First of all, for the first research question, one-way repeated ANOVA was performed on the results of the elementary GEPT listening test scores to determine whether there were significant differences among the three groups on their listening comprehension performance after the intervention. Secondly, the listening logs from autonomous listening groups were analyzed both quantitatively as well as qualitatively in order to understand the influences of autonomous listening.
Concerning the listening log, in the first section ‘Today’s Achievement’, the total of the hours the participants listened in class were kept as a baseline to see whether the participant was a faster or a slower listener. Next, the contents from the second section ‘Language Learning and Reflection’ were the most important sources for the second research question.
What they ‘picked up’ in their listening experiences was categorized into five areas, and the tally of them provided insight for what they thought they had learned. The problem types along with the learners’ own solutions or desired help from the teacher were also scrutinized according to the quality and the quantity of the details they documented, which suggested how they developed autonomy along the way. The third section ‘Listen after Class’ was a counterpart of ‘Today’s Achievement’. The total of listening hours at home was
accumulated to determine whether the learners extended their learning out of campus. As for the fourth part ‘Opinions about Autonomous Listening’, the views on or attitudes towards this form of listening practice were collected to understand their general perception of this listening program.
Lastly, the responses from the interviews were firstly transcribed and then inspected to compare: what the two experimental groups had in common as and what they thought differently. Both the shared and diverged opinions were discussed to discover the acts of teacher guidance and to explore the meaning of teachers’ existence.
CHAPTER FOUR RESULTS and DISCUSSIONS
Chapter Four is divided into three sections. Section 4.1 indicates the outcomes of the listening comprehension tests after the ten-week intervention with a view to answer the first research question: whether there are any significant differences in listening ability when learners received different listening instruction. The second research question—what are the perceptions of the listening program—involves complicated cognizance, so it is explored in two sections. Section 4.2 exhibits the qualitative data that documented in listening logs and Section 4.3 demonstrated the opinions of the functions of the teacher in the follow-up
interviews; both are followed by related discussions.
4.1 Effectiveness of Autonomous Listening
4.1.1 Performance of the Listening Comprehension Tests
The outcomes of the pretest and the posttest of GEPT listening comprehension tests are presented below in Table 4.1. The GAL group scored 12.64 in the pretest with the SD of 4.29, and 15.06 in the posttest with the SD of 4.39; the gain is 2.42 points. The AL group scored 13.43 in the pretest (SD=4.52) and 15.14 in the posttest (SD=5.40); the gain is 1.71 points. The control group scored 13.66 in the pretest (SD=4.91) and 15.35 in posttest (SD=5.82); the gain is 2.11 points.
Table 4.1
Descriptive Statistics of Listening Tests for the GAL, AL and Control Groups
Group Pretest Posttest Gains
M (SD)
CI 95% M
(SD)
CI 95%
LL UL LL UL
GAL
(n=33) 12.64 (4.29) 11.05 14.22 15.06 (4.39) 13.25 16.87 2.42
AL
(n=35) 13.43 (4.52) 11.89 14.97 15.14 (5.40) 13.39 16.90 1.71
Control
(n=32) 13.66 (4.91) 12.05 15.26 15.35 (5.82) 14.04 17.71 2.11
Table 4.2
One-way repeated ANOVA for Effects of Listening Instruction on Listening Comprehension
Sources SS df MS F p
Between subjects 27.34 2 13.67 .317 .729
Within subjects 4177.76 97 43.07
Total 4205.10 99
In order to determine the effect of listening instruction on learner’s listening ability, repeated one-way ANOVA was applied to analyze whether there was a significant difference among the three groups in their performance in the listening tests after ten weeks. The alpha level was set at .05. As can be seen in Table 4.2, the F value is .317, and p value is greater than 0.05. From the results of Tables 1 and 2 together, we learned that although the average gain of the GAL group (2.42) was slightly larger than the average gains of the AL group (1.71) and the control group (2.11) respectively, the analysis of ANOVA reveals that the difference among the three groups was not significant (F(2.97)=0.317, p>.05).
4.1.2 Discussion on the Effectiveness of Autonomous Listening
Based on the results from the pretests and the posttests presented above, it is discovered
that although the three groups showed significant improvement in their listening performance individually, there was no significant difference in their gains for receiving different listening instruction. Verdugo and Belmonte (2007) have similar findings: the experimental group that listened to audiobooks did pick up vocabulary and could answer the questions correctly despite fairly significant differences from the control group. This phenomenon seems to illustrate Renandya and Farrell’s (2010) assertions that easy and relaxing listening
experiences could work as well as formal listening strategy training. Still, three possible reasons for no significant differences could be scrutinized.
