Professional Documents
Culture Documents
DLLS 2000
70
categories: (1) actual linguistic performance improvements achieved by students and (2) changes in their attitudes
toward pronunciation instruction.
Our research design was loosely patterned after Derwing, Munro, and
Wiebe's (1998) study. We, however,
added an extra rating task to the two
they used. We also conducted a more
detailed analysis of changes in students'
performance and attitudes.
METHOD AND PROCEDURES
Partici pants
The fourteen students (six males,
eight females) in the primary group were
enrolled in an English Language Center
(ELC) level 5/6 listening-speaking class
at Brigham Young University. They
received communicative pronunciation
instruction from one teacher and more
general, speech-oriented instruction from
a second instructor. A second, smaller
comparison group of eight ELC students
(five males, three females) were in a parallel level 5/6 listening-speaking class
taught by a third teacher.
These ELC students came from a variety of native-language backgrounds. The
fourteen students in the primary group
consisted of native speakers of Spanish
(4), Korean (4), Portuguese (2), Japanese
(2), Russian (1), and Chinese (1). The
eight students in the comparison group
were native speakers of Spanish (3),
Portuguese (2), Japanese (1), Korean (1),
and Russian (1).
Instructional Treatment
As noted above, the actual instruction
provided to these ESL students varied
according to the class in which they were
enrolled.
Communicative Pronunciation Teaching
(CPT) Class
The primary group of ELC students
was enrolled in a class that was team
taught by two teachers Monday through
6.
7.
8.
9.
10.
I TS
EFFECTS ON LEARNERS
71
72
I TS
Sentence Reading
The General Linear Model analysis
determined that there were significant
differences in the sentence-reading data
for all three variables-time, group, and
subskill. The time variable compared performances in January and April, and the
difference was significant at the p < .05
level. On this same task, differences
according to group (with three levels)
and subskill (with six levels) were significant at an even higher level (p < .01).
A post hoc analysis of the means and
standard deviations in each of the categories (see results in Table 2) reveals pronunciation improvements for the group
that received communicative pronunciation teaching. From January to April, the
average accentedness rating of students
in the CPT group declined from 6.36 to
6.15 (or .21 points on the nine-point scale)
in subskill area 1, the "Overall
Comprehensibility and Accentedness"
category. (A decrease in score means that
accentedness went down or, in other
words, that pronunciation improved). In
contrast, in this same category, the group
that did not receive communicative
4.15
920.28
4.49
.043
.000
.001
Storytelling
F value Probability
2.45
1150.07
15.95
.118
.000
.000
Passage Reading
F value Probability
2.93
1637.42
9.70
.087
.000
.000
73
74
Times
Groups
1. LIS class
with CPT
N=14
Mean
2. LIS class
without CPT
N=8
Mean
SD
SD
3. Native English
Speakers
N=3
Mean
SO
1. Overall
January
comprehensibility April
and accentedness Difference
6.36
6.15
0.21
0.51
1.00
5.19
5.17
0.02
0.73
0.54
1.00
1.03
-0.03
0.00
0.05
2. Fluency
in general
January
April
Difference
6.26
6.05
0.21
0.77
1.12
5.13
5.10
0.03
0.75
0.60
1.03
1.03
0.00
0.05
0.05
3. Vowel sounds
January
April
Difference
6.45
6.31
0.14
0.44
0.78
5.84
5.51
0.33
0.88
0.35
1.03
1.06
-0.03
0.05
0.05
