A study on improving the pronunciation of first-Year English majors during a speech training course at Quy Nhon University

33
QUY NHON UNIVERSITY
SCIENCEJOURNAL OF
Tạp chí Khoa học - Trường Đại học Quy Nhơn, 2019, 13(6), 33-42
Nghiên cứu cải thiện khả năng phát âm 
của sinh viên chuyên ngữ năm nhất 
tại trường Đại học Quy Nhơn trong khóa Luyện âm
Nguyễn Hoài Dung*, Bùi Thị Minh Nguyệt, Lê Thị Thanh Tâm
Khoa Ngoại Ngữ, Trường Đại học Quy Nhơn
Ngày nhận bài: 10/09/2019; Ngày nhận đăng: 10/10/2019
TÓM TẮT
Phát âm là yếu tố đóng vai trò then chốt trong kĩ năng nói. Tuy nhiên các lớp học tiếng Anh chính quy ở 
trường phổ thông tại Việt Nam vẫn chưa dành sự chú ý đúng mức cho phương diện này cho đến khi các em học 
chuyên sâu hơn trong khóa học Luyện âm ở đại học cho sinh viên chuyên ngữ năm thứ nhất. Nghiên cứu này 
nhằm điều tra hiệu quả của việc áp dụng của hai phương pháp hỗ trợ Shadowing (phương pháp nghe và lặp lại) 
và Feedback (phương pháp phản hồi) ngoài chương trình chính khóa của sinh viên nhằm cải thiện khả năng phát 
âm của sinh viên chuyên ngữ, đặc biệt là sinh viên ngành sư phạm. Thực nghiệm kéo dài trong một học kì, tiến 
hành trên 67 sinh viên năm thứ nhất, khoa Ngoại ngữ, trường Đại học Quy Nhơn. Những người tham gia được chia 
thành 2 nhóm, nhóm thực nghiệm (n=43) và nhóm đối chứng (n=24). Dữ liệu được thu thập qua các bài kiểm tra 
trước và sau thực nghiệm, sau đó được phân tích thống kê. Kết quả cho thấy các sinh viên trong nhóm thực nghiệm 
đã có nhiều tiến bộ sau khóa học, đặc biệt trong việc thể hiện trọng âm của câu, điều này cho thấy tính hiệu quả 
của các phương pháp thực nghiệm. Kết quả phân tích cũng cho thấy sự tiến bộ khiêm tốn của nhóm thực nghiệm 
so với nhóm đối chứng trong việc khắc phục lỗi phát âm âm cuối (“ed”, “s/es”) và nối âm đã chỉ ra một thực tế là 
tinh thần tự học của sinh viên còn thấp. 
Keywords: Phát âm, Shadowing (phương pháp nghe và lặp lại), Feedback (phương pháp phản hồi).
*Corresponding author. 
Email: nguyenhoaidung@qnu.edu.vn
34
TRƯỜNG ĐẠI HỌC QUY NHƠN
KHOA HỌCTẠP CHÍ
Journal of Science - Quy Nhon University, 2019, 13(6), 33-42
A study on improving the pronunciation of first-year 
English majors during a speech training course 
at Quy Nhon University
Nguyen Hoai Dung*, Bui Thi Minh Nguyet, Le Thi Thanh Tam
Department of Foreign Languages, Quy Nhon University
Received: 10/09/2019; Accepted: 10/10/2019
ABSTRACT
As far as the speaking skill is concerned, pronunciation is a key area. However, in Vietnamese formal EFL 
classes, scan attention is paid to this aspect until the learners enter higher education, with Speech Training being 
introduced in the first year. This study aims at investigating the effect of using the techniques of Shadowing and 
Directing Effective Feedback in company with the formal course of Speech Training to upgrade the pronunciation 
of English majors at university, especially those who are trained to be teachers of English. An experiment was 
conducted, spanning one semester. It involved 67 first-year students in the Department of Foreign Languages, Quy 
Nhon University. The participants formed a treatment group (n =43) and a control group (n = 24). The data were 
collected by means of post-test and pre-test and statistically analyzed. The results indicate an improvement in the 
learners’ performance, especially in sentence stress, suggesting the effectiveness of the measures undertaken. The 
data obtained also revealed that the employment of the techniques for the treatment group did not result in vast 
difference from the control group as expected in bettering the students’ articulation of final sounds and aspects of 
connected speech, which suggests the students’ low level of autonomy. 
