Effectiveness of using pronunciation software in teaching pronunciation for efl learners

VIETNAM NATIONAL UNIVERSITY, HANOI UNIVERSITY OF LANGUAGES AND INTERNATIONAL STUDIES FACULTY OF ENGLISH LANGUAGE TEACHER EDUCATION GRADUATION PAPER EFFECTIVENESS OF USING PRONUNCIATIO

VIETNAM NATIONAL UNIVERSITY, HANOI

UNIVERSITY OF LANGUAGES AND INTERNATIONAL STUDIES

FACULTY OF ENGLISH LANGUAGE TEACHER EDUCATION

GRADUATION PAPER

EFFECTIVENESS OF USING PRONUNCIATION SOFTWARE IN TEACHING PRONUNCIATION

FOR EFL LEARNERS

Supervisor: Nguyễn Huy Hoàng, MA

Student: Nguyễn Minh Hằng Course: QH2013

ĐẠI HỌC QUỐC GIA HÀ NỘI

TRƯỜNG ĐẠI HỌC NGOẠI NGỮ KHOA SƯ PHẠM TIẾNG ANH

KHÓA LUẬN TỐT NGHIỆP

HIỆU QUẢ CỦA ỨNG DỤNG PHẦN MỀM TRONG VIỆC DẠY PHÁT ÂM CHO NGƯỜI HỌC

TIẾNG ANH NHƯ MỘT NGOẠI NGỮ

Giáo viên hướng dẫn: ThS Nguyễn Huy Hoàng Sinh viên: Nguyễn Minh Hằng

Khóa: QH2013

I hereby state that I: Nguyễn Minh Hằng, class QH2013.F1.E1, being a candidate for the degree of Bachelor of Arts (programme) accept the requirements of the College relating to the retention and use of Bachelor’s Graduation Paper deposited

in the library

In terms of these conditions, I agree that the origin of my paper deposited in the library should be accessible for the purposes of study and research, in accordance with the normal conditions established by the librarian for the care, loan or reproduction of the paper

Nguyễn Minh Hằng

May 4, 2017

My sincere thanks also go to the teachers of Business Program in English I, and the teachers in Fast-track Program for allowing me time to carry out the research

in their classes I would also want to thank Ms Do Thu Huong, Mr Nguyen Tuan Anh and Dr Nguyen Chi Duc for their insightful comments on my work, which helped me a lot in completing this thesis paper

At the same time, I would like to give my heartfelt thanks to the first-year students in Business Program, the Fast-track students of class 16E2 and my dearest students at Project 9 Plus for participating in this study, either in the main course or

in the pilot stage, although it involved them staying after class and other complicated requirements

I would also like to thank my dearest friends, especially my colleagues in Project 9 Plus: Thuy Duong, Thu Ha and Huong Tra, for staying with me through thick and thin, for cheering me up when I was down, and for tolerating my unseasonable mood swings during the months of doing research Without their support, I wonder if I could finish this research paper

Last but not least, my sincere thanks extend to my family, who gave me unconditional love and support whenever I needed, which gave me the indescribable sense of security even when I felt like giving up A special thank was to my little brother He was not only a great source of entertainment, but also was a wonderful young man who loved me more than anything in this world

ABSTRACT

The use of technology in language teaching has been a topic of special interest for researchers all over the globe for many years This research investigates the effectiveness of integrating a commercial software named Pronunciation Power in teaching pronunciation for EFL learners Adopting a true-experimental research approach, this study involved the participation of 53 university students, making up two groups Both groups experienced a six-week pronunciation training course with the same content, but one group were trained using pronunciation software while the other group received instruction in the traditional method Pretest and posttest were employed to investigate the progress of each group after training The result was that both groups’ pronunciation ability improved significantly after six weeks, and the group that studied with Pronunciation Power made remarkably better progress than the other group The post-course interview showed students’ positive evaluation of computer assisted pronunciation training and their wish to integrate computer

