As a preface to the analyses intended to shed light on the issues addressed by the research questions presented in Chapter 1, an overview of the quantifications made for each of the variables is presented in turn here (i.e, interaction, comprehension, and lexical acquisition). Appendix F contains a tabular representation of these quantifications, and their descriptive statistics, arranged by task. Additionally, Appendix G displays the scoring for each subject on each new lexical item.
The present study is an initial attempt to investigate these variables with young learners and was designed so that the students would feel relaxed and free to interact with the teacher at will during the treatment phase by conducting the tasks in such a manner as to reflect regular classroom activities rather than experimental situations. An identical pilot study carried out the previous school-year proved successful in that end. However, the guidelines set for the teacher, which did not allow for unsolicited feedback to the learners concerning the accuracy of their task completion, proved to be too stifling for the learners in the present study and seemed to restrict interaction in Round 1. Therefore, as mentioned in Chapter 3, Round 2 allowed for unsolicited feedback from the teacher.
Another substantial difference between rounds is that while Round 1
introduced six new lexical items to the learners, Round 2 introduced nine.
Therefore, comparisons across rounds are not made.
Interaction was scored according to the procedures described in the Scoring section of Chapter 3. Please see Table 2 for a tabular representation of the interaction scores from all four tasks and their descriptive statistics.
The most apparent characteristic of the interaction scores is the wide range of differences. These differences are not limited to individual differences among subjects; differences also appear among the tasks, and individual subjects interact differently in different tasks.
Individual differences among subjects can be seen regardless of task or setting. While subject number four (S4) consistently interacted with the teacher in order to negotiate the meanings of the new lexical items, S7 and S8 scored no interaction points in any of the four tasks.
Differences among the tasks is evident in the mean interaction scores. Task 1A (Individual Zoo Keeper) stimulated the least interaction with 6 of the 10 subjects completing the task with no negotiation of meanings for the new lexical items at all. Task 2A (Individual Snack Time) produced the highest mean interaction score, although the standard deviation is also the highest as 4 subjects completed the task with no negotiation.
Examples of individual subjects interacting differently in different tasks are also apparent. S2 completed the individual task in Round 1 (Task 1A) with no negotiation but had the highest interaction score for the individual task in Round 2 (Task 2A). In Round 2, S9 completed the individual task (Task 2A) with no negotiation but had the highest interaction score in the group task (Task 2B).
Table 2. Interaction Scores for All Information-gap Tasks *
----------------------------------------------------------
Subject Task
1A Task 1B Task
2A Task 2B
n=10
----------------------------------------------------------
S1
2
7
--
--
S2 0 2 24 12
S3 0 0 1 1
S4 15 8 22 11
S5 1 3 15 3
S6 0 0 5 3
S7 0 0 0 0
S8 0 0 0 0
S9 1 5 0 21
S10 0 3 7 11
S11
--
--
0
3
_____ _____
_____ _____
m 1.9 2.8 7.4 6.5
SD
4.65
3.01
9.50 6.93
---------------------------------------------------------
* Tasks 1A & 1B introduced 6 new lexical items.
Tasks 2A & 2B introduced
9 new lexical items.
Comprehension is considerably better in the group task in Round 1 (Task 1B) with overall comprehension being 80%, and the subjects demonstrating comprehension of 68.3% of the new lexical items presented (i.e., the flash-cards to be selected).
Despite the fact that a greater number of new lexical items was introduced in Round 2, higher comprehension scores were achieved than in Round 1. The individual task (Task 2A) yielded the highest comprehension scores of all the tasks with 8 subjects scoring 100% comprehension. Subjects performed equally well selecting the flash-cards as they did placing them.
The subjects demonstrated a high level of comprehension of the directions in the group task in Round 2 (Task 2B) also, with 5 subjects scoring 100%. S7's score of 33.3% stands out as being extraordinarily low as compared to the other subjects, all of whom scored at least 75% comprehension.