To begin with, the three groups were not exactly homogeneous in their composition.
On the basis of their pretest results, their average scores were not considerably varied, but the differences within subjects were quite observable. We found that the numbers of weaker listeners (those who scored between 6-15 points) were 27 and 25 for the AL group and the control group respectively, while the number was 29 for the GAL group. Moreover, the number of better listeners (those who scored more than 20) was 5 for both the AL group and the control group, but only 2 for the GAL group. That is to say, the GAL group had more lower-achievers and fewer higher achievers compared to the other two groups, which may not generate broad progress scope in their average listening performance.
Nevertheless, it was observed that in the posttests the number of less skilled listeners in the GAL group decreased 75 percent (from 8 to 2), more than both the AL group and the control group (from 8 to 6 and from 10 to 6 respectively). The gap of the standard deviations only widened .10 for the GAL group while it extended .88 and .91 for the other two groups, which meant that after the ten-week guided autonomous listening program, the degree of discrepancy changed slightly for the whole class. This phenomenon that teacher guidance facilitated in the enhancement of weaker learners seems to collaborate with the Vandergrift & Tafaghodtari’s (2010) and Cross’s (2011) findings in which teacher instruction,
or the provision of listening strategies, has a stronger positive impact on poorer listeners than on skilled ones.
The second reason may be attributed to the error or incomprehensiveness resulted from a single posttest to determine the progress in listening. In the retrospective interviews, one of the participants admitted that he did not feel like taking the test, saying “I listened carefully to every question for the first test, but for the other I felt sleepy, in bad mood...and I didn’t feel like listening.” Another interviewee stated that he spent much time keeping thinking the answer to a certain question. “It ended up that I had no choice but rush an answer since I totally missed the next question,” he said. By contrast, the training of pre-listening activities appeared perfectly useful for the control group to overcome their test-taking hitches, which was beneficial in improving their performance on the tests. In Chang and Millett’s research (2013), the question types in the posttests were consistently presented in every session for students. The know-how with the test types could successfully reduce the chances of errors caused by unfamiliarity in taking a test, which may imply the indispensability of familiarizing learners with test situations. Therefore factors like unstable personal physical conditions or the lack of practicing test-taking skills could both affect the validity of a test.
Thirdly, the fruit of autonomous listening may need longer time to yield even though ten weeks1 were acceptable compared to previous studies. However, considering the interaction of the nature of listening as well as the nature of autonomous learning, ten weeks could appear insufficient, as stated in Verdugo and Belmonte’s (2007) investigation, which proceeded for twenty-two weeks. For one thing, language competence is not to be developed overnight;
successful learners tend to devote themselves in language learning for quite a long period of time. For another, autonomous listening expected learners’ escalation in autonomy so that
1 Note: Listening instruction lasted eight weeks for Goh & Taib (2006) and Kemp (2010), thirteen weeks Chang
& Millett, (2013) and Vandergrift & Tafaghodtari (2013), and twenty-two weeks for Chang (2011) and Verdugo
& Belmonte (2007).
they would be more focused in class and more motivated in pursuing exposure to the target language outside the classroom, which also requires abundant time to establish or to adjust according to idiosyncrasy of each language learner.
Meanwhile, autonomous learning has much to do with the learning context (Allwright, 1988; Dickson, 1987; Benson, 2001), ranging from a microscopic classroom setting to a national educational system. Concerning the classroom environment, a high
student-to-teacher ratio (around 34:1) could lead to the inability for instructors to take care of everything, and individual advice might thus not be easily given (Verdugo & Belmonte, 2007).
Additionally, many participants of studies on autonomous learning were tertiary students, who were allowed to have fewer class hours and thus more flexibility in their study plans, and were able to afford more time in pursuit of further engagement in learning. By contrast, most of the secondary school learners in Taiwan have to spend over nine hours at school. At the same time, the trend of attending cram schools is almost universal, which could compress the length of voluntary devotion to learning after school. Under such a circumstance, the power of autonomous listening could be compromised to a considerable degree.
4.2 Perceptions of the Program 4.2.1 Results from the Listening Logs Self-perceived Acquisition
The linguistic gains from listening activities were recorded by the participants
themselves without any advance-appointed minimum of recording required by the teacher.
Table 4.3 shows the frequency and percentage of the learning points from each group’s voluntary record.