software in traditional training class

TABLE OF CONTENTS

ACKNOWLEDGEMENTS i

ABSTRACT ii

TABLE OF CONTENTS iii

LIST OF TABLES v

LIST OF ABBREVIATIONS vi

CHAPTER 1: INTRODUCTION 1

1.1 Background of the study 1

1.2 Research aims and research questions 2

1.3 Research design 3

1.4 Significance of the study 3

1.5 Scope of the study 3

1.6 Outline of the study 3

CHAPTER 2: LITERATURE REVIEW 5

2.1 Key concepts 5

2.1.1 Pronunciation 5

2.1.2 Pronunciation teaching 5

2.1.3 Approaches to teaching pronunciation 6

2.1.4 The use of CALL in teaching pronunciation 8

2.2 The integration of software in pronunciation training 8

CHAPTER 3: METHODOLOGY 17

3.1 Research questions 17

3.2 Participants 17

3.3 Teaching materials 18

3.4 Data collection instruments 19

3.4.1 Tests 19

3.4.2 Interviews 21

3.5 Procedure of data collection 21

3.5.1 Selection of participants 21

3.5.2 Pretest 22

3.5.3 Training 23

3.5.4 Posttest 24

3.6 Procedure of data analysis 24

3.6.1 Quantitative data 24

3.6.2 Qualitative data 25

3.7 Ethical considerations 25

CHAPTER 4: RESULTS AND DISCUSSION 27

4.1 Quantitative data 27

4.2 Qualitative data 29

4.2.1 Students’ evaluation of their progress 29

4.2.2 Pronunciation Power software 31

4.2.3 The training course 32

4.3 Discussion 33

4.3.1 Effectiveness of using software in teaching pronunciation 33

4.3.2 Students’ opinion of CAPT 35

CHAPTER 5: CONCLUSION 39

5.1 Summary of findings 39

5.2 Pedagogical implications 39

5.2 Limitations of the study 40

5.3 Suggestions for future research 41

5.3.1 Suggestions for improvement 41

5.3.2 Suggestions of future research directions 42

REFERENCES 44 APPENDICES

LIST OF TABLES

Page Table 3.1 Pretest score of two groups as compared using independent

sample t-test

Table 4.1 Inter-rater score as calculated by Pearson coeffiecient r

Table 4.2 Paired-sample t-test score of Control group

Table 4.3 Paired-sample t-test score of Experimental group

Table 4.5 Posttest score of two groups as compared using

independent sample t-test

LIST OF ABBREVIATIONS

ASR: automatic speech recognition

CALL: computer assisted language learning

CAPT: computer assisted pronunciation training

CG: control group

EFL: English as a foreign language

EG: experimental group

ELT: English language teaching

ESL: English as a second language

FELTE: Faculty of English Language Teacher Education ICT: Information communication technology

L2: Second language

ULIS: University of Languages and International Studies VSTEP: Vietnamese standardized test of English proficiency

CHAPTER 1: INTRODUCTION

This chapter presents the background of the study, which includes an overview

of the history of researching into pronunciation related issues as well as the rising popularity of computer assisted language learning on a national scale Also in this chapter, significance and scope of the study are specified, followed by the outline of the study

1.1 Background of the study

Pronunciation is of utmost importance to communication due to its pivotal role

in facilitating the exchange of information Pronunciation can profoundly affect the perception of listeners and at the same time, reflect speakers’ identity (Liu, 2008) Pronunciation teaching, therefore, is also of great significance The history of pronunciation teaching has reportedly experienced numerous ups and downs, as some

approaches like the audiolingualism “elevated pronunication to a pinnacle of importance” (Levis, 2005, p 369); while approaches such as communicative

language teaching actually “banished [the teaching of pronunciation] to irrelevance”

(Levis, 2005, p 369) However, recent empirical studies have presented a different picture, according to which, pronunciation instruction is necessary and can have positive effects if it is done appropriately

How to maximize the effectiveness of pronunciation instruction has been a central concern for a long time One solution that has piqued researchers’ interest was the use of information communication technology (ICT) in teaching pronunciation This technological era has witnessed numerous innovations in all aspects of life, including education The availability of modern digital devices has provided teachers with powerful tools to facilitate the process of teaching and learning pronunciation

In fact, for several decades, researchers have been interested investigating the effectiveness of utilizing computer-assisted pronunciation teaching (CAPT) (Molholt, 1988) The application of technology in language teaching, or specifically

in pronunciation training, is a promising approach that needs further investigation A large body of research has been carried out in a variety of contexts to investigate the

efficacy of pronunciation teaching when it comes to the application of technology Research mostly focused on certain outstanding features of the pronunciation softwares, which includes (but not limited to): automatic speech recognition feedback (Coniam, 1999; Kawai & Hirose, 2000; Neri, Cucchiarini & Strik, 2006), visual feedback (Hardison, 2004; Hirata, 2004; Neri, Cucchiarini & Strik, 2002) Regarding the effectiveness of pronunciation software in improving students’ pronunciation competence, multiple aspects of pronunciation have also been investigated, including both segmental and suprasegmental elements, in terms of accuracy, intelligibility and accentedness

In Vietnam, the rising popularity of applying ICT in teaching English can be evidenced by the government’s launch of Project 2020 that aims at improving the overall foreign language proficiency on a national scale, via many acts including applying ICT in foreign language teaching (Đề án Ngoại ngữ Quốc gia đến năm 2020, n.d.) At the same time, the fact that the majority of Vietnamese teachers of English are under-qualified, coupled with the unmet demand for English teachers (Yen, 2016) has made ICT a promising solution to the problem of ensuring English teaching quality in Vietnam, especially in pronunciation training However, the matter of how effective the application of ICT in teaching pronunciation is has yet to be universally concurred At the same time, little research has been done to investigate the effectiveness of utilizing ICT in Vietnam These facts serve as the primary motivation for me to conduct this study to investigate the effectiveness of computer-assisted pronunciation training, particularly in the context of Vietnam

1.2 Research aims and research questions

The research aims at investigating the effectiveness of computerized pronunciation training via the use of a commercial software called Pronunciation Power, and students’ opinion of this method of pronunciation teaching At the same time, the research is also designed to propose some recommendations for the future use of pronunciation software in the case of Vietnam

In other words, the research seeks the answers to the questions:

(1) Do students with exposure to pronunciation software make more progress than those who experience traditional teaching method?

(2) What are students’ opinions of the use of CAPT in learning pronunciation?