-------------------------------------------------------
Table 3. Comprehension Scores for Round 1 Information-gap Tasks *
-------------------------------------------------------
Table 4. Comprehension Scores for Round 2 Information-gap Tasks *
-------------------------------------------------------
Lexical acquisition was scored according to the procedures described in the Scoring section of Chapter 3. Posttests were identical to the pretests and measured both productive and receptive vocabulary knowledge. The acquisition of the new lexical items presented in the tasks is of particular interest. Please see Table 5 for a tabular representation of these lexical acquisition scores and their descriptive statistics.
The most apparent characteristic of the lexical acquisition scores is the low gain-scores achieved in productive vocabulary knowledge. The fact that these scores are so low, and in one case nonexistent (Task 1B), precludes further analysis. Receptive scores, however, are considerably higher and warrant further investigation. In both rounds of the study, 7 days after completing the tasks, the subjects were able to identify at least one fourth of the new lexical items which had been presented.
Turning to overall lexical acquisition for all words presented in the pretests and posttests, again there is little or no productive lexical acquisition demonstrated by either the experimental group (who participated in the treatment phase) or the control group (who did not participate in the treatment phase). Therefore, no analysis of productive lexical acquisition scores is undertaken. However, the receptive lexical acquisition scores are somewhat higher, and further investigation is warranted; please see Table 6, for the descriptive statistics of the overall receptive lexical acquisition scores for Group (experimental and control) and Test (pretest and posttest) in both rounds of the study.
In comparing the degree of receptive lexical acquisition for the experimental group with that of the control group for all words included in the pretest and posttest in Round 1, it was found that there are no significant main effects for the Group effect, for the Test effect, or for the Group by Test interaction. Please see Table 7 for the results of the Repeated-measures ANOVA for overall receptive lexical acquisition in Round 1. Despite the fact that the Experimental group had a 2.5 point gain between the pretest and posttest while the Control group had a very small loss (which is probably due to chance) between the pretest and posttest the interaction effect was not statistically significant. Figure 8 illustrates the Group by Test interaction found in Round 1.
In comparing the degree of receptive lexical acquisition for the experimental group with that of the control group for all words included in the pretest and posttest in Round 2, it was found that there was no significant main effect for Group. However, the effects for Test and the Group by Test interaction were both significant. Please see Table 8 for the results of the Repeated-measures ANOVA for overall receptive lexical acquisition in Round 2. Additionally, Figure 9 illustrates the significant Group by Test interaction found in the Round 2 results. The significant interaction effect was due to the fact that the Experimental group had a 4.6 point gain between the pretest and posttest, while the Control group had no gain between the pretest and posttest.
-----------------------------------------------------------
Table 5. Acquisition of New Lexical Items Presented in All Information-gap Tasks
-----------------------------------------------------------
Table 6. Descriptive Statistics of Receptive Acquisition
-----------------------------------------------------------
Table 7. Group by Test Repeated-measures Analysis
of Variance Procedure for Round 1
-----------------------------------------------------------
Figure 8. Group by Test Interaction for Round 1
-----------------------------------------------------------
Table 8. Group by Test Repeated-measures Analysis of Variance
-----------------------------------------------------------
Figure 9. Group by Test Interaction for Round 2
-----------------------------------------------------------
Multivariate analysis of variance procedures indicated that there are
no significant differences between the individual and group settings for
any of the measures (i.e., interaction, comprehension, and lexical acquisition)
in Round 1 or Round 2; please see Table9. In the univariate tests, displayed
in Table 10, the effect for comprehension in Round 1 appears to be significant
at p<.05. In other words, it appears that in Round 1, the subjects'
comprehension is significantly higher in the group setting than it is in
the individual setting. However, since the overall multivariate tests are
not significant, this finding can only be interpreted as spurious (i.e.,
due to chance alone).
