Results and Discussion
This chapter will present the results based on the research questions proposed in chapter one. In addition, the findings in the present study would be compared with those of the former studies.
Results
Research Question 1
To what extent does task-based instruction make a difference in
vocabulary learning?
After the intervention of task-based instruction (TBI), the performance of vocabulary learning in the posttest between the control and experimental groups was compared. When Independent-Samples t-test was conducted, Levene’s Test was F(2, 65) = 3.45, p = .07 which didn’t reach the specified .05 significance level, i.e., the two groups were homogeneous group. Results in the posttest showed that the mean scores in the experimental and control groups were 9.03 and 6.38 respectively. The t value was t (65) = 2.36, p = .02, indicating there was significant difference on vocabulary performance between two groups. In addition, the standard deviations in the
experimental and control groups were 5.50 and 3.69 respectively; the minimum score in the control group was 2.00 and the experimental group one was 3.00; the maximum
score in the control group and experimental group was 19.00 and 25.00 respectively.
The score distribution of the minimum and maximum scores in two groups was wide, revealing large performance range. Means, standard deviations, minimum, and
maximum values regarding students’ performance between two groups were shown in Table 6.
Table 6
Descriptive Statistics of the Posttest Between Two Groups
Groups n M SD Min. Max. t df Sig.
Control 37 6.38 3.69 2 19 2.36 65 .02*
Experiment 30 9.03 5.50 3 25
Note. *p < .05. **p < .01.
After the TBI intervention, the Paired-Samples t-test was conducted to examine whether there were differences within the experimental group on vocabulary
performance. Results showed that the experimental group adopting TBI method got higher mean score in the posttest and there were significant differences between the pretest and posttest. Thus, in the experimental group, the t value was t (29) = 3.30 and
p < .01, indicating there was significant performance difference between pretest and
posttest. Besides, in the pretest, the mean score in the experimental group was 6.47
with SD = 2.33. In the posttest, the mean score in the experimental group was 9.03 with SD = 5.50. In addition, the distance of the gain score between pretest and posttest was 2.57, indicating participants who have made a big progress in their posttest. The scores in the pretest (Appendix G) and posttest (Appendix H) depicted the
performance of the experimental group. The following Table 7 has shown the descriptive statistics of the pretest and posttest in the experimental group.
Table 7
Descriptive Statistics of the Pretest and Posttest in the Experimental Group
Tests n M SD Gain
Score
t df Sig.
Pretest 30 6.47 9.03 2.57 3.30 29 .00**
Posttest 30 2.33 5.50
Note. *p < .05. **p < .01.
On the other hand, the Paired-Samples t-test in the control group showed that there was no significant difference between pretest and posttest based on t (36) = -2.07, p = .05. Regarding the mean score in the pretest, it was 7.65 with SD = 2.50; in the posttest, the mean score was 6.38 with SD = 3.69. Besides, the distance of the gain score between pretest and posttest was -1.27. Obviously, participants in the control group performed less well in their posttest than pretest. The following Table 8 was the descriptive statistics of the pretest and posttest in the control group.
Table 8
Descriptive Statistics of the Pretest and Posttest in the Control Group
Tests n M SD Gain
Score
t df Sig.
Pretest 37 7.65 2.50 -1.27 -2.07 36 .05
Posttest 37 6.38 3.69
Concerning the aforementioned posttest results, it could be concluded that in traditional lecture classes teachers spent a lot of time on teaching English without peer interaction activities involved. It was a kind of teacher-centered English class without learners’ interaction and practice. This could be a reason that learners didn’t perform well on the posttest. Especially, learners in most of their time in the classroom were used to sitting quietly to listen to the teacher’s teaching and passively for learning.
Viewing from learners’ mean scores between pretest and posttest, traditional lecture teaching couldn’t improve learners’ vocabulary performance, nor did it help learners to learn vocabulary. On the contrary, TBI was a learner-centered class and learners were provided more opportunities to speak, discuss, and practice target language. In particular, they could interact with their group members freely even though they couldn’t speak or pronounce the target language well. Consequently, TBI has brought participants significant learning effects in vocabulary learning which was distinctly different from traditional lecture instruction.
Research Question 2
In which aspects of task-based instruction help students in vocabulary
learning?