Key words: Pronunciation, Shadowing, Feedback.
1. INTRODUCTION
For decades, English has been the foreign 
language that interests Vietnamese people of all 
ages. People learn English for business, for fun, 
for travelling, for international relationships, for 
reading foreign books, for scientific research, 
for overseas survival and so on. Courses of 
English are available nationwide, formally and 
informally. In universities and colleges, English 
departments often attract the most students who 
are trained in formal courses of standard English, 
which is especially essential for classes of 
teachers-to-be. Standard English is easily found 
through spoken English, where pronunciation 
including sounds, connected speech and 
*Tác giả liên hệ chính. 
Email: nguyenhoaidung@qnu.edu.vn
intonation, plays the key role. Nevertheless, it is 
a fact that the speaking skill in terms of fluency 
and naturalness and pronunciation in particular 
of English majors at Quy Nhon University has 
still been far from satisfaction. 
At Quy Nhon University, as regards the 
Department of Foreign Languages, although 
the students, like elsewhere, started to learn 
English at the 6th form when they entered the 
junior high school, some even started at primary 
school, they have not had a formal course of 
pronunciation until they become students of the 
English department. The Speech Training course 
lasts 30 periods (or 2 credits); students take an 
oral examination on pronunciation at the end of 
35
QUY NHON UNIVERSITY
SCIENCEJOURNAL OF
Journal of Science - Quy Nhon University, 2019, 13(6), 33-42
the course. Like other Vietnamese ESL students, 
English majors at Quy Nhon University share 
many of the same pronunciation errors that ESL 
students from elsewhere do, as reported in some 
previous studies (Nguyen Thi Thu Thao, 2007; 
Nguyen Thi Hang; 2014; Ha Cam Tam, 2005). 
The most common mistake relates to ending 
sounds. This is probably because Vietnamese 
has no inflectional endings like -ed and -s/es 
and the ending consonants in Vietnamese are 
not pronounced, so Vietnamese learners easily 
make mistakes in pronouncing English final 
consonants. They may be confused about /t/ and 
/d/ for the past form -ed or /s/ and /z/ for the 
inflectional ending -s. They even sometimes do 
not enunciate those ending sounds at all. 
This article reports a study aimed at 
improving the students’ spoken English through 
promoting their autonomy. The study was aimed 
to explore how effective some extra measures 
are in improving students’ pronunciation in the 
Speech Training course at Quy Nhon University. 
The research questions are:
1. To what extent can the application of 
Shadowing and Directing Effective Feedback 
help improve the English majors’ pronunciation?
2. What implications can be drawn from 
the experiment?
2. THEORETICAL BACKGROUND 
2.1. The Speech Shadowing Technique 
The Speech Shadowing Technique (also 
Shadowing) is a language learning method 
developed by Alexander Arguelles, an American 
Professor, who defines Shadowing as “an 
advanced learning technique where you listen to 
a text in your target language, and then speak 
it aloud at the same time as the native speaker” 
(Argüelles, 2009a). Tamai (2002, p. 181) sees 
Shadowing as “an act or task of listening in 
which the learner tracks the heard speech and 
repeats it as exactly as possible while listening 
attentively to the incoming information”. 
Generally, Shadowing is learning by imitating 
an auditory source as model.
 In his “Shadowing Discussed”, Arguelles 
(2009b) states that this technique is very effective 
to learn new languages because: it helps in 
better pronunciation; it improves vocabulary; it 
gains fluency of the language; and it creates an 
impression of the sentence structures in mind.
