs Consequently, academic writing in English requires authors to organize and structure English sentences coherently within a paragraph to articulate their ideas effectively (Mwangi, 2017).

Academic English writing aims to convey information in the Englishlanguage objectively (Day, 2023), to present arguments based on the evidence

being considered, but not according to author’s biases or opinions That means

all claims must be supported by relevant evidence leaving writers bound tofollow the writing process comprising many specific conventions of content,structure, and style (Flower & Hayes, 1981) For this reason, choosing accuratevocabulary and avoiding using informal language, including slang, acronyms,cliches, and conversational phrases are important (Muhartoyo & Wijaya, 2014)

2.2 Characteristics of good academic English writing

Well-written academic English writing encompasses several importantaspects that contribute to the effective communication of research (Dafouz,2020) It should possess a clear structure, with coherent transitional phrases,sentence patterns for presenting and expressing viewpoints, providing focusedinformation, and specialized vocabulary (Howarth, 2013)

One crucial aspect of being well-organized is to ensure that the writingfollows a clear structure comprising an introduction, body, and conclusion(Lundstrom & Baker, 2009) The progression of ideas and paragraphing should

be logical and coherent, with each paragraph connected to the overall argument

and supported by relevant examples It is also important to present variousviewpoints and interpret the existing literature by comparing and contrastingdifferent authors’ arguments (McKenna & Zucker, 2014)

Greetham (2001) emphasizes the significance of clarity and coherence

in the introduction, paragraphs, and conclusion The introduction serves as a

vital component, as it reveals the writer’s interpretation of the research topic and/or question and outlines the roadmap for the paper The paragraphs in

academic writing should align with the introduced roadmap Greetham suggests

including a topic sentence at the beginning of each paragraph to provide a briefpreview of its content (McKenna & Zucker, 2014).

Another important characteristic of successful academic English writing is relevance to the research topic, as highlighted by Davies (2008) Writers must exercise caution when selecting information to include in their

research, as not all interesting information may be pertinent to the researchquestion at hand When reviewing the literature, it is important to compare andcontrast different ideas and perspectives (Van Laar et al., 2017) Additionally,online resources can be valuable in providing precise definitions of keywordssuch as analyze, define, criticize, discuss, describe, explain, justify, andillustrate (Eysenbach, 2009)

The final characteristic of good academic writing is accuracy, whichencompasses high standards of grammar, word choice, vocabulary, spelling, andpunctuation (Polio, 1997) Paying attention to these writing sub-skills, includingpunctuation, is crucial Brown and Hood (1998) provide examples, exercises,

and strategies to practice in order to avoid common mistakes that can lead to confusion For instance, the placement of a comma can significantly alter the

meaning of a sentence

In summary, good academic English writing involves clear expression

of research, precision, and effective organization It requires a well-structured paper with a clear introduction, body, and conclusion The ideas and paragraphs

should be logically connected, supported with examples, and incorporatemultiple viewpoints The writing should be relevant to the research topic,accurately interpret the literature, and compare and contrast different arguments.Additionally, accuracy in grammar, word choice, vocabulary, spelling, and

punctuation is essential For example, the use of comma in the following

statements changes the meaning as shown below:

1 The passengers, who traveled early, were killed in the accident (It means allpassengers)

2 The passengers who traveled early were killed in the accident (It means only some passengers) (Altakhaineh, 2008).

Jordan (1986) also provides many examples and exercises that reveal the

importance of grammar, vocabulary, spelling, and punctuation becausemisusing tenses changes meaning and spelling mistakes confuse (Singh et al.,2017) As well, Jordan (1986) distinguishes the differences between some verbs

32 66 33 66

that lead to confusion such as “lend” and “borrow”, “rise” and “arise”, “raise”

99 66

and “increase”, “make” and “do”, and “say” and “tell” as a part of taking care

of word choice and the use of words

In short, to write academic English well, the writer needs to apply the characteristics (Akkaya & Aydin, 2018) discussed so far in this section, that is, organization, relevance to questions, and accuracy, as well as using words and

syntax carefully and precisely to avoid confusing or misleading readers

2.3 Difficulties in writing academic English

Being aware of common problems encountered in academic writing is important to developing writing skills and producing a complete article Alfaki