Table 4.3 The Learning Points Recorded
Learning points GAL AL Control
N % N % N %
Pronunciation 136 18% 217 19% 105 24%
Intonation 166 22% 227 20% 98 22%
Vocabulary 291 39% 511 45% 110 25%
Phrases/Expressions 148 20% 146 13% 61 14%
Grammar 12 2% 40 4% 47 11%
Others 2 0% 0 0% 23 5%
TOTAL 755 100% 1141 100% 444 100%
As presented in Table 4.3, we found that the amounts of the participants in each group recording what they thought they had learned has shown some inconsistency. The
autonomous listening (AL) group has documented more than a thousand items during the listening sessions, the guided autonomous listening (GAL) group more than seven hundred, and the control group less than five hundred. That is, the AL group recorded the most tallies of learning items, approximately as 1.5 times as the GAL group did, and even as 2.5 times as the control group did.
Nevertheless, with regard to what linguistic features the participants thought they had acquired, a strikingly similar pattern between the two autonomous listening groups is
demonstrated. Both GAL and AL groups marked new vocabulary as their biggest gain in the autonomous listening activities while they learned least on grammatical rules. Intonation as
well as pronunciation accounts for equally matched amount, with 40% in total. A major distinction would be that the AL group seemed to pick up more single words along the intervention while the GAL group absorbed more chunks or common expressions.
Compared with the two AL groups, the control group acquired approximately the same amount on pronunciation and intonation as they did, but the percentage of vocabulary is far less (20% lower than the AL group and 13% than the GAL group), and the percentage of learning grammar is about seven percent higher. Another five percent of acquisition in their listening course marked by the control group are mostly test-taking skills, such as note taking or making prediction based on pictures or questions.
Self-perceived Listening Difficulties
Just like the linguistic gains from listening activities, the encounters of difficulties were recorded by the participants themselves with a minimum of one entry set by the teacher; they noted down things they wanted to note. The frequency and percentage of various forms of their observed problems were displayed in Table 4.4.
The number of the total recorded listening problems from the three groups did not vary greatly, yet the AL group had the most tallies. During the ten weeks, it is discovered that the distributions of difficulty types between the GAL and AL groups look relatively similar: item B Unable to form a mental representation from words heard takes the largest proportion (more than 30%), secondly item A neglect the next part when thinking about meaning (more than 20%), item C do not understand of subsequent part due to earlier questions accounts for 14%, and item E confused about the key ideas in the message takes about 10%. Item D understand words but not the intended message takes 11% in the GAL group and 16% in the AL group.
Table 4.4 The Listening Problems Recorded
Listening Problems GAL AL Control
N % N % N %
A. Neglect the next part when thinking
about meaning
114 27% 101 21% 80 20%
B. Unable to form a mental
representation from words heard
134 32% 176 37% 86 22%
C. Do not understand of subsequent
parts of input due to earlier problems
58 14% 66 14% 60 15%
D. Understand words but not the
intended message
46 11% 74 16% 57 14%
E. Confused about the key ideas in the
message
43 10% 42 9% 33 8%
F. Others 23 6% 16 3% 80 20%
TOTAL 418 100% 475 100% 396 100%
On the other hand, the distribution of difficulty types of the control group is obviously different from the two experimental groups. While the rates of items B, C, D and E did not differ greatly compared to the AL group and the GAL group, they seemed to encounter fewer challenges in Item B Unable to form a mental representation from words heard (22%) in comparison with the two AL groups (32% and 37% respectively). Next, they reported an unignorable number of other types of difficulties, which accounts for one fifth. These
difficulties raised by the participants in the control group include getting distracted or lost during listening, not being able to catch up with the speed of delivery, and misspelling or not being spelling words etc.
Listening Hours after Class
In addition to the different distributions of the two AL listening groups in terms of their self-recorded linguistic improvement and troubles in listening, their voluntary listening practice at home also diverged, as we can see from Figure 4.1. In the first week, the
participants from both the GAL group and the AL group listened less than fifty minutes. For the GAL group, however, after the second week the listening hours has shown a steady
growth until the sixth week, and even a rapid growth to the ninth week though followed by a sharp drop in the last week. On the contrary, although the total of listening hours for the AL group surpassed that for the GAL group in the second week, there was only a mild increase before the fifth week, followed by a steady decrease to the last week.
Figure 4.1 After-class Listening Hours during the Ten-Week Intervention
0.
125.
250.
375.
500.
625.
1 2 3 4 5 6 7 8 9 10
Minutes
Listening hours a.er class
Guided AL AL
4.2.2 End-Semester Feedback after the Autonomous Listening Program
The open-ended feedback after the ten-week autonomous listening program from the experimental groups has been sorted by the teacher-researcher into three categories:
advantages, disadvantages, and suggestions for the teacher. Since the participants were simply required to give responses on what came to their mind first, the results would be focused in the count of the responses rather than the percentage. The results of advantages are summarized in Table 4.5.
