6.4.1 Discussion of the Results for Research Question Three
6.4.1 discussion of the results for research question three
research question three asked: what is the relationship between learners' metacognitive listening awareness and listening comprehension under the four different listening conditions?
the only study that has examined the relationship between learners' reported use of metacognitive awareness strategies and listening comprehension ability is the study carried out by vandergrift (2003a). in order to compare the results of the present study with vandergrift's results, a brief summary of vandergrift's study is reported here (see chapter two for a more detailed account of the study).
thirty-six junior high school students studying french as a second language took a listening comprehension test consisting of three short, authentic texts, and were classified as either more skilled or less skilled listeners according to their scores obtained from the test. think-aloud data from the students were recorded for the three different listening tasks and were transcribed and analyzed using a predefined taxonomy of listening comprehension strategies (vandergrift, 1997b). each coded report of a strategy was tabulated, and a listening strategy profile was created for each student. the results showed that almost all previously identified metacognitive strategies were used by the participants in his study (planning, directed attention, problem solving, etc. ). only evaluation strategies did not appear to be used, which may have been due to the language level of the students under study, as only the more advanced learners reported using these strategies (though even their use of evaluation strategies was minimal). overall, more skilled listeners used metacognitive strategies more frequently than less skilled listeners, and the difference between the two groups was significant. also, the more skilled listeners used more problem solving strategies, i. e. they were more likely to verify and correct their comprehension (if necessary) as they were listening. on the other hand, less skilled listeners reported using translation more than the more skilled listeners, a difference that also reached significance.
in order to discuss the results for research question three, it is useful to schematize the differences between the four groups. this is done in figure 10 below. it shows that group a differs from the other three groups in that it listened to the text only once. group b, like groups c and d, listened three times, but did not receive any listening training. groups c and d differed in the kind of training they received—group c received schema-raising training and group d inferencing training.
figure 10 listening conditions for the four groups
overall, the correlations between reported use of metacognitive awareness strategies and listening comprehension, as shown from tables 64 and 65, were on the low side. only 5 out of a possible of 40 correlations (i. e., 5 aspects x 4 groups x 2 cycles) reached statistical significance in listening comprehension. the maximum variance in the listening comprehension scores accounted for by any one aspect of metacognitive awareness was 11% (.337 for group b in cycle 2, as in table 65). this can be explained by the fact that the malq, which measures learners' overall use of metacognitive strategies and can be expected to have an impact over time on their listening comprehension, may not necessarily impact on their performance of specific listening comprehension activities, especially if they are quite challenging for the learners, as was the case in this study.
in order to have a clearer picture of the general pattern of correlations found for research question three, table 66 is presented.
table 66 general pattern of correlations between the participants'ma and lc
note. ma = metacognitive awareness; lc = listening comprehension; group a = listening one time; group b = listening three times; group c = schema raising+listening three times; group d = inferencing training+listening three times; ns = not significant.
the following points have emerged from table 66:
(a) reported use of planning-evaluation, directed attention and problem solving awareness were not related to listening comprehension
it can be clearly seen from the table that planning-evaluation strategies, directed attention strategies and problem solving strategies were not related to listening comprehension scores. the explanation might lie in the fact that the learners had little opportunity to prepare themselves for the listening tasks, to evaluate the results of their listening or to undertake inferencing in the kinds of listening tasks used in this study. in other words, even if students generally had a preference for such strategies, they had no real chance to apply them in tasks of such a short duration. this has implications for the testing of listening comprehension, as in many standard tests (such as the toefl) the listening tasks are also of a very short duration.
(b) person knowledge was related to listening comprehension scores but only for groups a and b
in cycle1, the correlation for person knowledge and listening comprehension approached significance (r=.297) in group b. the correlation for group a was not significant but was positive (r=.176). person knowledge awareness was significantly correlated with listening comprehension for both of these groups in cycle 2. in the case of the other two groups (c and d) no relationship between person knowledge and listening comprehension was evident. what distinguishes groups a and b from groups c and d is whether or not they received any kind of listening training. thus it is possible that, with no training, learners are influenced by their perceptions of the difficulty of second language learning and of their own self-efficacy, but that training compensates for perceived self-efficacy.