(2015) believes that students often encounter problems with syntactic functions,

such as grammar, sentence structure, expression, and mechanics Accordingly,

Al Fadda (2012) supports the view that combining words into sentences inwritten text is a common difficulty that learners encounter They believe thatthey can translate a written sentence into the target language easily as long asthey can produce a complete sentence in their mother tongue (L1) However,

the difficulty for learners in mastering writing skills arises when the perceiving

and understanding academic factors differ in each country (Kudryavtseva,

2015) Rahmat (2020) classified six challenges in academic English writing,

including vocabulary, grammar, motivation, time, paragraph organization, and

coherence

2.3.1 Vocabulary

English has the largest vocabulary of any world language, to the extentthat educated native speakers are likely to know about 70,000 words (Nation &Anthony, 2016) This leaves ESL learners with a large challenge (Callahan,2013) Furthermore, in English, some synonyms are sometimes used within

different contexts and are interchanged to avoid repeating words (Phoocharoensil, 2010) However, if some words are replaced in the same

context, they will change the content of the sentence Therefore, when studentsare learning academic writing in English, they often err when seeking the wordssuitable for each context (Bailey, 2014) In addition, if students lack vocabulary,they will not know what words they need to use, and so will tend to writesentences with incorrect structures and thus struggle to make sense (Qamariah

& Aini, 2021)

2.3.2 Grammar

According to Hajana (2006), the best way to improve students’grammar is to use their writing for analysis and discussion of common errors.The starting point for students’ progress would come from producing a new text

learners should dedicate ample time to develop a deeper understanding ratherthan viewing grammar as a complex issue Grammar is divided into tenses,subject-verb agreement, syntax, and various components of speech to determine

Managing time proficiently is important to education’s teaching and

learning in general (Graham, 2020) However, meeting the deadlines for

studying sufficiently for exams is commonly stressful for students when time

runs short during their preparation and they lose confidence and becomeconfused Therefore, students should practice more, so they adequate controltheir time In other words, students should spend more time on writing

instruction and devote at least double the amount of time they expected.

2.3.5 Paragraph organization

Because organizing paragraphs in writing consists of much more than

a random combination of sentences; they need to be linked logically and

coherently A typical paragraph includes a topic sentence, supporting evidenceand details, an analysis, and a conclusion (McKenna et al., 2007) However,many students face difficulty in their writing when they do not know how toarrange paragraphs so that the content is presented clearly and convincingly.After all, a well-organized paragraph will help readers better understand the

ideas presented (Rahman et al., 2022) Then again, many students can be

unfamiliar with the order of sentences in their paragraphs, thus affecting thestructure not only of paragraphs but also of the entire essay if the reader cannotgrasp what the writer intended to convey

2.3.6 Coherence

RahmtAllah (2020) said that coherence is extremely necessary inparagraphs so readers understand the structure and meaning of the entireparagraph because the ideas are organized Besides coherence, writing also

needs cohesion, which refers to how well sentences and paragraphs are linked

to the words that precede and follow them (McKenna et al., 2007)

In short, every student faces a specific problem for a specific reason when it comes to academic writing The problems listed above are the most

typical problems that students often encounter They affect not only theiracademic English writing abilities but also their academic performance of, andmotivation to write English.

2.4 The benefits of ChatGPT in supporting academic English writing

ChatGPT supports students in maximizing their academic Englishwriting abilities by providing highly accurate sample articles, quick evaluations,

outlining assistance, idea generation, summarizing materials, and essay writing

providing information that helped researchers Zhai (2022) concludes by

suggesting that students need to be able to use AI tools to perform professional

tasks while the education sector needs to create conditions for students to

improve critical thinking skills To achieve that, teachers need to design learning

content related to AI (Yang, 2022) to create conditions wherein students practice

critical thinking in solving real-life problems Therefore, users need to focus on

creativity and different thinking so that ChatGPT maximizes its capabilities and users do not lose their inherent abilities Thanks to its ability to synthesize

information, ChatGPT helps to give a complete answer to a question instead ofjust providing results (Hulman et al., 2023), or numbers for users to synthesizethemselves like previous search engines In short, users can use it to assist incompleting all steps of the academic writing process, from suggesting topics towriting articles, to finding resources, and even editing the writing