With the results that TBI group outperformed its counterpart group, in order to further examine which aspects of TBI help students in vocabulary learning, students’
performance on the posttest was examined. Results showed that there was significant difference in multiple-choice vocabulary aspect but there were no significant
differences on part of speech and cloze aspects in the posttest after the intervention of the TBI. In the multiple choice performance, it was t (65) = 3.31, p < .01; part of speech was t (65) = 1.04, p > .05; cloze was t (65) = .61, p > .05 respectively. Means, standard deviations, minimum, and maximum values toward students’ performance between two groups were shown in Table 9.
Table 9
Descriptive Statistics of the Posttest Between Control and Experimental Groups
Groups n M SD Min. Max. t df Sig.
The results of Q2 showed that there was significant difference in
multiple-choice vocabulary performance after the intervention of the TBI because TBI brought learners the effects of vocabulary learning. However, there were no
significant differences on part of speech or cloze items, but the performance in the experimental group regarding part of speech and cloze items was still far better than those in the control group. Since the experimental group’s performance was better than control group, in order to further understanding in which aspects did TBI make a difference within the experimental group, the performance in the pre- and posttest was examined. When Paired-Samples t-test was performed to examine the participants’
performance, the results showed that in multiple-choice vocabulary aspect t (29) = 3.04, p < .05, showing there was significant difference when students learned vocabulary by using TBI method. Next, the results in the part of speech aspect showed that the t (29) = 1.84, p > .05, indicating there was no significant difference between the pretest and posttest in the experimental group. In other words, there was no difference in terms of vocabulary performance through TBI approach;
Nevertheless, learners in the posttest got more mean score 1.07 than the pretest one only .67. Lastly, the results in the cloze aspect disclosed that the t value was t (29) = 2.28, p < .05, revealing there was significant difference between the pretest and posttest in the experimental group. As to the gain scores, the distances in the three sections (multiple-choice, part of speech, and cloze items) were 1.43, .40, and .73
respectively; it showed that no matter in which section participants in the
experimental group they have made a progress in their vocabulary learning after the treatment of TBI. Overall, results revealed that the intervention of the TBI brought learners significant vocabulary learning difference in multiple-choice and cloze aspect in the experimental group. Means, standard deviations, and p values between the pretest and posttest in the experimental group were shown in Table 10.
Table 10
Descriptive Statistics of the Pretest and Posttest in the Experimental Group
Groups n M SD Gain
Moreover, Paired-Samples t-test was also conducted in the control group. The results showed that there were significant differences between pretest and posttest in multiple-choice and part of speech aspects but indicated the negative relationship no matter in multiple-choice (t (36) = -2.56, p = .02) or part of speech (t (36) = -3.12, p
= .00) aspects in the posttest of the control group. The results displayed that only
mean scores in multiple choice, part of speech, and cloze aspects were 5.16 with SD = 1.75, 1.51 with SD = 1.19, and .97 with SD = 1.21 respectively; the posttest mean scores were 4.11 with SD = 2.33, .73 with SD = 1.24, and 1.54 with SD = 1.48 respectively. Concerning the gain scores, participants in the posttest didn’t perform well in their posttest, especially in multiple-choice and part of speech aspects.
Between pretest and posttest the gain score distances in multiple-choice and part of speech aspects were -1.05 and -.78 respectively. The gain score only in cloze aspect performed better than the former mentioned two aspects; the distance between pretest and posttest was .57, i.e., participants in cloze aspect have made a progress in their posttest when they were conducted by using traditional lecture instruction. The following Table 11 was the descriptive statistics between the pretest and posttest in the control group.
Table 11
Descriptive Statistics of the Pretest and Posttest in the Control Group
Groups n M SD Gain
In conclusion, the Paired-Samples t-test showed that TBI method has brought learners the learning effects in multiple choice and cloze aspects in the experimental group. It was different from control group’s performance that participants in cloze aspect performed better than that in multiple-choice and part of speech aspects.
Overall, when experimental group was compared with control group by using
Independent-Samples t-test, there was significant difference on multiple-choice but no significant differences in both part of speech and cloze parts after the intervention of the TBI between the two groups. Some possible reasons were delineated as follows.