(c) mental translation awareness is negatively related to groups b, c and d, but not a factor for group a
in cycle 1 mental translation awareness was negatively correlated with listening comprehension for all groups. in other words, learners whoreported using mental translation in the malq were likely to be less successful in carrying out the listening activity.
it should be noted, however, that the relationship between mental translation and listening comprehension varied among the groups. in cycle 1, the correlations for groups b and c were significant (r=-.326 and-.315), but for group a it was near zero (r=-.033). in cycle 2 mental translation awareness was negatively correlated with listening comprehension for all the three-time listening groups and the correlation for group c was significant (r=-.321), but for group a the correlation, though not significant, became positive (r=.257).
groups b, c and d all differed from group a in that they involved listening to the text three times. it can then be summarized that learners in all the three-time listening groups who reported using mental translation in the malq were likely to be less successful in carrying out the listening activity. in the case of group a in cycle 1 none of the learners might have been able to make use of mental translation even if they had been inclined to do so because they were under pressure to comprehend the text immediately. in cycle 2, possibly because they were more familiar with the type of listening task, there was, then, a stronger tendency to try to make use of mental translation and this helped them in some limited way.
overall though, the results show that mental translation did not assist listening comprehension; rather it was related to the listening comprehension of the less successful learners. translation arguably involves only surface mapping between languages and generally fails to activate conceptual processes (swaffar, 1988) and thus those learners who used this strategy may have been able to only interact with the text very superficially. vandergrift (2003a) found that his less skilled listeners reported using mental translation significantly more than the more skilled listeners. o'bryan and hegelheimer (2009) also found that the lowest-proficiency students demonstrated an increase in the use of mental translation strategies after a one-semester's listening course, aresult that matches vandergrift's (1997b) finding. when they translate on-line, less skilled listeners are incapable of keeping up with the incoming input, and they experience greater difficulty holding meaning in memory, a problem also noted by goh (2000). to address this problem they resort to mental translation but, as the results of vandergrift's and this study both show this does not assist comprehension.
research question three asked: what is the relationship between learners' metacognitive listening awareness and listening comprehension under the four different listening conditions?
the only study that has examined the relationship between learners' reported use of metacognitive awareness strategies and listening comprehension ability is the study carried out by vandergrift (2003a). in order to compare the results of the present study with vandergrift's results, a brief summary of vandergrift's study is reported here (see chapter two for a more detailed account of the study).
thirty-six junior high school students studying french as a second language took a listening comprehension test consisting of three short, authentic texts, and were classified as either more skilled or less skilled listeners according to their scores obtained from the test. think-aloud data from the students were recorded for the three different listening tasks and were transcribed and analyzed using a predefined taxonomy of listening comprehension strategies (vandergrift, 1997b). each coded report of a strategy was tabulated, and a listening strategy profile was created for each student. the results showed that almost all previously identified metacognitive strategies were used by the participants in his study (planning, directed attention, problem solving, etc. ). only evaluation strategies did not appear to be used, which may have been due to the language level of the students under study, as only the more advanced learners reported using these strategies (though even their use of evaluation strategies was minimal). overall, more skilled listeners used metacognitive strategies more frequently than less skilled listeners, and the difference between the two groups was significant. also, the more skilled listeners used more problem solving strategies, i. e. they were more likely to verify and correct their comprehension (if necessary) as they were listening. on the other hand, less skilled listeners reported using translation more than the more skilled listeners, a difference that also reached significance.
in order to discuss the results for research question three, it is useful to schematize the differences between the four groups. this is done in figure 10 below. it shows that group a differs from the other three groups in that it listened to the text only once. group b, like groups c and d, listened three times, but did not receive any listening training. groups c and d differed in the kind of training they received—group c received schema-raising training and group d inferencing training.