Secondly, ChatGPT offers students a swift means to self-assess theirwriting, bypassing the need for immediate teacher feedback, thus fostering moreeffective learning and academic achievement (Elbana & Armstrong, 2024) This

capability enables scalable content personalization, previously a time-intensive

process, and has diverse applications in conversational chatbots, automation,and data analytics Some educators are exploring ChatGPT as an adjunct toolfor self-directed study in academic English writing, empowering students toindependently evaluate their essays without instructor input (Javaid et al.,

2023) Students can then use the AI to request specific edits or revisions,

enhancing both efficiency and learning outcomes Additionally, in creativewriting, ChatGPT functions as an efficient word organizer, optimizing students’time while maintaining learning effectiveness

Thirdly, ChatGPT helps establish an outline, a suitable structure, and

vocabulary at the required level Once a topic is available, the writer needs to

create an outline for the article before starting to write This is the stage where

writers typically engage in brainstorming to generate ideas, carefully selecting

and organizing them in a manner that reflects their topic It is a process thatdemands extensive reading and a solid grasp of the subject matter Therefore,

creating an outline refers to creating the “backbone” for the article (Sinha et al.,

2015), which is an extremely important step but challenging, leaving many

students tentative about overcoming it However, thanks to ChatGPT, the answers are given in just a few seconds ChatGPT can fully provide the main

parts of an article's outline From there, the writer can choose all or some of the

main parts to apply to their article Currently, the literature research is exploring

how ChatGPT supports students’ writing abilities For example, how doesChatGPT support writing with all the steps from preparing an outline, providing

suitable structure and vocabulary for each specific level, editing content, and

making suitable suggestions (Su, 2023) Yan (2023) also seeks to ascertain the

purpose of understanding and evaluating EFL students’ behavior when exposed

to ChatGPT while learning writing in class Whereas the results show thatChatGPT successfully guides and supports EFL students in writing, it is not

fully supported nor completely replaces their writing Besides, ChatGPT can

provide EFL students with training programs that provide reasonable errorcorrection suggestions (Young & Shishido, 2023) Thanks to the ability tosynthesize information, ChatGPT helps provide a complete answer to a questioninstead of just providing results and numbers for users to synthesize themselveslike previous search engines Consequently, users can apply it to complete all

steps of academic writing, from making topic suggestions to writing papers,

finding materials, or even helping with editing

Fourthly, ChatGPT helps provide ideas about the topic in that, based

on their teacher’s requests, students can give a command for ChatGPT to

suggest suitable topics (Rahma & Watanobe, 2023) The more detailed the

description in the command, the more accurately the tool will provide results

that meet the specified requirements For example, with the requirement to write

an introduction about a career in the field of information technology, studentscan add criteria such as “related to security”; “professions currently sought after

by companies” From there, the results will display the best options

Fifthly, ChatGPT helps search for documents, synthesize andsummarize them This is an extremely important step in research in general andacademic writing in particular With traditional search engines like Google,when writers type in keywords, related results will be provided The next step

is to read, synthesize, and select relevant information from accepted official

sources ChatGPT with just the command “please provide documents related to

the topic X”, the tool will provide sources along with the Abstract of those

documents However, notably, ChatGPT’s data source covers all from 2021 In

academic English writing, the skills of “paraphrasing” - rewriting sentencesusing different vocabulary and structures while still retaining the meaning of the

original sentence - and summarizing are two important skills that help writers

analyze and synthesize information (Ruslan et al., 2020) This is also the skill

that most students are most “afraid” to do Since the advent of AI tools like

ChatGPT, academic writing has become easier by simply providing the originaltext and giving pertinent commands For example, a writer can ask, “Summarizethe following paragraph in 150 words” to get the desired result within seconds

Sixthly, ChatGPT helps students write a sample essay Students or

teachers can consider asking ChatGPT to provide a sample essay on a certain

topic, with specific requirements for the number of words and allowedparagraphs Mohammad (2023) researched the authenticity of sample essayswritten by ChatGPT using Turnitin, an anti-plagiarism tool The results showthat most sample papers so written are considered to be less similar to previous

papers This means they are not considered plagiarism if teachers use software

to scan for it and thus will challenge teachers who ask students to write essays

in which 24 students were randomly selected from universities in Karachi for individual interviews The sample size was small but diverse, including final-

year students from the sciences and humanities The data were divided into four

sections to display varying levels of feedback, ranging from positive to negative.