1.3 Research design

This study is a true-experimental research which adopts a pretest-posttest design with a mixed approach, making use of both qualitative and quantitative data

to answer the research questions

1.4 Significance of the study

Considering the central role of pronunciation in successful communication with regards to mutual intelligibility, the teaching of pronunciation should be paid attention to in order to foster the process of information exchange (Ha, 2005) Research into the Vietnamese pronunciation of English language has shown that segmental errors, such as mispronouncing initial consonant clusters sp-, br-, fr- or str-

as well as consonant sounds (mostly fricatives and affricates), are frequent mistakes that can obstruct comprehension (Tweedy, 2012) Therefore, in the context of Vietnam, where students’ pronunciation is reportedly not intelligible enough (Ha, 2005), innovations in pronunciation teaching are highly expected

1.5 Scope of the study

Pronunciation is a very broad term; as a result, the research did not aim to cover all aspects of pronunciation within six weeks of teaching Therefore, only 10 most problematic sounds and three aspects of intonation are focused on in this research At the same time, participants in the study are limited to first year students

in Business English program, at University of Languages and International Studies

1.6 Outline of the study

The research paper consists of 5 chapters

Chapter 1 presents the background of the study, in which a brief overview of research history in the field of pronunciation teaching was introduced Also in chapter

1, significance and scope of the study are included

Chapter 2 provides a review of literature, in which key concepts are defined, approaches to pronunciation teaching is included, and the use of computer assisted pronunciation teaching is analysed through various materials At the same time, a review of previous studies that investigated similar matters is also included as a base for the development of the current research

Chapter 3 states research questions, as well as giving information about participants, data collection instruments, teaching materials, data collection procedure and data analysis procedure

Chapter 4 presents the results that answer the two research questions, together with the discussion of the effectiveness of integrating pronunciation software in training, as well as participants’ opinions of the course

Chapter 5 discusses the conclusion of the research, discusses the limitation of this study and offers suggestions for future research on the use of software in

pronunciation training

CHAPTER 2: LITERATURE REVIEW

This chapter includes a brief analysis of key concepts, including pronunciation teaching, different approaches to pronunciation teaching, and an overview of the use

of computer assisted language learning (CALL) programs in teaching pronunciation Also in this chapter, a review of previous studies on similar issues is provided together with an evaluation of strengths and weaknesses of the studies

2.1 Key concepts

2.1.1 Pronunciation

Pronunciation is defined as the production of sounds that are utilized to convey meaning (Stanculea, 2015), including segmental (referring to sounds) and suprasegmental (connected with elements of speech, such as: intonation, rhythm or stress) aspects Pronunciation is of utmost importance to EFL learners due to its central role in vocal communication, the perception of listeners and the recognition

of speaker identity (Liu, 2008), in the sense that pronunciation can foster intelligibility and help “identify membership in a special community” (Moghaddam, Nasiri, Zarea, & Sepehrinia, 2012, p 215)

in ELT has led to the question of what should be the goal of pronunciation teaching and whether it should be taught explicitly (Rajadurai, 2007) Controversies surrounding the appropriate techniques and assessment methods continue to

demotivate educators from integrating pronunciation in English teaching Another reason for the negligence in pronunciation training is probably its potential collision with social identity Since it is believed that pronunciation is a crucial part of cultural and individual identities, teaching pronunciation to L2 students can “go against the grain” and threaten their self-image (Rajadurai, 2007, p.3) However, a large number

of empirical studies have highlighted the effectiveness of explicit pronunciation instructions in the classrooms (Liu, 2008; Derwing & Rossiter, 2003) Different approaches to pronunciation teaching present different views as to the appropriate content and the methodology utilized in class (Liu, 2008), which is presented in the following section

Although the application of CAPT systems seems to be a promising approach towards pronunciation teaching, it “remains in its infancy” (Levis, 2007) Nevertheless, considering the potential benefits that CAPT can bring to the pronunciation classrooms, research should be done in order to evaluate the actual effectiveness of CAPT programs in specific contexts as well as finding suggestions for further development in this discipline In recent years, various studies have been carried out to measure the effectiveness of different ASR-based CAPT programs The results, however, were inconsistent and yet to be localized to many specific contexts including Vietnam Therefore, this current research is conducted for the purpose of analysing whether CAPT is effective in improving students’ pronunciation of certain sounds in a specific Vietnamese context, as well as suggesting an appropriate orientation for the development of pronunciation teaching in general

2.1.3 Approaches to teaching pronunciation

Traditionally, the teaching of pronunciation mostly conformed to the nativeness principle, which, according to Levis (2005), was “the dominant paradigm

in pronunciation teaching before 1960s” (p 370) Those supporting nativeness principle hold the belief that “it is both possible and desirable [for adult learners] to achieve nativelike pronunciation in a foreign language” (Levis, 2005, p 370) However, the approach has been criticized as it is impractical, unrealistic and socioculturally inappropriate (Rajadurai, 2007) Recent research reveals that whether

an individual can develop nativelike pronunciation depends on whether the language acquisition occurred before adulthood Specifically, in order to obtain the nativelike pronunciation, learners must start learning in a critical period before childhood, which leads to a conclusion that nativelike pronunciation is unattainable and rather an unrealistic goal to aim at (Levis, 2005; Rajadurai, 2007) At the same time, conforming to a native model is unnecessary as Inner circle English does not always occur in non-native contexts Plus, as language is an indispensible part of a culture, forcing non-native learners to perform according to native norms is a threat to an individual’s or a community’s identity (Rajadurai, 2007) Last but not least, nativeness approach is unrealistic as it contradicts the prevailing lingua franca status

of English, according to which, English is being “reshaped to express local cultures and identities” (Rajadurai, 2007, p 5)