Murphey (2004, p. 21) suggests three 
kinds of Shadowing:
l Complete Shadowing: Using this 
Shadowing, learners are to imitate every single 
word spoken by the model. 
l Selective Shadowing: Here, there is 
a selection of what the model says to shadow. 
Usually, only key inputs are chosen to imitate. 
l Interactive Shadowing: With this type, 
practitioners are allowed to add their own 
comments into the original conversation to make 
it more natural. 
2.2. Directing Feedback Technique
Hattie (1999, p. 9) describes Feedback 
as “the most powerful single moderator that 
enhances achievement”. Similarly, Gibbs 
& Simpson (2004, p. 2) state that Effective 
Feedback is “more strongly and consistently 
related to achievement than any other teaching 
behavior [...] this relationship is consistent 
regardless of grade, socioeconomic status, 
race, or school setting. Feedback can improve 
a student’s confidence, self-awareness and 
enthusiasm for learning” and finally, they add 
that “feedback is something that every student 
can benefit from, whether it is offered digitally, 
verbally, or through the traditional written 
annotations on an assignment”.
According to Winstone and Boud (2019, 
p.9), to be effective, feedback should be 
educative, be timely, consider the individual 
needs of the student, and be aimed at a specific 
skill or knowledge.
2.3. Phonetics and Phonology: Relevant aspects
The presentation of the following basic 
issues in English phonetics and phonology 
is heavily based on the widely-circulated 
coursebook ‘English Phonetics and Phonology’ 
by Peter Roach (2004).
36
TRƯỜNG ĐẠI HỌC QUY NHƠN
KHOA HỌCTẠP CHÍ
Tạp chí Khoa học - Trường Đại học Quy Nhơn, 2019, 13(6), 33-42
2.3.1. Stress
Stress, as defined in Oxford Advanced 
Learner’s Dictionary (2018), is “an extra force 
used when pronouncing a particular word or 
syllable”. The stress placed on syllables within 
words is called word stress or lexical stress, as 
in ‘holiday, a’lone, admi’ration, confi’dential. 
There is no rule about which syllable is stressed 
in a word with more than one syllable. Learners 
of English as a foreign language (EFL) will need 
to learn the stress of words by heart.
The stress placed on words within 
sentences is called Sentence stress or prosodic 
stress. In spoken language, grammatical words 
(auxiliary verbs, prepositions, pronouns, 
articles,) usually do not receive any stress. 
Lexical words, however, (nouns, verbs, 
adjectives, adverbs,) must have at least one 
stressed syllable. For example, The ‘children are 
‘waiting for their ‘teacher.
2.3.2. Final consonants 
Roach (2004, p. 35) states that “Consonant 
is a basic speech sound in which the breath 
is at least partly obstructed and which can be 
combined with a vowel to form a syllable”. 
Consonants may occur at the beginning, in the 
middle or at the end of a word. Final consonants 
include -ed-ending and -s/es-endings, which are 
usually referred to as ‘inflectional endings’.
2.3.3. Consonant clusters
Structurally, an English syllable consists 
of an onset, a center, and a coda, with the 
onset and the coda being optional. Whereas the 
center is always realized by a pure vowel or a 
diphthong, the onset and the coda is realized by 
one or more consonants, referred to as consonant 
clusters. For example, spy, stay, sky, sphere, 
small, snow, sleep, swear, suit; ‘split’, ‘stream’, 
‘square’; bump, bent, bank, belt, ask; bets, beds, 
backed, bagged, eighth etc.
2.3.4. Connected Speech
Connected speech is spoken language 
used in a continuous sequence, commonly used 
in casual, informal conversations.
There is often a significant difference 
between the way words are pronounced in 
isolation and the way they are pronounced in 
the context of connected speech. Roach (12:35) 
introduces four aspects of connected speech: 
Elision, Assimilation, Linking, and Intrusion. 
3. METHODOLOGY 
3.1. Research design 
In order to address the aim and objectives, 
an experimental design was used to measure the 
extent of effectiveness of the chosen techniques. 
This is also a mixed-methods study, in which 
the data were collected both qualitatively and 
quantitatively. 
