Proficiency Level
In section 4.2, it was shown that enhanced background knowledge pre-teaching and enhanced vocabulary pre-instruction were, as a whole, equally effective to junior high school students’ listening comprehension. But it was also found that student’s proficiency level played an important role in determining their listening performance.
In other words, it was very likely that proficiency level could moderate the effects of listening supports. Therefore, in this section, the interaction between types of treatment and students’ proficiency level is investigated in order to answer research question 4 and 5, specifically termed as the following:
RQ4: Is there an interaction between types of pre-listening supports and participants’ proficiency level?
RQ5: When participants’ proficiency level is taken into consideration, which type of pre-listening support is more helpful?
In order to answer these two questions, another method of grouping was used.
Depending on students’ achievement scores, the top 30% students were categorized into high-proficient learners (HL) and the low 30% counterparts were labeled as low-proficient learners (LL). Afterwards, two-way ANCOVA between-groups analysis was conducted to see if there was a significant difference in post-listening comprehension test scores between HL and LL in their response to three kinds of treatments. Table 4-9 showed that no significant interaction was found between types of listening support and proficiency level (F=.07, df=2, p>.05). As to the two independent variables, the effect of participants’ proficiency level on listening comprehension did not reach significance as well (F=.04, df=1, p>.05). However, significant difference was found in the main effect of pre-listening support types (F=3.58, df=2, p<.05, eta squared=.12). These findings suggested that the effects of both enhanced background knowledge and vocabulary pre-teaching did not depend on the level of students’ proficiency; in other words, the single factor of enhanced pre-listening supports was strong enough to contribute to effective listening comprehension. Also, the main effect of treatments had reached a large effect size, accounting for 12% of the variance on students’ listening comprehension scores.
Briefly speaking, junior high students in the present study benefited from the two types of pre-listening support activities, regardless of how proficient they were.
Table 4-9: ANCOVA of Between-Group Effects
Source SS’ df MS’ F Eta Square
Treatment 14.47 2 7.23 3.58* .12
Proficiency Level .09 1 .09 .04 .00
Treatment × Proficiency Level
.27 2 .14 .07 .00
*P<.05
Although the interaction between types of pre-listening supports and students’
proficiency level did not reach a significant level, some implications could still be drawn by examining the adjusted means on listening comprehension tests that each proficiency group in three treatment classes scored after removing the influence of covariate. Table 4-10 showed the original mean scores for HL and LL groups in three treatment classes and the corresponding adjusted scores. It seemed that background knowledge pre-instruction was the most effective pre-listening support across proficiency groups and vocabulary pre-teaching came the next. (background high=8.41 > voc high=7.51 > control high=7.24; background low=8.22 > voc low=7.52 > control low=6.88).
Table 4-10: Means and Adjusted Means for Each Sub-Group
Group N Mean Adjusted Means Std. Error
High 12 8.92 8.41 .56
Background
Low 10 7.90 8.22 .51
High 6 8.00 7.51 .69
Vocabulary
Low 13 7.00 7.52 .55
High 12 7.75 7.24 .56
Control
Low 7 6.14 6.88 .77
Graph 4-7 further presented the visual image, with upper line representing high-proficient students and lower line representing low-proficient students in three treatment classes. Some important implications were derived from these data. First, two bands of proficiency level students showed a similar pattern on their listening performance; that is, they scored the highest in enhanced background knowledge class and the lowest in control class and the mean score difference between these two types of treatment reached 1.17 and 1.34 points (10% v.s. 11%) respectively. Second, high- and low- proficient students almost scored the same means under the treatment of vocabulary pre-teaching; in fact, low proficient students even scored slightly better than high proficient counterparts (mean=7.52 v.s. mean=7.51). These findings suggested that regardless of their proficiency level, the participants in the present study benefited most from the help of enhanced background knowledge pre-teaching.
They also suggested that enhanced vocabulary pre-instruction might work better for low-proficiency students. High proficient learners did not respond to this type of treatment as well as they did to the enhanced background knowledge support, judging from the fact that their mean scores dropped more dramatically from background treatment to vocabulary treatment. Third, the mean score differences between high- and low- proficiency students were relatively small in enhanced background class (mean dif.=.19) and enhanced vocabulary class (mean dif.=.01) when compared to that in the control class (mean dif.=.36). This suggested that the application of pre-listening supports in class could help alleviate the commonly-occurring bimodal curve phenomenon in Taiwan. Such advantage can mean that teachers could address their teaching to a wider range of students and would be less worried about those students in extreme ends to find the lecture either too difficult or too easy.
Graph 4-7: Adjusted Means of Listening Scores
In conclusion, the above findings answered research question 4 and 5. That is, there was no significant interaction between types of pre-listening supports and participants’ proficiency level. Nevertheless, when the statistically adjusted means for each sub-group were further analyzed, results showed that participants in the present study benefited the most from the enhanced background knowledge support no matter how proficient they were. As far as the enhanced vocabulary pre-teaching was concerned, it seemed that this type of support worked better for lower proficient learners than higher proficient ones. In addition, in terms of the fact that the mean score difference narrowed between high- and low- proficient students when pre-listening supports were given, it may suggest that teachers could alleviate the tricky bimodal curve situation by incorporating into the class these two types of the pre-listening supports.