The first reason was related to the participants who did not learn English in TBI method before. Therefore, some of the participants seemly couldn’t comprehend and listen to the teacher’s indications carefully, so the effects of the task-based instruction were limited. Observing from the participants’ group discussion tasks, the researcher found that some of the participants among different groups forgot the importance of group work sometimes, so they didn’t work very hard when they were participating in their discussion activities; some of the group members in different groups kept
chatting with their group members or kept silence at times. The second reason was related to the three tasks in the present study in which every group member had to play different roles according to the role assignment each time. Every group member should be responsible for certain their group work based on their roles and some of
the group members felt that their own task scope was too difficult to handle. Thus, they couldn’t perform well when undertaking their tasks; some of them in different groups delayed their progress in which their groups couldn’t finish their tasks as expected. Therefore, their performances in the posttest were influenced.
The third reason was related to the skills’ training such as the use of translation (Grace, 1998; Hunt & Beglar, 2002; Laufer & Shmueli, 1997; Liao & Chiang, 2003;
Prince, 1996), dictionary (Grabe & Stoller, 1997; Hsieh, 2000; Hu, 2006; Nation, 2002b; Summer, 1989; Yang, 2005), and guessing (Carter & McCarthy, 1991; Nation, 2002a; Nation & Coady, 1991; Nattinger, 1989; Oxford, 1990; Schmitt, 2000). The participants maybe were not familiar with the translation structure and they had no idea on which meaning was the appropriate one when they were looking up the target words in the dictionary; nor could they guess cloze target words or part of speech of words wisely. That is to say, it was probably related to the participants’ insufficient vocabulary knowledge (Huang, 2004; Laufer, 1997), insufficient grammar ability (Bensoussan, 1983), and unskillful strategy usage. Besides, observing from the participants’ frequent looking up dictionaries, the researcher found it seemly revealed that learners’ vocabulary size was not enough because they didn’t know a lot of basic and frequent occurring words in reading texts.
The fourth reason was related to the self-confidence and autonomy of learners.
During the task-based instruction intervention, the teacher requested those participants to write down what the teacher taught them each time and hoped they would review what the teacher has taught them after class but they couldn’t do well. Since the participants were low-level learners, based on two vocabulary tests, classroom observation, and school English teachers’ statement for their English learning performance, they lacked of self-confidence to learn English well even though the teacher and their school English teachers often encouraged them. Besides, during the intervention of TBI, the teacher requested every group to write down their answers after they discussed their tasks. Most of the groups have written down their sentence translation in multiple-choice vocabulary task, but they were not confident in writing down their discussion results of reading texts’ translation such as in part of speech and cloze procedure tasks. The researcher found part of speech reading (Soden Hensler &
Bech, 2006) and cloze procedure (Kobayash, 2002) in reading texts seemly too difficult for them because the sentences in the contexts were longer than the single sentence in multiple-choice vocabulary task. Namely, students could handle
individual sentence structure, but they had difficulty in comprehending longer texts.
Research Question 3
To what extent does task-based instruction make a difference in receptive
and productive vocabulary learning?
Due to the results that TBI group performed better than its counterpart group, in order to further explore whether there was a significant difference in receptive and productive vocabulary learning between experimental and control groups, students’
performance on the posttest was examined. Independent Samples t-tests were carried out to examine the differences of vocabulary learning between the receptive
(multiple-choice) and productive (part of speech and cloze) aspects. Results displayed that the experiment group outperformed than its counterpart group in receptive aspect but not in productive aspect, p < .01 and > .05 respectively, indicating there was significant difference on receptive aspect but not in productive aspect. Means, standard deviations, minimum, and maximum values forward students answering accurately between two groups were shown in Table 12.
Table 12
Descriptive Statistics of the Posttest Between Two Groups
Groups n M SD Min. Max. t df Sig.
From Table 12, the results of the performance of productive vocabulary learning in the posttest showed that there was no significant difference between
and productive aspects, results showed that TBI brought more effect on receptive than productive vocabulary knowledge.