4. Consonant
sounds
January
April
Difference
6.63
6.49
0.14
0.61
0.72
5.57
5.74
-0.17
0.79
0.71
1.03
1.03
0.00
0.05
0.05
5. Consonant
clusters
January
April
Difference
6.85
6.56
0.29
0.65
0.92
5.91
5.76
0.15
0.92
0.62
1.00
1.00
0.00
0.00
0.00
6. Stress and
intonation
January
April
Difference
6.44
6.09
0.35
0.56
0.84
5.54
5.25
0.29
0.45
0.60
1.08
1.00
0.08
0.08
0.00
0.22
0.11
0.00
Storytelling
For the storytelling task, a somewhat
different picture emerges from the statistical analysis. According to the General
Linear Model statistics, the differences
between January and April scores were
not significant. The group and subskill
variables, however, were still significant,
I TS
EFFECTS ON LEARNERS
Times
Croups
1. LIS class
with CPT
N=14
Mean
1. Overall
January
comprehensibility April
and accentedness Difference
2. LIS class
without CPT
N=8
Mean
SD
SD
3. Native English
Speakers
N=3
Mean
SD
5.86
5.85
0.01
1.17
0.92
5.27
5.07
0.20
0.89
0.84
1.00
1.00
0.00
0.00
0.00
2. Fluency
in general
January
April
Difference
5.68
5.71
-0.03
1.23
.99
5.17
5.14
0.03
0.93
0.96
1.03
1.00
0.03
0.05
0.00
3. Vowel sounds
January
April
Difference
5.80
5.73
0.07
0.85
0.66
5.24
5.23
0.01
0.60
0.62
1.06
1.03
0.03
0.05
0.05
4. Consonant
sounds
January
April
Difference
5.83
5.95
-0.12
0.93
0.75
5.48
5.42
0.06
0.88
0.96
1.00
1.00
0.00
0.00
0.00
5. Consonant
clusters
January
April
Difference
6.14
6.21
-0.07
1.07
0.79
5.73
5.49
0.24
0.88
0.90
1.03
1.00
0.03
0.05
0.00
6. Stress and
intonation
January
April
Difference
6.01
5.94
0.07
1.18
0.96
5.28
5.11
0.17
1.10
0.91
1.03
1.00
0.03
0.05
0.00
7. Reduction and
blending
January
April
Difference
6.92
6.81
0.11
1.14
0.97
6.34
6.25
0.09
1.00
0.77
1.17
1.17
0.00
0.22
0.22
0.01
0.11
0.02
75
76
Times
Groups
1. LIS class
with CPT
N=14
Mean
2. LIS class
without CPT
N=8
Mean
SD
SD
3. Native English
Speakers
N=3
Mean
SO
1. Overall
January
comprehensibility April
and accentedness Difference
5.77
5.55
0.22
0.83
1.23
4.85
4.88
-0.03
0.59
0.35
1.00
1.00
0.00
0.00
0.00
2. Fluency
in general
January
April
Difference
5.41
5.40
0.01
0.84
1.29
4.58
4.75
-0.17
0.38
0.64
1.00
1.00
0.00
0.00
0.00
3. Vowel sounds
January
April
Difference
5.72
5.54
0.18
0.51
0.87
5.00
4.90
0.10
0.74
0.35
1.03
1.00
0.03
0.05
0.00
4. Consonant
sounds
January
April
Difference
5.74
5.43
0.31
0.66
0.92
5.08
5.00
0.08
0.48
0.24
1.00
1.00
0.00
0.00
0.00
5. Consonant
clusters
January
April
Difference
6.08
5.71
0.37
0.69
0.92
5.32
5.27
0.05
0.49
0.30
1.00
1.00
0.00
0.00
0.00
6. Stress (word
and sentence
level)
January
April
Difference
5.90
5.81
0.09
0.81
1.20
5.08
5.23
-0.15
0.34
0.41
1.00
1.00
0.00
0.00
0.00
7. Intonation
January
April
Difference
5.86
5.87
-0.01
0.79
1.16
5.02
5.28
-0.26
0.50
0.59
1.03
1.00
0.03
0.05
0.00
8. Reduction
and blending
January
April
Difference
6.58
6.45
0.13
0.76
1.29
5.65
5.75
-0.10
0.79
0.23
1.00
1.00
0.00
0.00
0.00
9. Segmentation
(appropriate
pausing)
January
April
Difference
5.97
5.71
0.26
0.72
1.30
4.88
5.05
-0.17
0.57
0.79
1.03
1.00
0.03
0.05
0.00
0.17
-0.07
0.01
ing of it. It is difficult to rate pronunciation when listening to storytelling sampIes because the language of each story is
original and varies from speaker to
speaker in content, structure, and lexicon.
It is easy for raters to get distracted by
what the speaker is saying or by grammar or vocabulary mistakes. In this
regard, our observation, which may
benefit future researchers, is that storytelling is not the best task for evaluating
pronunciation improvement.