The results indicated that students believed ChatGPT would be beneficial foracademic English writing if it were used under appropriately supervised or

controlled conditions.

Tossell and Tenhundfeld (2024) study students’ perceptions of usingChatGPT in writing academic texts in essay format Using a pre-post studydesign, they collected data from 24 participants to evaluate how well ChatGPT’s

assisted students to complete assignments, explore the educational value and

impact of the study, and its impact on learning Their quantitative and thematicanalyses transformed students’ learning and yielded mixed feedback The

results show that participants found ChatGPT valuable for learning Being

comfortable with its ethical and benevolent aspects, they increasingly used it

Mahyoob and Alblwi (2023) developed a proposed framework forhuman-like language processing of academic text by ChatGPT This studytested the ability of ChatGPT to successfully produce academic writing The

proposed framework includes six principles (relevance, completeness,

boundaries, verifiability, cognition, and redundancy) related to discovering the

accuracy and proficiency of this algorithm-generated text The researchers used

ChatGPT to obtain several academic texts and passages of different genres asanswers to several text-based academic queries The results show that, although

ChatGPT has exceptional capabilities, serious shortcomings remain.

2.5.2 In Vietnam

Hang (2023) studied EFL teachers’ views on using ChatGPT in writing classes at Van Lang University to find what EFL teachers think about using

ChatGPT in writing sessions and to gather ideas for its implementation The

study involved 20 EFL university teachers using ChatGPT in their language

curriculum and teaching many writing courses An online survey and structuredinterviews were used to collect quantitative and qualitative data The resultsshow that these EFL lecturers are excited about using ChatGPT in writing

lessons Furthermore, the study emphasizes the need for professional training of instructors, increasing users’ knowledge of ChatGPT’s restrictions and possible dangers, and that using chatbots properly needs to be assured to contribute to its

successful implementation

Dang and Nguyen (2024) studied the current status of ChatGPTapplications during the research and learning of students at Vietnam NationalUniversity, Ho Chi Minh City The survey subjects were students of six memberuniversities of Vietnam National University, Ho Chi Minh City However, the

article also pointed out five constraints that students of Vietnam National

University, Ho Chi Minh City often encounter when using ChatGPT It said that,

while ChatGPT has become useful for students’ studying and researching, its use requires consideration, responsibility, and the necessary preparation of

practical skills This study provided the basis for further research to improvestudents’ experiences when using ChatGPT

2.6 Conclusion

In summary, the use of ChatGPT in academic English writing hasattracted the attention of educators and researchers globally since the release of

ChatGPT The literature review has identified some perceptions and practices

of using ChatGPT for academic English writing among EFL students in variouscontexts In Vietnam, however, few studies were conducted on this matterparticularly at An Giang University In order to reach this gap, this present studyaims to explore English-majored students’ perceptions and practices of usingChatGPT in academic English writing at An Giang University The design ofresearch, data collection and analysis, as well as proposes hypotheses and

models to answer the research questions will be discussed in the next chapter.

CHAPTER 3: RESEARCH METHODOLOGY

Chapter 3 describes this research method This chapter first presentsthe research design, then the participants, the data analysis method, the datacollection procedure, validity and reliability of the study, and finally the ethicalconsiderations

Mixed-methods research is conducted when the researcher has bothquantitative and qualitative data, and both types of data together provide a betterunderstanding of the research problem than either one of the two types(Halcomb, 2015) Mixed-methods research is suitable if the researcher wants tobuild on the strengths of both quantitative and qualitative data (Edmonds, 2016).Quantitative data, such as specific numbers for statistical analysis, producequantitative results and can provide useful information if a researcher needs todescribe a trend in a large population However, qualitative data, such as open-ended interviews or semi-structured interviews that provide the actual words of

the people in the study, offer a variety of perspectives on the research topic and provide a concrete picture of the situation (Blandford, 2013).