In contrast to the nativeness principle, intelligibility principle puts more focus

on the communicative function of language, according to which, learners just need to

be “understandable” (Levis, 2005) This approach comprises two main kinds of

understanding: intelligibility (the degree to which listeners can decode the words spoken) and comprehensibility (whether the speaker’s message is successfully

communicated to the listeners) (Levis, 2007) Since it is believed that not all elements

of pronunciation have the same effect on understanding, instruction should only focus

on factors that foster intelligibility and ignore those that are “unhelpful” (Levis, 2005,

p 371) Advocates of intelligibility principle maintain that pronunciation teaching should only focus on suprasegmentals instead of segmental factors as in accuracy-oriented approaches However, complete ignorance of segmentals is not the answer

to better intelligibility Adopting the Lingua Franca Core theory, Jenifer Jenkins (2000) has specified several areas of pronunciation that facilitates intelligibility, which includes segmental factors (Setter, 2008) This “renewed emphasis on segmentals” (Levis, 2005, p 371) has been met with interest and attention from researchers and teachers in various contexts, representing the preference for a more balanced approach to teaching and learning pronunciation

2.1.4 The use of CALL in teaching pronunciation

Computer-assisted language learning (CALL) is defined as the application of technology-related methods and techniques in the practice of teaching and learning a language (Gruba, 2006; Hismanoglu, 2011) CALL consists of a variety of

“technological applications and approaches […] such as the proliferation of tools used in virtual learning environment and web-based tools” (Levis, 2007, p 184), which makes CALL an ideal tool for pronunciation training (Cordier, 2009) Various studies have investigated the effectiveness of computer assisted pronunciation training (CAPT) software (Hismanoglu, 2011), especially automatic speech recognition (ASR-based) programs The benefits of these programs mostly concern the capability of addressing students’ individual needs, maximizing learner autonomy and eliminating classroom anxiety (Elimat & AbuSeileek, 2014) This is due to the nature of computers that allow “individualized instruction, frequent practice through listening discrimination […] and automatic visual support” (Levis, 2007, p 184), making computers almost “ideally suited” (p 184) to the learning of pronunciation CAPT, therefore, is a promising approach towards pronunciation teaching In fact, a variety of empirical studies have shown pretty much the same result: CAPT can be effective when it comes to pronunciation instruction, provided that it is utilized wisely (Levis, 2007)

According to Levis (2007), a CAPT system should possess the following qualities:

(1) Provide a substantial amount of meaningful input, including the use of multiple models and accurate articulatory instruction; (2) give learners a reason and desire to practice through rule-oriented practice and realistic materials; [and] (3) provide immediate, useful feedback, especially for those features that are most important for intelligibility (p 186)

2.2 The integration of software in pronunciation training

In this part, six studies into the integration of software in pronunciation training, with contrasting results, were closely investigated, with a special attention

to their methodology, from which the researcher points out some strengths and

possible weaknesses that may influence the reliability of data This acts as a theoretical and empirical base for the construction of the current research

Hardison (2004) investigated the use of computer-assisted programs in French prosody training with the participation of 16 native English speakers whose only source of language input was in-class instruction All participants showed strong motivations to learn French in order to study abroad and/or for travelling purposes The research adopted both qualitative and quantitative approaches with the use of pretest-posttest design and surveys After the posttest, participants were also asked to produce new sentences to “test generalization of training” (Hardison, 2004, p 7) without knowing that they were being tested on that part As for the training session, participants are trained individually using real-time computerized pitch display They were given cards containing sentences, then were asked to look away from the cards and re-produce the sentences at a conversational pace Their speech was then recorded and displayed in real time After that, native speaker’s speech was played and also visualized on the screen The participant’s and the native speaker’s speech contours were then overlaid in contrasting colors Participants kept practicing until they could produce the speech pattern that resembled the version produced by native speakers In the pretest and posttest, participants were required to do similar tasks as the practice, but this time, they receive no visual or auditory feedback Questionnaires included 5 open-ended questions and must be done anonymously

After 13 fourteen-minute sessions of training, students had shown significant improvement in prosody as well as developing greater awareness of prosodic matters, which was observed in the posttest and in students’ responses to questionnaires at the end of the training The result was highly reliable as raters are required to give scores

to both filtered and unfiltered versions of students’ responses, thanks to which the researcher could minimize the influence of segmental factors on raters’ marking However, the questionnaire was not well-designed as it only contained open-ended

questions, which may be answered with a general response Questions like “What do

you feel you focused your attention on during the 3 weeks of practice?” is vague and

can be answered by providing a general statement In fact, the result showed that

students did provide such general answer as “pronunciation.” It would be better were

the questionnaire to conclude statements from a variety of aspects, which would then

be rated by participants on a Likert scale, together with free space for comments In that way, the researcher would be able to avoid too general responses while he could easily see students’ thoughts and attitudes towards the integration of pronunciation software in studying