After considering various ways to upgrade 
the students’ English pronunciation, we decided 
to employ the Speech Shadowing technique 
in combination with the Directing Effective 
Feedback technique as the measures to enhance 
students’ pronunciation.
3.2. Participants
The experiment was carried out on two 
pedagogical classes of 67 first-year English 
majors (Course 41th) at Quy Nhon University. 
Forty-three students in the treatment group 
(Class A) had an extra pronunciation exercise to 
record and got the teacher’s feedback on their 
performances weekly while 24 students in the 
control group (Class B) did not. A Welch’s t-test 
(Welch’s Test for Unequal Variances) revealed 
that there was no statistically significant 
difference between the two groups in terms of 
the three pronunciation skills investigated prior 
to the experiment (p > 0.05).
3.3. Instruments
3.3.1. Pre/ Post-Test
In order to assess the participants’ 
pronunciation, a pre-test was given to the two 
groups right before the Speech Training course 
started, in which the students were required 
to read aloud a printed short dialogue after a 
preparation of 10 minutes. 
37
QUY NHON UNIVERSITY
SCIENCEJOURNAL OF
Journal of Science - Quy Nhon University, 2019, 13(6), 33-42
For the sake of the test’s reliability, the 
same dialogue was subsequently used as the 
post-test in the last week of the course. All of the 
participants’ performances during the Pre/Post-
test were recorded with a mobile phone for later 
analysis.
Both the pre-test and post-test were 
accessed by the instructor of the course, 
according to the same rubrics. 
Although all the phonetic and phonological 
features in the dialogue produced by the English-
majors of the groups are examined, only the 
following aspects were taken into consideration 
in the survey: the pronunciation of Ending 
Consonants, Connected Speech; Sentence Stress.
The participants’ performances were all 
recorded with mobile phones. The audio files 
were then labeled as A
n
.Pre or B
n
.Pre (for those 
of the pre-test) and A
n
.Post or B
n
.Post (for those 
of the post-test). These were analyzed afterwards 
in terms of the sentence stress, connected speech 
and the pronunciation of final consonants. A 
represents experimental group; B represents 
control group, and n indicates the ordinal number 
of the participants.
3.3.2. Materials 
During the course of ten weeks, ten short 
texts or dialogues were chosen as the extra 
weekly pronunciation task for the students in the 
treatment group to shadow. The five first texts 
ranges from 50-88 and the last five are of 97-
139 words, all of which were taken from the 
course books of Speech Training and Language 
Skills 1.1-1.2 (Handcock, 2003; Falla & Davies, 
2012a; Falla & Davies, 2012b) since they are 
appropriate to the participants’ level in terms 
of vocabulary and the audio files are all high 
quality recordings by native English speakers 
with enough clarity and medium speed. 
3.4. The Experiment 
During the course of the experiment, a 
mixed utilization of informal, formative and 
constructive feedback is mainly used in chats on 
Facebook messengers to grasp areas of weakness 
or strengths, and to encourage a focus on future 
improvement. Formative feedback helps students 
to improve and prevent them from making the 
same mistakes again. The experiment spanned 
through one semester.
3.5. Data analysis 
Although all features of phonetics and 
phonology in the texts that the treatment group 
performed were examined and fed back, only 
mistakes relating to Sentence Stress, Final 
Consonants, and Connected Speech that both 
groups revealed in the two tests by both groups 
were analyzed and reported. In case of Connected 
Speech, the items in the sub-group Natural, are 
in fact, not exactly mistakes, but refer to the fact 
that the students could connect words naturally 
as native speakers.
Next, the data were processed with the 
software IBM SPSS Statistics 20, of which the 
Welch’s t-test determined whether the two groups 
were statistically different after the treatment, 
while The Wilcoxon Sign Rank test showed 
whether there was any significant difference 
in the mean scores within the group under the 
experiment itself. 