This research method combines paradigms, allowing researchers toinvestigate from both inductive and deductive perspectives (Jogulu & Pansiri,

2011) However, this method also has significant limitations For example, it may be quite complex to plan this mixed-method design and may take more

time and resources to carry out this form of research

In this study, the quantitative and qualitative parts were conducted

sequentially to collect data on students’ perceptions and practices throughquestionnaires and semi-structured interviews A questionnaire was used tocollect quantitative data to explore English major students’ perceptions ofChatGPT and its benefits for academic English writing The researcherdistributed questionnaires to participants in person and on Google Formsthrough the social network The researcher did not limit the number ofrespondents and collected all responses from the voluntary participants Afterthe questionnaire was collected and analyzed, qualitative data were collected

through semi-structured interviews to explore broader and deeper insights intoperceptions and practices of using ChatGPT for academic English writing.

The data collection process from participants through the questionnaire

survey and semi-structured interviews was conducted in the following sequence Firstly, the researcher utilized a questionnaire survey to understand the difficulties that students encountered in academic English writing.

Additionally, the questionnaire included questions aimed at exploring students'perceptions and practices of using ChatGPT as a supporting tool for academicEnglish writing Secondly, based on the information gathered from thequestionnaire, a group of six qualified and voluntary students was selected forparticipation in semi-structured interviews Subsequently, these six students

took part in semi-structured interviews to grasp the difficulties as well as their

perceptions and practices regarding the use of ChatGPT for academic English

writing Each interview lasted for 30 minutes and was conducted in Vietnamese.

Finally, the researcher synthesized the information from the questionnaire

survey and semi-structured interviews and presented the findings.

Therefore, a mixed-methods design was considered to be the mostsuitable approach to answering the research questions in this study

3.2 Population, Sampling and Participants

The population is a statistical term used to refer to the entire set ofindividuals or objects that researchers are interested in studying (Guest, 2014)

This study focuses on English-majored students, including English Language

and English Language Teacher Education students from freshmen, sophomores,and juniors at the Faculty of Foreign Languages at An Giang University, with atotal population size of 900 students

Estimating the required sample size for a research project is crucialbecause it can determine the success or failure of the study (Chow et al., 2017).Additionally, selecting an appropriate sample size is important to ensure the

reliability and generalizability of the results (Boddy, 2016) The sample size,

calculated with 97% confidence level, a 7% error margin, and a 50% responsedistribution, was 176 participants

A sampling method is used in research to select a subset of individuals

or elements from the population to form a study sample (Guest, 2014) The mainpurpose of a sampling method is to ensure the representativeness and

generalizability of the sample, increase reliability, reduce errors, and save time and resources to draw valuable inferences and conclusions about the population

(Mujere, 2016) This study used simple random sampling method to select theparticipants from each year and major, making sure that each subgroup had

3.2.1 Questionnaire participants

The study included 176 EFL students studying a bachelor’s program at

AGU, including English Language Teacher Education (ELTE) and English

Language (EL) programs All participants were full-time English majors On average, they have studied English as a foreign language for 10 years, following

the curriculum offered by the Ministry of Education and Training Throughoutthe study, confidentiality and anonymity measures have been applied to protectparticipants’ privacy

The participants in the survey were freshmen, sophomores, and juniorsfrom the Faculty of Foreign Languages at An Giang University The researcherwanted to select these students for this research because they’re in the process

of studying academic English reading and writing as well as learning andapplying learning support tools like ChatGPT to their learning process

176 participants from 11 classes volunteered to participate in theresearch Table 1 illustrates the demographics of the research participants,including the total number of participants, their major, their school year, andtheir gender

Table 1 Description of the participants

Prefer not to say 6 3.4%

Table 1 shows that the total number of participants is 176 The number

of participants who were enrolled in the ELTE program (N=31) is lower than

that of the EL program (N=145) The participants varied in terms of school

years, including freshmen (N=56), sophomores (N=23), and juniors (N=97).