Neri, Cucchiarini, and Strik (2006) conducted a research on the effectiveness

of ASR-based corrective feedback on pronunciation Studying a group of 30 immigrants who were learning Dutch as a foreign language, the researchers divided them into three groups: one experimental group (EG) with 15 students, and two control groups (CG1 and CG2) with 10 and 5 students respectively EG received training in segmental aspects of Dutch pronunciation with the integration of CAPT, while those in CG1 were trained with a CAPT system that had its feedback function reduced, and CG2 received no extra training besides their regular Dutch class Students’ pretest results were compared with the posttest after 4 training sessions that last for 30 minutes – 1 hour each It was observed that while all groups had made progress in segmental accuracy, the EG group had shown the greatest improvement

At the same time, from students’ self-assessment, students trained with CAPT claimed the effectiveness of CAPT in improving their pronunciation as well as their awareness of pronunciation problems However, the reliability of the research is to

be questioned as the training time was too short (from 2-4 hours in total), and the tests were claimed to be somewhat non-representative of the training content (Neri, Cucchiarini & Strik, 2006) At the same time, since students are from different countries, the inconsistency in learning backgrounds may result in them making different levels of progress

Also studying the effects of pronunciation software on pronunciation in general, Kılıçkaya (2011) carried out research on the integration of accent reduction and text-to-speech software Adopting a quasi-experimental research design, the study aimed at investigating the effect of the software on students’ pronunciation 35 students, aged 22-28, were divided into three groups: one control group (CG) that

followed traditional instructions, experimental group 1 (EG1) that were trained using accent reduction software (Pronunciation Power) and experimental group 2 (EG2) using both accent reduction and text-to-speech software (Text Aloud MP3 and NeoSpeech voices)

The training lasted for 16 weeks with 48 hours of meeting between the instructor and the groups The content of the course was adapted from the book Tree

or Three? by Ann Baker (1993) Regarding the pretest and the posttest, there are three main sections in the tests First, participants are required to pronounce 20 “most frequently mispronounced words” (Kılıçkaya (2011, p.2) The second part includes

10 sentences including these words Finally, the last section is picture description in which test-takers are asked to describe children’s activities The responses were rated

on a five-point Linkert scale

By comparing the results from the pretest and posttest, it was observed that students in EG1 and EG2 made better progress than those in CG, with significant improvement mostly seen at sentence level Semi-structure interviews with two experimental groups revealed students’ positive attitude towards the integration of software in pronunciation training, since they claim that the programs provided visual support, self-pacing practice and immediate feedback, which helped them improve the pronunciation both at word and sentence levels as well as their spelling (Kılıçkaya, 2011)

There, however, still exist some problematic points in the study First of all, the researcher aimed at investigating the effectiveness of the programs on pronunciation in general, which is a rather ambiguous target, since pronunciation is

a broad term that includes numerous aspects, each needing a different approach to teaching The result, therefore, was unclear as to which aspect of pronunciation has students made progress on It would be better if the study paid more attention to a specific area of pronunciation rather than on the whole Secondly, the rating scale utilized in this study was vague, as it was unclear what the researcher wanted to rate, for example, intelligibility or accuracy in participants’ speeches Therefore, the

research needed a more reliable rating scale in order to accurately measure students’ performances

Unlike the aforementioned research that focus on suprasegmentals and pronunciation in general, Pearson, Pickering, and Silva (2011) investigated the influence of CAPT on segmental aspects of pronunciation, which is the

“improvement of Vietnamese learner production of English syllable margins” (Pearson, Pickering & Silva, 2011) 13 Vietnamese students who were living in America were recruited as participants in the research Regarding their learning background, these participants had studied English in Vietnam from 2 to 8 years The placement test, together with a proficiency interview, had determined the levels of the students, which ranged from pre-intermediate to upper-intermediate They were studying in an intensive English course at a university in southeastern America

The training consisted of eight 30-minute sessions, using American English as

a model for studying Materials for training were developed by the researchers, one

of whom (who was a native speaker of English) recorded his voice to use as waveform models, and as a base to design a learning package for students As for the pretest and the posttest, participants were checked on their perception and productions of syllable margins There were three main parts in each test: 30 multiple choice questions to test perception, 55 items to test production and a speaking task All items in the test of production were monosyllabic to avoid the influence of suprasegmental factors in rating Students’ speeches were recorded and rated by two raters according to the level of being “target-like” or “non-targetlike.”