Table 4.5 Advantages Proposed by AL and GAL Learners in the Feedback
Advantages
AL (N=35) GAL (N=33) Total (N=68)
N % N % N %
Speed of delivery 6 18% 4 11% 10 15%
Pronunciation/intonation 4 12% 6 17% 10 15%
Vocabulary/expressions 12 35% 11 31% 23 34%
Listening 12 35% 17 49% 29 43%
Speaking 4 12% 2 6% 6 9%
Right to choose 5 15% 4 11% 9 13%
Enjoyment 9 26% 5 14% 14 21%
Motivation increase 2 6% 6 17% 8 12%
As can be seen in Table 4.5, the first and foremost idea that came to their mind is listening (with a sum of 29 counts), ensued by words/expressions (23 counts). The
participants also implied the specific benefits in listening abilities, such as speed of delivery
(10 counts) and intonation or accent (10 counts). Moreover some of them indicated the benefits with regard to autonomous learning contain enjoyment (14 counts), right to choose (9 counts), and motivation increase (8 counts). Another thing worth noting is that 6 of them indicated feeling improvement in speaking skills.
Despite that the participants stated many merits of autonomous listening, the diagrams below would only center on ‘listening’ and ‘autonomy’. Table 4.6 displays the quotes describing the assets of autonomous listening on listening: general listening improvement, following speech delivery, pronunciation and intonation, and vocabulary and expressions.
Since some of the quotes actually involved with various facets in learning listening, they were labeled to a single category in order to keep the intactness of the original version. Table 4.7 presents the quotes describing the values of autonomous listening on autonomy, i.e. pleasure, flexibility, and growth in motivation.
Table 4.6 The Advantages of Listening Learning from Learners’ Quotes General Listening Improvement
l “I remember the first time I listened I had much trouble understanding, but as time went by I found there’s less and less that I didn’t understand.”
l “Thanks you for providing this listening program; I sincerely considered my listening improved.”
l “This ten-week listening program helped me improve a lot in the midterm listening test!”
l “I found that I can understand more of the English radio program every Thursday morning, and sometimes I can even understand videos without Chinese subtitles! I really hope we can spend more time [on this].”
l “I could have one session relaxing myself and found my listening improved.”
Following Speech Delivery
l “I think the best thing in this program is that I could catch up with the fast speed a little bit, at least the vocabulary in my head wouldn’t be completely disconnected to what I heard.” (speed)
l “[This listening course] enabled me to catch up the speed [of speech] as well as roughly get the meaning of sentences.” (speed)
Pronunciation and Intonation
l “During this program, I think my biggest gain is getting familiar with native speakers’ pronunciation and intonation, making me understand what they are talking about more easily in the future.” (pronunciation and intonation)
l “We can learn new words or sentences that we couldn’t learn from our textbooks, and get familiar with how they [native speakers] speak English.”
(intonation)
l “[This listening course] helped me learn how to pronounce some vocabulary.”
(pronunciation)
Vocabulary and Expressions
l “We could learn some words and expressions with interesting animation along with English pronunciation. Although some videos are fast [in speech delivery] and some words are difficult, listening is more important.” (words and expressions)
l “… Also if there were some interesting sentences I would pick them up.”
(expressions)
l “Whenever I watched TV or listened to music, I found familiar words easily.”
(words)
Table 4.7 The Advantages of Autonomous Learning from Learners’ Quotes Enjoyment
l “This activity is amazing; it makes me more comfortable than general listening practice.”
l “And it’s not rigid as common listening practice.”
l “I could enjoy what I like and enhance my (listening) ability at the same time.
It’s just like killing two birds with one stone.”
l “This approach is fun, not boring at all.”
l “And I’m starting to consider English wasn’t as boring and rigid as I thought before.”
Flexibility
l “This approach allows advanced classmates learn more, and those weaker can learn at a slower pace.”
l “We can find something interests us and suits our levels.”
l “I think this approach is excellent because I can learn what I want to listen and control the speed myself. Since my English is relatively poor, I can choose something easier to listen to, progressing little by little. I can replay the parts I don’t understand by myself.”
Motivation Increase
l “Choosing the videos we like increased my motivation to learn English.”
l “This approach made me love English more. I hope we can still have this program in our third grade.”
l “Listening is fun: we can watch interesting animation as well as nice English songs. I would try hard to listen more.”