Moreover, table | also provides information about the gender of participantsincluding male (N=50), female (N=120), and prefer not to say (N=6)

In addition to the general characteristics described above, the

information regarding their Writing course The detailed information isdemonstrated in Table 2 below.

Table 2 Students’ information related to the Writing course

N=176

Items Detail Numbers | Percent

At what level they have | Level A2 73 41.5%

successfully completed Level B2.1 47 26.7%

the Writing course Level C1.1 48 27.3%

Other 8 4.5%

Their average grade for | From 9 to 10 (Excellent) 3 1.7%

the Writing course in From 8 to 8.9 (Good) 38 22.2%

their most recent From 7 to 7.9 (Fair) 65 36.9%

semester (semester 1) From 5 to 6.9 (Average) 64 36.4%

Below 5 5 2.8%

Table 2 shows two specific pieces of information: the proficiency levels

achieved by the participants and the average scores in the Writing course in themost recent semester In terms of Writing course levels, the highest levelattained by the participants is Level A2, with a percentage of 41.5% Level C1.1

and Level B2.1 have nearly equal percentages, accounting for 27.3% and 26.7% respectively Additionally, some participants reached proficiency levels other

than those mentioned above, accounting for 4.5% of the participants

In terms of the average scores in the Writing course in the most recent

semester, the majority of participants fell into two categories: “From 7 to 7.9”

(Fair) and “From 5 to 6.9” (Average), with relatively equal percentages of

36.9% and 36.4% respectively The next category, “From 8 to 8.9” (Good),

accounted for 22.2% of the participants Interestingly, the percentages ofparticipants in the categories “Below 5” and “From 9 to 10” (Excellent) were

relatively equal, accounting for 2.8% and 1.7% respectively.

3.2.2 Semi-structured interview participants

Based on the information provided in the questionnaire survey, six students were selected to participate in the research interviews The selection criteria were based on their GPA, which ranged from 7.0 to 8.9 The main

purpose of this selection was to ensure diversity in the research outcomes and

to avoid having all students from the same class.

and Writing courses for the English major Before their participation, they wereprovided with comprehensive information about the research's purpose,procedures, potential risks, and benefits They were also given the assurance

that they had the right to withdraw from the study at any time without any obligations.

The scheduling of the semi-structured interviews was carefullyarranged to minimize any conflicts with the participants’ regular school hours.The main priority was to ensure that their involvement in the research did nothave any negative impact on the quality of their academic studies Therefore,the interviews were scheduled during their available free time, such as betweenclasses or after school hours, to avoid causing distractions and interfering withtheir ability to focus on their regular coursework

Table 3 Description of the participants for the interview

Interviewee School year Gender Major

Student 1 Seniors Female ELTE Student 2 Seniors Male EL Student 3 Juniors Female ELTE

Student 4 Juniors Female EL

Student 5 Freshmen Female ELTE Student 6 Freshmen Male EL

3.3 Data collection instruments

Data was collected by means of the questionnaire, and semi-structured

interviews The detailed information on these two instruments is described in

the following subsections

3.3.1 Questionnaire

A questionnaire in scientific research is a data collection tool used togather information about the opinions, emotions, and behaviors of participants

through a series of pre-designed questions (Rahi, 2017) The questionnaire can

be administered directly through communication applications, depending on the research objectives and practical circumstances Additionally, a questionnaire

allows for data collection from a large group of participants and can be

completed according to a convenient schedule for participants to increase the

opportunity to collect data from busy or distant individuals (Marshall, 2005).The questionnaire in this study consisted of four sections (see Appendix 1) andadopted Likert scale statements (Strongly Agree= 5, Agree= 4, Neutral= 3,Disagree= 2, Strongly Disagree= 1) except in the first section

Section I includes 5 questions about Personal Information Thesequestions aim to gather information about the participants’ major, school year,gender, proficiency level in writing, and their most recent GPA in the Writing

course.