By comparing students’ performances in the pretest and posttest, it was concluded that students had made significant progress in pronunciation However, it

is not clear whether the progress was thanks to the integration of software, since the students were living in America, an English speaking environment As a result, students’ progress could be largely attributed to the constant exposure to native English language

On the other hand, not all studies reached a final conclusion as to the effectiveness of CAPT While much literature has shown remarkable effectiveness of

pronunciation software, which can be evidenced by students’ significant improvement in the posttests compared to those being trained with traditional method, some studies have proved some not-as-positive result

Stenson, Downing, Smith and Smith (1992) studied the integration of IBM Speech Viewer in pronunciation training for international teaching assistants (ITAs) The program was primarily used in medical field for the purpose of treating children with different types of communication disorders by visualizing sounds, thus drawing users’ attention to the possible influence of vocalization and encouraging them to produce appropriate types of speech

Pretest and posttest were also employed to evaluate students’ improvement as compared to the result of those receiving traditional training The pretest and posttest were the SPEAK tests that the ITAs had to take (but failed) before the course and retake after the course finishes Also included in the posttest was the mimic test that includes technical terms in isolation, in phrases or in sentences Test-takers, during the test, listened to native speakers’ model of speech, then were required to produce similar speeches that were as close as possible to the model in terms of pronunciation

of sounds, stress and intonation

The result revealed that, although both the experimental and the control groups showed certain degree of improvement in the test result, there was little significant difference between the results of two participant groups However, this can be attributed to a variety of factors, including the participants’ insufficient exposure to the software to make the best use of it or the ineffectiveness of Speech Viewer itself

as perceived by the users At the same time, the research demonstrated a possible lack

of focus since participants, over a limited time span, were trained in all aspects of pronunciation, instead of one or several aspects The use of pretest and posttest were problematic, because the SPEAK test’s content was independent from the the content

of the training course (the training content was determined based on researchers’ assumptions) As a result, the test may not accurately reflect the effectiveness of training To improve on this point, it is recommended that the researchers made use

of a specially designed test that was closely related to the course content In addition,

the researchers also admitted the lack of established reliability and validity of the mimic test (Stenson, Downing, Smith and Smith, 1992, p.17)

Therefore, this research finding was limited and thus, could not be the basis for generalization for the effectiveness of CAPT as a whole

Research by Liu (2008) also shared the same result The researcher studied the effectiveness of teaching pronunciation using Pronunciation Power, with the participation of 18 graduate students in the US Participants were intentionally recruited from different majors at Iowa State University so as to make sure the experiment could best resemble the authentic environment at that university The participants were varied in terms of age, nationality and academic backgrounds The division of participants into experimental group and control group was based on their own wishes

The training content for two participant groups was developed from the data taken from Pronunciation Power software In addition to that, the teaching materials used for the control group which followed teacher-led instruction was adopted from three other books and a DVD set for animated illustration of sound movements and examples of target sounds This research was basically a revised version of a similar one by Seferoglu (2005) with some changes in methodology, such as the choice of raters with no previous exposure to accented speech and no experience in language training The researcher argued that the choice was made intentionally for the purpose

of avoiding influence from personal experience and assumptions by professionals, thus ensuring the objectivity of the result

Regarding the findings of this study, after comparing the result of the pretest and the posttest, it was concluded that there was little significant improvement in the pronunciation of both groups, as rated in terms of comprehensibility and accentedness Data gathered from the post-class questionnaires revealed that the participants showed little clear preference for any type of pronunciation training, although the overall tone of the responses was rather positive Generally, the participants acknowledge the usefulness of the commercial pronunciation software in teaching pronunciation, and believed that it could replace some parts of human

instructions in regular classrooms About certain features of the program, participants gave relatively low scores to the human demonstration and the speech analysis functions of the software as little instruction was given explain the use of these features

There exist several limitations in Liu’s study (2008) First of all, it is the choice

of participants that is problematic Although the study aimed at investigating ESL class, half of them were EFL students from Vietnam, China and Thailand At the same time, the striking inconsistence in their personal backgrounds and academic majors may make it hard to design appropriate activities and learning content for the whole group In addition, as participants were given the freedom to choose to study

in the group they wanted, it was possible that they have already hold certain preference to either type of classroom, which can affect the objectiveness of the study, especially in the analysis of the secondary research question about course evaluation Secondly, another factor that could have affected the result, as perceived by the researcher, was the lack of suprasegmental instructions of the software Pronunciation Power Most of the practice and instructions included in this program were about segmental elements Therefore, participants may not show remarkable progress in the rating of comprehensibility and accentedness due to the lack of suprasegmental training, which otherwise should be the focus of the whole course Since students did not have enough practice in suprasegmentals, their pronunciation after the course could still be heavily accented, which made it challenging for the raters to figure out their speech due to their lack of exposure to these accents In addition, the rating scale also needed revising for better assessment result As raters were asked to rate the comprehensibility of students’ speech on a scale of 1 to 9, the rating was not only highly subjective, but also too general that students’ progress of even 5% could not

be recognized in scores (Liu, 2008)

This current research

Based on the aforementioned analysis of related studies in this field, the current research inherited several favorable qualities, such as the use of two raters to enhance data reliability or the inclusion of both perception and production of speech

in the tests At the same time, attention has been paid to areas that need improvements

in the previous studies Specifically, this research made use of semi-structured interview instead of questionnaires, so that the interviewer can gain better insights into participants’ opinions of CAPT At the same time, the research was carried out with a clear focus on improving students’ pronunciation of 10 sounds and 3 suprasegmental aspects (intonation, sentence stress and linking sound) Regarding the choice of participants, they were intentionally chosen from the same program (Business English) at university to minimize inconsistency in backgrounds or

exposure to English

CHAPTER 3: METHODOLOGY

This chapter presents the research questions, as well as providing information about participants, teaching materials and data collection instruments At the same time, data collection procedure and data analysis procedure are also described in details in this chapter

3.1 Research questions

This true-experimental research employed both quantitative and qualitative approach Interviews and experiments with the pretest-posttest design are utilized to investigate the effectiveness of CAPT in teaching pronunciation as well as students’ perception of this method In order to investigate the effectiveness of CAPT programs

in teaching pronunciation as well as students’ evaluation of this integration, this research aims at providing answers to the two following questions:

(1) Do students with exposure to pronunciation software make more progress than those who experience traditional teaching method?