4. FINDINGS AND DISCUSSION 
4.1. Results of the Treatment and Control 
Groups in the Post-test
The study is aimed to explore whether the 
two techniques applied affect the performance 
of the treatment group in terms of Sentence 
Stress, Connected Speech and Final-word 
Consonants as compared with that of the 
control group. To this end, the Welch’s t-test 
was conducted with independent variables 
being the conditions (treatment versus control) 
and dependent variables being Sentence Stress, 
Wrong Connected Speech, Unnatural Connected 
Speech, Natural Connected Speech, -ed-Ending, 
s/es-Ending and Redundant -s-Ending of the 
post-test. Table 4.1 below presents students’ 
scores on all measures in the post-test.
38
TRƯỜNG ĐẠI HỌC QUY NHƠN
KHOA HỌCTẠP CHÍ
Tạp chí Khoa học - Trường Đại học Quy Nhơn, 2019, 13(6), 33-42
Table 1. Results of the Post-test of the Treatment and Control Groups
Sum of 
Squares
df
Mean 
Square
F Sig.
Sentence Stress
Between group
Within groups
Total 
15.195
180.656
195.851
 1
65
65
15.195
 2.779
5.467 .022
Wrong Connected 
Speech
Between group
Within groups
Total
.596
25.703
26.299
 1
65
65
.596
.395
1.507 .224
Unnatural Connected 
Speech
Between group
Within groups
Total
.904
101.275
102.179
 1
65
65
.904
1.558
.580 .449
Natural Connected 
Speech
Between group
Within groups
Total
11.767
228.950
240.716
 1
65
65
11.767
3.522
3.341 .072
ED Ending
Between group
Within groups
Total
.153
20.563
20.716
 1
65
65
.153
.316
.485 .489
S/ES Ending
Between group
Within groups
Total
11.689
258.819
270.507
 1
65
65
11.689
3.936
2.936 .091
Redundant-S Ending
Between group
Within groups
Total
1.692
126.965
128.657
 1
65
65
1.962
3.982
.866 .356
Figures in Table 1 show that after the 
treatment, there was a statistically significant 
difference between the treatment group and 
the control group in the scores of the Sentence 
Stress (p = 0.022 < 0.05). The mean scores for 
each group in Table 2 below confirmed that 
the experimental group made significantly less 
Sentence Stress mistakes than the control group.
Table 2. Pre-test and Post-test Mean Scores of Sentence Stress
N
Pre-test Post-test
Mean
Std. 
Deviation
Mean
Std. 
Deviation
Sentence 
Stress
Treatment Group 43 3.07 2.685 1.47 1.638
Control group 24 4.33 3.749 2.46 1.719
Total 67 3.52 3.140 1.82 1.723
To gain more insights into the mean scores 
of the treatment and control groups in the Post-
test, let us have a look at Table 3.
39
QUY NHON UNIVERSITY
SCIENCEJOURNAL OF
Journal of Science - Quy Nhon University, 2019, 13(6), 33-42
Table 3. Descriptive Statistics on the Post-test across Groups
Measures Conditions N Mean Std. Deviation
Sentence Stress Mistakes
Treatment
Control
43
24
1.47
2.46
1.638
1.719
Wrong Connected Speech Mistakes
Treatment
Control
43
24
.51
.71
.703
.464
Unnatural Connected Speech Mistakes
Treatment
Control
43
24
.67
.92
1.017
1.586
Natural Connected Speech cases
Treatment
Control
43
24
2.79
1.92
2.065
1.472
ED Ending Mistakes
Treatment
Control
43
24
.56
.46
.548
.588
S/ES Ending Mistakes
Treatment
Control
43
24
2.84
3.71
1.902
2.156
Redundant-S Ending Mistakes
Treatment
Control
43
24
1.42
1.75
1.159
1.751
The highest mean score (3.71) was 
registered when the comparison students make 
more -s/es-ending mistakes in the post-test.
A closer look at those figures revealed 
that the mistakes students make the most is 
the final consonant mistakes with s/es-ending 
accounting for the highest mean score of the 
three subtypes regardless of the conditions they 
were assigned into. 
4.2. Results of the Treatment Group in the