Section II comprises 4 questions related to difficulties in academicEnglish writing Specifically, it explores the types of academic English writing

essays, specific difficulties encountered during the process of learning academic

English writing, and sources sought for improving academic English writingskills

Section III contains 3 questions focusing on participants’ perceptions

of using ChatGPT for academic English writing purposes These questions aim

to explore students' awareness of ChatGPT and its level of assistance.

Section IV consists of 4 questions that focus on exploring the practices

of English-majored students in using ChatGPT for academic English writing

These questions aim to gather information about the types of assignments

students seek assistance from ChatGPT, the useful features of ChatGPT, and itspotential future usage

Table 4 Questionnaire structure

Section Content Number of

Items

Section I Personal Information 5

Section II Difficulties in academic English writing 4

Section III Perceptions of using ChatGPT for 3

academic English writing

Section IV Practices of using ChatGPT for 4

academic English writing

Below are the information tables for 3 main sections includingDifficulties in academic English writing; Perceptions of using ChatGPT foracademic English writing; Practices of using ChatGPT for academic Englishwriting

Table 5 Questionnaire structure about the difficulties that

English-majored students face in academic English writing

Content Number of Items

Types of academic English writing essays |

Difficulties in academic English writing 1

Seeking support 2

Table 6 Questionnaire structure about students’ perceptions of using

ChatGPT in academic English writing

Content Number of ItemsUnderstanding about ChatGPT 1

Evaluation of ChatGPT’s utility 2

Table 7 Questionnaire structure about students’ practices of using

ChatGPT for academic English writing

Content Number of Items

Types of tasks using ChatGPT |

Helpful features of ChatGPT 1

Using in the future 2

To facilitate accessibility and prevent misunderstandings among

participants, the questionnaire was in both English and Vietnamese languages

The purpose of using both languages for the questionnaire was to ensure that

participants could understand and respond to the questions effortlessly andaccurately (Tsang et al., 2017) Using Vietnamese in the questionnaire ensuredthat students could comprehend the meaning of the questions and select the most

suitable answers quickly and accurately Using English helped participants become familiar with reading and answering questions in English, thereby

enhancing their proficiency in academic English writing In summary, the use

of both languages in the questionnaire ensured the accuracy and effectiveness

of the data collection process

Before distributing the questionnaire to the participants, all the

questions were carefully reviewed and modified by advisors and researchers.

Subsequently, a pilot test of the survey was conducted with 5 EL and ELTE

students from the Faculty of Foreign Languages at An Giang University This

pilot test aimed to ensure that the questions were presented in a specific, clear

The reliability of the questionnaire was examined by SPSS version

20.0 The Cronbach's alpha values for the scale of English-majored students’

perceptions and practices of using ChatGPT in academic English writing were

0.906 and 0.910, respectively (Table 8) whereas the accepted value is 0.6

(Sekaran & Bougie, 2016) The results confirmed the reliability of the

questionnaire.

Therefore, the questionnaire went through a thorough process oftranslation, review, and pilot testing to ensure its coherence and suitability forthe research study

Table 8 Reliability Statistics of the pilot study about English-majored

students’ perceptions and practices of using ChatGPT in academic

English writing

Reliability Statistics

Cronbach's Cronbach's N of Items

Alpha Alpha Based on

Standardized

Items

3.3.2 Semi-structured interview

Semi-structured interviews (see Appendix 3) were conducted in the

subsequent phase to gather more in-depth insights and understanding of Englishmajor students’ perceptions and practices regarding the use of ChatGPT for

academic English writing The semi-structured interviews involved six selected students from the Faculty of Foreign Languages, who were native Vietnamese speakers Using their native language aimed to ensure that the participants could easily express their thoughts and attitudes, thus enhancing their ability to

provide information and making the interview process more comfortable

To maintain confidentiality and adhere to ethical principles, theparticipants were assigned anonymous names The interviews took place at thelibrary of An Giang University and lasted approximately 30 minutes each Theinterviews began with introductory questions to build rapport Subsequently, the

participants were encouraged to share their perceptions and practices regarding

the use of ChatGPT for academic English writing purposes The ultimate

objective of the interviews was to identify the challenges students face in the process of academic English writing, particularly how they have utilized

ChatGPT to support their academic writing journey

The researcher would record the interview sessions with all theparticipants for analysis purposes.