Passage Reading
The third speaking task used in this
study-but which was not used by
Derwing, Munro, and Wiebe (1998) in
their research-was passage reading. The
results here (see Table 4) were similar to
those produced by the sentence-reading
task. The General Linear Model analysis
of variance found that there were differences according to time, but these were
significant only at the p < .1 level (mean-
I TS
EFFECTS
()\J
77
LEARNERS
Sentence Readin~
Class with
CPT
Jan.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
5.61
4.69
6.19
5.32
6.21
6.36
6.94
6.56
7.06
7.40
6.20
6.43
7.15
7.53
6.39
6.64
6.85
6.38
6.32
5.06
6.47
6.35
6.38
6.04
7.56
7.03
5.85
5.92
Mean Difference
Class w/out
CPT
1
2
3
4
5
6
7
8
Apr.
Jan.
Apr.
Diff.
0.92
0.87
0.15
-0.38
-0.34
-0.23
-0.38
-0.25
0.47
1.26
0.12
0.34
0.53
0.07
0.23
Diff.
6.26
5.96
0.30
3.94 4.86 -0.92
5.94
5.86
0.08
5.57 5.74 -0.17
5.67 -0.34
5.33
5.90
4.90
1.00
0.07
5.67 5.60
0.82
5.61
4.79
Mean Difference 0.11
Passa~e Readin~
Storytellin~
Jan.
Apr.
5.43
5.55
4.64
5.30
4.60
5.58
7.23
7.07
7.09
6.65
6.62
6.52
6.44
7.10
6.33
5.86
5.77
6.15
4.65
6.25
7.15
6.82
5.69
5.92
7.20
6.06
4.75
4.24
Mean Difference
Jan.
Apr.
Diff.
-0.12
-0.66
-0.98
0.16
0.44
0.10
0.66
0.47
-0.38
-1.60
0.33
-0.23
1.14
0.51
-0.01
Diff.
5.17
5.31 -0.14
4.51
4.19
0.32
6.55
6.55
0.00
5.44
5.63 -0.19
4.55
4.95 -0.40
5.07
4.89
0.18
6.85
0.29
6.56
5.88
5.02
0.86
Mean Difference 0.12
Jan.
Apr.
5.20
4.53
4.35
3.67
7.00
5.69
6.19
5.79
6.76
7.05
5.88
5.56
7.12
6.97
6.17
5.57
5.96
6.42
5.79
4.70
5.93
5.76
6.26
4.44
6.28
6.95
5.64
4.94
Mean Difference
Jan.
Apr.
Diff.
0.67
0.68
-1.31
-0.40
0.29
0.32
0.15
-0.60
-0.46
1.09
0.17
1.82
-0.67
0.70
0.18
Diff.
4.66
4.96 -0.30
5.69 -0.11
5.58
5.22 -0.79
4.43
0.61
5.44
4.83
4.93
0.23
5.16
Not available
Not available
Not available
Mean Difference -0.07
78
Individual Variation
In their research on the effects of different types of pronunciation teaching,
Yule and Macdonald (1995, 346-48)
found a wide range of individual student
reactions ranging from improvement in
pronunciation to deterioration (see also
Yule, Hoffman, and Damico 1987). Other
researchers and teachers (Miller 2000)
have also discovered that the relationship
between instruction and improvement is
not a simple, straightforward one.
Leather (1983, 205) explains that pronunciation change is a complicated process
involving many variables, such as age
and attitudes.
Looking at individual student scores
in our research (see Table 5), we see a
similar pattern-some individuals made
much more progress than others. There
was a fairly consistent and dramatic
increase in standard deviations (or
spread of scores) in the CPT group from
January to April. A similar pattern cannot
be found in the comparison group. This
phenomenon suggests that communicative pronunciation teaching benefitted
some students more than others, thus
spreading their scores more.
The class that received communicative pronunciation training showed a
large increase in the standard deviation
for nearly every pronunciation subskill
for both the sentence reading and passage reading tasks (see Tables 2 and 4). In
many cases, the standard deviation went
up by half or more and in a few cases it
nearly doubled. For about a third of the
CPT students, the score increases were
dramatic (close to half a point or more).
These students improved much more
than the class mean (.23 for sentence
reading, .18 for passage reading). On the
I TS
EFFECTS ON LEARNERS
79
80
REFERENCES
Acton, William. 1984. Changing fossilized pronunciation.
I TS
EFFECTS ON LEARNERS
(2): 145-58.
Scarcella, Robin
c.,
(2): 117-22.
81