4.2 Posttest
4.2.1 Learners’ Learning Achievement Score
4.2.1 Learners’ Learning Achievement Scores
4.2.1.1 Paired samples t test
Paired samples t-test was conducted to examine the treatment effect on participants’ near-synonymous vocabulary knowledge for the control group and experimental group individually. The means and standard deviations for within group differences of the posttests are presented in Table 4.2 for each group. An alpha level at p = .05 was considered significant.
Table 4.2 Paired sample T-test in the two groups
The results showed that participants in the experimental group made significant improvement by 17.07 points in the posttest (t = 8.505, p < .05). As for the control group, participants also made significant progress in posttest (t = 4.525, p < .05), the mean score of the posttest increased by 6.65 points compared with that of the pretest.
In other words, participants in both groups showed great progress in their
near-synonymous vocabulary knowledge after near-synonym instruction with or without corpus-based data-driven learning. The pretest mean scores of the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
50
experimental group were lower than that of the control group; however, the posttest mean scores of the experimental group were higher than that of the control group.
To sum up, to answer the first research question, there were significant differences in learners’ posttest scores between the experimental group and the control group. The corpus-based data-driven learning group outperformed the traditional in-class teaching group in their learning achievement scores. Such a finding confirms study of (Sun &Wang, 2003; Chan & Liou, 2005; Koosha &
Jafarpour, 2006; Yeh, Liou & Li, 2007; Alex, 2009; Ewa, 2011; Serkan, 2011) that data-driven learning enhances participants’ second language vocabulary acquisition and data-driven learning approach performed better than traditional teaching
methodology. The possible reason might be that the data-driven learning approach helped participants pay attention to the word forms and retrieve the word meanings from their memory. Moreover, another possible reason may be that data-driven learning is not only a process-oriented but a product-oriented approach. Participants made effort to investigate the target words with his or her classmates and infer the meanings and usages within contexts. Participants had a sense of participation through the process and had a strong impression on the target words after the whole process.
DDL not only promotes creativity and self-discovery learning among learners, but also presents to the learners by multiple exposures within contexts (Batstone, 1995).
The results also echoed the statement that “learners have a big improvement after corpus-based teaching instructions (Tribble & Jones, 1990; John, 1994; Tribble, 1996;
Kettemann, 1995).”
‧
4.2.1.2 Independent sample t-test
Since the treatment yielded significant effects in the posttest of data-driven learning in the experimental and the control group, their specific improvements in high and low achievers were further examined.
Independent sample t-test was applied to examine the treatment effects on posttest between high and low achievers in the control and the experimental groups.
An alpha level set at p = .05 was considered significant. The means and standard deviations of the between-group high and low achievers’ differences are shown in Table 4.3 and Table 4.4.
Table 4.3 Independent sample T-test of posttest of high achievers in the two groups Group Mean SD df t Sig.(2-tailed)
Table 4.4 Independent sample T-test of posttest of low achievers in the two groups Group Mean SD df t Sig.(2-tailed)
In the aspect of high achievers, the results showed that the differences between the control group and the experimental group were significant (t = 2.735, p < .025).
The mean score of the experimental group (M = 69.47) outperformed that of the control group (M = 58.40), which indicates that when learning the near-synonyms, EFL high proficiency learners perform better with data-driven learning approach as scaffolding than high achievers who were taught by the traditional method.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
52
This finding is not in contradiction with that of the empirical study in previous research. For high and low achievers with DDL approach, this finding confirms the study of Lin (2005), though there are important differences regarding other aspects of the study. This result lends some credence to the hypothesis that different language proficiency and different language competence might have impact on the effect of data-driven learning (e.g. Tseng, 2002). High achievers in DDL group seem to be indicative of the fact that inductive learning can lead to greater improvement to second language vocabulary learning. However, the data suggests that DDL does not only improve learners’ receptive skills, but also improve their productive skills, which was not found in Lin’s study.
In the aspect of low achievers, the results showed that the mean score of the experimental group (M = 59.53) was slightly higher than that of the control group (M = 54.92). However, the differences between the control group and the
experimental group were non-significant (t = 1.748, p >.05), which rejected the hypothesis that when learning the near-synonyms, EFL low proficiency learners perform better with data-driven learning approach than low achievers who were taught by the traditional method.
Since using DDL to teach low achievers has been a controversial issue, the direct comparisons must be treated with caution. The present study enhances the previous studies by providing a much more detailed examination of DDL. In Hadley’s (2002) study, he was assured by colleagues in Japan that any attempt to use DDL with beginners was doomed to failure even though these same colleagues had never tried to use DDL to learn vocabulary. Other than Hadley, some researchers (Cobb, 1999;
Allan, 2006; Koosha and Jafarpour, 2006) also assumed that DDL is not helpful for low achievers. The result of present study conforms to the assumption. The possible
‧
reason is that low achievers might simply not have sufficient analytical and linguistic skills to cope with the complexity of authentic data of a second or foreign language.
The deeper reasons were further investigated in the qualitative results.
4.2.1.3 Summary of result
Results of the pretest and posttest comparisons revealed that both groups improved their vocabulary knowledge, after receiving near-synonymous instructions, with or without using data-driven learning approach. Although the control group made some progress in the posttest, the gains of the experimental group were
significantly greater than those of the control group. According to Figure 4.1, we can see that before the data-driven instructions, the participants’ entering behaviors were the same which meant they had the same vocabulary knowledge before receiving the teaching treatment. Later, after eight weeks of data-driven learning training, the participants in the experiment group acquired more vocabulary knowledge than those of the control group.
Figure 4.1 Pre- and Post-test Scores
The results of between-group comparison of both high achievers showed that high proficiency EFL learners improved their performance in vocabulary knowledge,
0
‧
after being treated with or without data-driven learning approach. Data-driven learning approach and traditional teaching method both had influence on learners’
vocabulary knowledge. High achievers in the experimental groups made more progress than those of the control group. Overall speaking, the participants in the experiemtal group outperformed those of the control group.
Figure 4.2 showed the progress of high achievers and low achievers in their pre-and post-tests. The high achievers’ mean score in the experiment group had a lot of progess (incresed from 52.53 to 69.47) than that of the control group (incresed from 52.8 to 58.4 ). Similarly, the low achievers’ mean score in the experiment group had a lot of improvement (increased from 45.06 to 60.12) than that of in the control group (increased from 44.62 to 54.92). The data showed that no matter who the participants were, the high or low achievers made a lot of imporvement in the experimental group rather than the control group. Additionally, when comparing the between-group progress gains, the high achievers’ mean score still had more
improvement (increased from 58.04 to 69.47) than low achivers’ mean score (increased from 54.92 to 59.53). This result suggests that DDL might have much more influence on high achievers than low achievers.
Figure 4.2 Pre- and Post-test Results of Progress Gains
16.94
high achievers (mean scors) low achievers (mean score) experiment group control group
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
55
Based on the results of within-group comparison or between-groups, it showed that data-driven learning is a more ideal scaffolding teaching methodology to cultivate EFL learners’ knowledge of vocabulary knowledge of near-synonyms. The present study provides a valid proof in agreement with the fact that “data-driven learning enhanced participants’ second language vocabulary acquisition.”