(2) What are students’ opinions of the use of CAPT in learning pronunciation?

3.2 Participants

The research was conducted among 53 first year students, 42 females and 11 males, msajoring in Business English at the University of Languages and International Studies They were all 19 of age when I conducted this study Students were divided into two groups: Experimental group (EG) with 27 participants and Control group (CG) with 26 participants Both groups studied the same content designed with reference to the official course syllabus for first year mainstream students While Control group experienced traditional lessons, Experimental group studied with Pronunciation Power software, with the teacher working as teaching assistant only

The research is carried out in the second semester of the school year

2016-2017 In the first semester, all students have undergone pronunciation training, with

a focus on segmentals, as a compulsory part of their Social English subject at university However, via the researcher’s observation at the beginning of the second

semester, the participants’ pronunciation was still problematic, which could seriously obstructed listeners’ comprehension Therefore in this research, segmental training was still included, coupled with suprasegmental one, as this is the main focus of pronunciation training in their formal study

3.3 Teaching materials

Pronunciation Power is a commercial software containing two disks for two different levels: Disk 1 is suitable for students up to intermediate level of proficiency, while Disk 2 is appropriate for students at higher levels The main difference between Disk 1 and Disk 2 is the types and difficulty level of exercises within each As all students are in their first year at university, I suppose Disk 1 is more suitable for them

Regarding the content, Pronunciation Power software includes:

of the articulators’ positions and movements as well as the flow of air Students will also have a chance to observe how the sound is actually produced by a native speaker Accompanied with the illustration of each

sound are the tips for pronunciation

match the waveform of a given sound, distinguish between two sounds in minimal pairs, listen to sentences and re-produce the statements resembling those given in the program, record their voices and compare the samples Attention is given to both segmental and suprasegmental elements of

speech

purposes Words containing the sounds in training can be classified in terms

of themes, sounds or alphabetically arranged like a traditional dictionary Together with each word is the transcription of sounds, meaning and an

illustration

 recognition games, in which users hear the words and have to choose the pictures that show the meaning of the words The games can be

challenging due to the time pressure (See Appendix 5)

Although the software is regarded as well-designed for the purpose of segmental training for students, suprasegmental practice is somehow neglected and does not receive sufficient attention from the developers Instruction on suprasegmental training is included, but very briefly and inconsistently, as a small practice part in the instruction of individual sounds Therefore, it is hard for users to choose specific suprasegmental aspects to practice

As a result, the researcher has added further explanation and instruction in terms of suprasegmental training so as to boost students’ understanding, better exploit the given instruction and exercises, thus improve the effectiveness of learning However, students mostly worked with the software, and additional instruction by the researcher was just minor and based on the content of the lessons in the software

3.4 Data collection instruments

Part A: Listening discrimination

Participants are required to identify the sounds spoken in minimal pairs The purpose of this part is to test students’ ability to distinguish the target sounds from similar ones In this part, there are 20 minimal pairs (2 pairs for each target sound), each includes one word with the target sound and another with similar sounds With each correct answer, participants get 0.5 point Total score for this part is 10

Part B: Pronouncing words with target sounds

Participants are asked to pronounce words containing the target sounds This part is to test students’ ability to pronounce the sounds in 20 individual words (2 words for each target sound) as well as their ability to put stress on the correct syllable If the participant can pronounce the target sound in a word correctly, he/she gets 0.25 point even if the participant mispronounces other sounds in the same word With each correct word stress, he/she gets 0.25 point Total score for this part is 10

Part C: Role-play

This part is to test students’ intonation and pronunciation of individual sounds

by role-playing two short conversations Participants are tested in pairs, each in the pair roleplays either Student A or Student B Participants are given unlimited time for preparation before they speak Performances are rated from 1 (very poor) to 5 (excellent) on four criteria, namely: segmental accuracy, sentence stress, intonation and linking sounds, which are the four focuses of pronunciation training Maximum score for this part is 20

The marking rubrics used for part C were specifically developed to evaluate students’ progress after this course as the criteria in rubrics included just the main features of the training course With an intelligibility approach, the rubrics were intentionally made to balance between accuracy and comprehensibility, which can be seen in the description of each score that contains both the level of accuracy and level

of intelligibility For example, in order to achieve average score (3) for Linking sound criterion, test-takers should “demonstrate attempt to link sounds but mistakes are occasionally found; however, the errors do not affect listeners’ understanding of the speech.”