After the TBI intervention, the posttest scores in receptive and productive aspects between control and experimental groups had some differences. The
experimental group adopting TBI method obtained more mean score than the control group adopting traditional lecture instruction did. With regard to the participants of the control group in the receptive vocabulary learning aspect, multiple-choice vocabulary item, the mean score in the posttest was 4.11 with SD = 2.33; the
experimental group’s mean score was 6.17 with SD = 2.77. Also, the minimum score in the control group was .00 and the experimental group one was 3.00; the maximum score in the control group was 10.00 and the experimental group one was 13.00. As to the participants of the control group in the productive aspect like part of speech and cloze procedure items, the mean score in the posttest was 2.27 with SD = 2.13; the experimental group one was 2.87 with SD = 3.10. The minimum scores in both groups were .00; the maximum score in the control group was 9.00 and the experimental group one was 12.00. Generally speaking, the distributions between minimum and maximum scores in both groups were wide, no matter in receptive or productive vocabulary learning aspects.
In terms of the above mentioned posttest results in receptive and productive
aspects, the researcher concluded some possible reasons about why there were significant learning effects in receptive aspect but not in productive aspect after the intervention of TBI. First of all, understanding a word was easier than producing a word (Nation, 1990), so receptive vocabulary learning was easier than productive vocabulary learning. The participants could choose the answers from the
multiple-choice questions and they didn’t need to write down the target words. All they had to do was to find out the correct answer from the four options in every multiple-choice question. In fact, Hughes (2003) indicated that learners’ scores were maybe from guessing when they were answering multiple-choice questions. Therefore, the various test outcomes from the test-takers would be unpredictable. Additionally, multiple-choice questions tested learners’ passive vocabulary (Madsen, 1983), but not active vocabulary; therefore, learners could find out the answer from sentential or contextual meanings which would be easier than they need to produce or write down target words. Next, productive vocabulary learning was more difficult than receptive vocabulary learning (Mondria and Wiersma, 2004). In the current study, the
participants had to write down their answers in blanks according to the contextual clues to decide which target answer was appropriate. Even though productive vocabulary learning in cloze procedure item has provided a target word list for the participants to find out one of the correct answers and write down the target word in
an appropriate blank, they still couldn’t make the right decision to write down or spell the correct target word in correct blank. The reason was when they were reading their cloze text that they had to possess vocabulary knowledge and grammatical
competence simultaneously; however, they couldn’t comprehend the contextual meanings and determine the proper target words to fill in the blanks respectively in the cloze context. Besides, when the participants decided one of the proper target words as their answer from the word list, they still needed to think about whether they needed to change the word’s form before they wrote down and filled in every blank in the cloze reading text. Also, productive vocabulary learning in part of speech item was somewhat difficult, too. Observing from the activity of task-based instruction, the researcher found that the participants in part of speech task still had a big problem that they couldn’t conquer. That is, they were not familiar with words’ family, affixes, grammatical categories, and sentence structures, even though the teacher in classes has taught them those concepts. Consequently, when they were looking up dictionary, they tended to choose the wrong words for both in part of speech and filled in the blanks. During the TBI classes, learners could not absorb the knowledge completely and promptly concerning what the teacher taught them and the teacher had to often repeat what she said. Maybe the task of part of speech of words which the teacher designed was too difficult to them, since the participants who were low-level learners.
And a major reason may be related to what Kojic-Sabo and Lightbown (1999) said that it really took time for learners to get considerable language knowledge or skills in L2. After all, learners in lexical acquisition aspect had to spend more time in order to use it.
Moreover, learners should spend more time in productive aspect when they were learning target words (Hsu, 2005; Mondria & Wiersma, 2004; Nation, 2002a;
Waring, 1997). Thus, the participants were seemly unwilling to spend much time on productive aspect out of frustrations. In particular, learners were seemly accustomed to receptive vocabulary learning but not productive vocabulary learning because those learners were used to test-based learning style. In the Entrance Examination of
Four-Year Technological Colleges, all questions were in multiple-choice form. Hence, this was one of the possible factors why they had better performance in receptive vocabulary learning aspect (multiple-choice part) but not productive vocabulary learning aspect. Also, observing from the participants’ task like part of speech of task in reading text, the researcher found that they didn’t review what she taught them
Four-Year Technological Colleges, all questions were in multiple-choice form. Hence, this was one of the possible factors why they had better performance in receptive vocabulary learning aspect (multiple-choice part) but not productive vocabulary learning aspect. Also, observing from the participants’ task like part of speech of task in reading text, the researcher found that they didn’t review what she taught them