Table 9 Question structure for semi-structured interviews

Content Number of

ItemsIntroduction to questions 1

Difficulties in academic English writing 1

Perceptions of ChatGPT in academic English writing 1

Practices of using ChatGPT for Academic English Writing 2

3.4 Data collection procedures

The researcher used an inductive method to analyze the data collectedfrom the interviews This approach allowed the researcher to distill a largeamount of information The process began with the researcher familiarizingthemselves with the data by repeatedly reading the transcripts Through thisthorough reading, the researcher identified key themes and compiled them into

a comprehensive report This process helped the researcher discover similarities

and differences between the interview transcripts and compare the results By

continuously engaging with the data in this way, the researcher was able to

identify and highlight significant patterns and insights from the interviews

Table 10 Summary of the data collection and analysis

Research Research Research Data Data

Questions participants instrument | collection | analysis

1 Which major | 176 English- | Questionnaire | 2 weeks SPSS

difficulties do majored (Google version

English- students Form) 20.0

majored

students face in 6 English- Semi- Around 30 |

Theme-academic majored structured minutes based English students interview for each analysis writing? interview

2 What do 176 English- | Questionnaire | 2 weeks SPSS

English- majored (Google version

majored students Form) 20.0

students think

about the 6 English- Semi- Around 30 | benefits and majored structured minutes based

Theme-drawbacks of students interview for each analysis

using ChatGPT interview

in academic

Englishwriting?

3 In what ways | 176English- | Questionnaire | 2 weeks SPSS

do English- majored (Google version

majored students Form) 20.0

students use

ChatGPT in 6 English- Semi- Around 30 |

Theme-academic majored structured minutes based

English students interview for each analysis

writing? interview

3.6 Validity and reliability of the study

The following factors have increased the value and reliability of the study.

Regarding internal validity, several methods were implemented to

ensure internal validity during the study The research designs were carefully

planned and executed to minimize potential confounding factors and biases that

employed, combining qualitative and quantitative data collection methods, to

provide a comprehensive understanding of the cognitive and practical aspects

of English major students’ use of ChatGPT for academic writing Furthermore,

the participants selected for the study were English major students from An

Giang University, ensuring that the sample accurately represented the targetpopulation Their participation was voluntary, and measures were taken to

ensure informed consent and confidentiality of their responses The survey

questionnaire and interview protocols were developed based on existing

literature and expert opinions in the field of language learning and technology.

Regarding external validity, the external validity of the study refers tothe generalizability of the findings to other contexts or populations Althoughthe study was conducted at An Giang University, the results can providevaluable insights into the cognitive and practical aspects of English major

students’ use of ChatGPT for academic writing in similar educational settings

or institutions

Table 11 Reliability values of Items related to English-majored students’

perceptions and practices of using ChatGPT in academic English writing

Reliability Statistics

Cronbach’s Cronbach’s N of Items

Alpha Alpha Based on

Standardized

Items

Regarding conclusion validity, conclusion validity refers to the extent

to which the conclusions drawn from the data accurately reflect the findings ofthe study To ensure conclusion validity, rigorous data analysis procedures were

followed, and the results were interpreted in light of the research questions and

objectives

3.7 Ethical considerations

The research was conducted ethically throughout the study process

Before commencing the research, participants were provided withdetailed information about the purpose of the study, as well as the measures

taken to ensure information security and their rights Additionally, participants

were informed that they could withdraw from the study at any time withoutfacing any negative consequences

Secondly, the data were treated confidentially The researcher ensuredthe confidentiality of personal information and participant responses Allcollected data was securely stored and only accessible by the research and the

advisor Participant identities were kept anonymous, and information was protected through the use of pseudonyms or participant codes instead of their

real names

Thirdly, the data were guaranteed The researcher maintained

transparency throughout the research process by accurately reporting the results

and acknowledging any limitations or biases that could affect the findings

By adhering to these ethical considerations, the researchers ensured the

safety and rights of the participants, maintained the integrity of the research, and

contributed to conducting responsible research.