Overall, the total score for three parts is 40, with 20 points for segmental accuracy and 20 points for the suprasegmentals

The test format and content were decided by the researcher, with reference to various studies in the same fields (Liu, 2008; Gambari, Kutigi & Fagbemi, 2014; Cordier, 2009) Words used for part 1 and 2 were taken from participants’ course book in the first semester

Before pretest and posttest were finalized and given to participants, they were piloted among 25 students, all of whom were English majored students, with 6 sophomore and 19 freshman students from Fast track program Pearson Correlation

Coefficient r was used to calculate the reliability of two tests Result showed that the two tests are highly consistent in the level of difficulty, with coefficient r as high as

0.86 As a result, the two tests can be used as pretest and posttest, and the result from which can be used to test participants’ progress after training

Additionally, since the pilot tests were carried out on second year students and first year students from fast track program, who are presumably better at pronunciation than the actual participants, so the scores on the real pretests may be even lower Therefore, the difficulty level of the two tests is probably appropriate for the participants in this study

3.4.2 Interviews

After the course, 11 students from Experimental group participated in a structured interview with the researcher, in which they were asked to self evaluate their progress (if any) in each aspect of pronunciation covered in the course, as well

semi-as commenting on the strengths and weaknesses of the program Pronunciation Power Students were also asked to compare the software integrated training with traditional method of pronunciation teaching, and make suggestions as to how to maximize the effectiveness of the program

In order to make it easy for students to express their thoughts and feelings, interviews were carried out in their mother tongue (Vietnamese)

3.5 Procedure of data collection

3.5.1 Selection of participants

Participants were intentionally chosen from two classes from the same program at university to minimize the discrepancy in the instruction received from their official subjects The researcher also purposefully selected participants among first year students at ULIS, as they were presumably at the lowest level of English proficiency, as compared to sophomores and higher

Before the selection of participants, the researcher had consulted the teachers who were directly in charge of teaching the two classes to see if the students were appropriate for the purpose of this current research Besides, the researcher also joined the two classes in their formal lessons to observe their pronunciation abilities Notes taken show that both classes faced almost the same problems with pronunciation, such as the pronunciation of some certain individual sounds, which later were selected as the target sounds of this research, and some suprasegmental problems in linking sounds, word stress, sentence stress or intonation

Seeing the potential of improvement in these students, the researcher decided

to select them as participants of this research Both groups had a semi-formal meeting with the researcher before joining the course, in which the research purposes, procedure, participants’ rights and responsibilities were explained in great details to the students They were asked to sign in a consent form if agreeing to partake this course, according to which, they could draw out of the program at any time Also in the same meeting, the structure of the pretest was described This test also acted as

an entrance test for participants, in the sense that those with higher score than 90% would not be chosen as participants, because they were assumed to have already excelled in pronunciation, and improvement (if any) would be hard to see

One group was randomly chosen as Control group to receive traditional instruction, and the other was the Experimental group that would experience pronunciation software as the primary means of instruction

3.5.2 Pretest

Before the beginning of the CALL course, all participants participated in a pretest For part A, all participants sat in one classroom and together listened to the words in 20 minimal pairs Students listened to the recording once only For part B and C, students were called out of the classroom to do the test All the answers were recorded for later analysis Students with remarkably high scores (who achieve more than 90% of the test, equivalent to more score 36/40) were not chosen

According to the two groups’ teachers, their pronunciation ability seemed to

be the same, and the difference (if any) was not remarkable Independent sample

t-test was also utilized to compare the two groups’ level of pronunciation ability, or specifically, to see if their levels were approximately the same at the beginning of the course

Table 3.1 Pretest score of two groups as compared using independent sample t-test

Sig tailed)

(2-Mean Difference

Std Error Difference Pretest Equal variances

3.5.3 Training

The class was taught by the researcher, a non-native English speaker I was a senior student at the University of Languages and International Studies (ULIS), majoring in English language teacher education At the time of research, I was a practicum teacher at Faculty of English Language Teacher Education, being in charge

of teaching Social English subject (including pronunciation) for 4 first year classes excluding the class of participants Regarding teaching experience, I had been working as both private tutor and formal class teacher for three years, working mostly

on speaking and listening skills

The training lasts for 6 weeks, including 6 in-class sessions; each session lasts for 30 minutes Students participate in an extra session before the course to get to know the program and familiarize themselves with its functions

Based on the data collected via the pre-course observation, coupled with the research results concerning the common pronunciation mistakes by Vietnamese speakers (Ha, 2005), a syllabus was designed Due to the limited scope of this research, the course cannot cover all elements in pronunciation, but just focuses on the sounds that students need to improve and the suprasegmental aspects that are the main content of pronunciation training in their formal study at university

The course mostly covers the pronunciation of fricatives and affricates, as these are some of the consonants that many Vietnamese students usually have trouble with (Ha, 2005)

3.5.4 Posttest

The posttest follows the same structure, and is at the same difficulty level as the pretest Like the pretest, information about the posttest was announced to all participants before they take the test

3.6 Procedure of data analysis

3.6.1 Quantitative data

The speaking performances of students during pretests and posttests were recorded, and then rated in terms of accuracy For part C, two raters with excellent English proficiency, which can be evidenced by their VSTEP Score of 9.5 (9.5 and

10 for Speaking skill), gave scores to students’ performances with reference to the marking rubrics specifically designed for assessment in this particular research Inter-

rater reliability was calculated using Pearson correlation (r)

The students’ scores in the pretest and posttest were compared to see if there

is any improvement after the course Paired-sample test and independent sample

t-test were utilized to compare the progress (if any) between two participant groups and within the groups themselves