figure 10 listening conditions for the four groups
overall, the correlations between reported use of metacognitive awareness strategies and listening comprehension, as shown from tables 64 and 65, were on the low side. only 5 out of a possible of 40 correlations (i. e., 5 aspects x 4 groups x 2 cycles) reached statistical significance in listening comprehension. the maximum variance in the listening comprehension scores accounted for by any one aspect of metacognitive awareness was 11% (.337 for group b in cycle 2, as in table 65). this can be explained by the fact that the malq, which measures learners' overall use of metacognitive strategies and can be expected to have an impact over time on their listening comprehension, may not necessarily impact on their performance of specific listening comprehension activities, especially if they are quite challenging for the learners, as was the case in this study.
in order to have a clearer picture of the general pattern of correlations found for research question three, table 66 is presented.
table 66 general pattern of correlations between the participants'ma and lc
note. ma = metacognitive awareness; lc = listening comprehension; group a = listening one time; group b = listening three times; group c = schema raising+listening three times; group d = inferencing training+listening three times; ns = not significant.
the following points have emerged from table 66:
(a) reported use of planning-evaluation, directed attention and problem solving awareness were not related to listening comprehension
it can be clearly seen from the table that planning-evaluation strategies, directed attention strategies and problem solving strategies were not related to listening comprehension scores. the explanation might lie in the fact that the learners had little opportunity to prepare themselves for the listening tasks, to evaluate the results of their listening or to undertake inferencing in the kinds of listening tasks used in this study. in other words, even if students generally had a preference for such strategies, they had no real chance to apply them in tasks of such a short duration. this has implications for the testing of listening comprehension, as in many standard tests (such as the toefl) the listening tasks are also of a very short duration.
(b) person knowledge was related to listening comprehension scores but only for groups a and b
in cycle1, the correlation for person knowledge and listening comprehension approached significance (r=.297) in group b. the correlation for group a was not significant but was positive (r=.176). person knowledge awareness was significantly correlated with listening comprehension for both of these groups in cycle 2. in the case of the other two groups (c and d) no relationship between person knowledge and listening comprehension was evident. what distinguishes groups a and b from groups c and d is whether or not they received any kind of listening training. thus it is possible that, with no training, learners are influenced by their perceptions of the difficulty of second language learning and of their own self-efficacy, but that training compensates for perceived self-efficacy.
(c) mental translation awareness is negatively related to groups b, c and d, but not a factor for group a
in cycle 1 mental translation awareness was negatively correlated with listening comprehension for all groups. in other words, learners whoreported using mental translation in the malq were likely to be less successful in carrying out the listening activity.
it should be noted, however, that the relationship between mental translation and listening comprehension varied among the groups. in cycle 1, the correlations for groups b and c were significant (r=-.326 and-.315), but for group a it was near zero (r=-.033). in cycle 2 mental translation awareness was negatively correlated with listening comprehension for all the three-time listening groups and the correlation for group c was significant (r=-.321), but for group a the correlation, though not significant, became positive (r=.257).
groups b, c and d all differed from group a in that they involved listening to the text three times. it can then be summarized that learners in all the three-time listening groups who reported using mental translation in the malq were likely to be less successful in carrying out the listening activity. in the case of group a in cycle 1 none of the learners might have been able to make use of mental translation even if they had been inclined to do so because they were under pressure to comprehend the text immediately. in cycle 2, possibly because they were more familiar with the type of listening task, there was, then, a stronger tendency to try to make use of mental translation and this helped them in some limited way.
overall though, the results show that mental translation did not assist listening comprehension; rather it was related to the listening comprehension of the less successful learners. translation arguably involves only surface mapping between languages and generally fails to activate conceptual processes (swaffar, 1988) and thus those learners who used this strategy may have been able to only interact with the text very superficially. vandergrift (2003a) found that his less skilled listeners reported using mental translation significantly more than the more skilled listeners. o'bryan and hegelheimer (2009) also found that the lowest-proficiency students demonstrated an increase in the use of mental translation strategies after a one-semester's listening course, aresult that matches vandergrift's (1997b) finding. when they translate on-line, less skilled listeners are incapable of keeping up with the incoming input, and they experience greater difficulty holding meaning in memory, a problem also noted by goh (2000). to address this problem they resort to mental translation but, as the results of vandergrift's and this study both show this does not assist comprehension.