-----------------------------------------------------------
Table 9. Multivariate Tests of Significance for
Setting
-----------------------------------------------------------
Table 10. Univariate F-tests for Setting with
(1,16) D. F.
-----------------------------------------------------------
Before addressing the relationship between interaction and comprehension, a brief summary of the findings of the quantification of these variables presented above is included here. Also, as stated above, although the two rounds of the study are similar, there are substantial differences (i.e., the teacher's guidelines for directing the tasks, and the number of new words introduced); therefore, comparisons across rounds are not made.
Concerning interaction, as displayed in Table 2, the mean interaction scores are as follows: Task1A=1.9, Task1B=2.8, Task2A=7.4, Task2B=6.5. In comparing the individual task to the group task in each round, as displayed in Tables 9 and 10, there is no significant difference between the amount of negotiated interaction that occurs in these two treatment settings.
Concerning comprehension, as displayed in Table 3 and Table 4, mean comprehension scores are as follows: Task1A=7.4, Task1B=9.6, Task2A=11.6, and Task2B=10.5 (a score of 12 represents 100% comprehension in each task). The group task in Round 1 (Task 1B) yielded comprehension scores which appear to be significantly higher than the individual task in Round 1 (Task 1A) based on the univariate test (Table 10). However, since the overall multivariate tests were not significant (Table 9), this finding is interpreted as spurious. In Round 2, the difference between the comprehension scores attained in these two settings is not statistically significant (Table 9 & Table 10).
Further statistical analysis allows the investigation of the relationship between the interaction scores obtained by the subjects during the tasks and the comprehension scores the subjects achieved. Correlations between the interaction scores and the comprehension scores using the Pearson product moment correlation coefficient were calculated. Results of these calculations are presented below in Table 11.
This analysis reveals that in the individual task in Round 1 (Task 1A)
there is a statistically significant positive correlation between
interaction and comprehension; that is, higher interaction scores are associated
with higher comprehension scores and lower interaction scores are associated
with lower comprehension scores in Task 1A. In the group task in Round
1 (Task 1B) the correlation between interaction and comprehension is not
statistically significant. In the individual task in Round 2 (Task 2A)
there is a statistically significant negative correlation between
interaction and comprehension; that is, higher interaction scores are associated
with lower comprehension scores, and conversely, lower interaction scores
are associated with higher comprehension scores in Task 2A. In the group
task in Round 2 (Task 2B) the correlation between interaction and comprehension
is not statistically significant.
-----------------------------------------------------------
Table 11. Correlation Between Interaction and Comprehension
-----------------------------------------------------------
The results reveal that the extent to which the students engaged in negotiated interaction with the teacher is significantly, but inconsistently, associated with their level of comprehension in the individual tasks but not associated with their comprehension in the group tasks. That is, in the individual tasks, in Round 1 there is a statistically significant positive correlation between interaction and comprehension while in Round 2 there is a statistically significant negative correlation; but in the group tasks, there is no statistically significant correlation in either round. Considering the fact that the analysis of the relationship between interaction and comprehension in the individual tasks resulted in a positive correlation in Round 1 and a negative correlation in Round 2, the present study offers support for two seemingly contradictory contentions. The fact that higher interaction scores are associated with higher comprehension scores in Round 1 supports the results of previous research (e.g., Pica et al., 1987) which suggest that negotiated interaction facilitates comprehension. The fact that higher interaction scores are associated with lower comprehension scores in Round 2 supports the results of previous research (e.g., Chaudron, 1982) which suggest that teachers' elaborations may actually confuse learners and inhibit comprehension.
Examples of instances in which negotiated interaction can lead to comprehension are shown in Excerpts 1, 2, and 3, presented in Chapter 3 (pp. 115, 116, & 117 respectively), as students displayed comprehension of the target word in each of these interactions. In each of these examples, students initiated a non-understanding routine which resolved the comprehension problem by making the input comprehensible.
Examples of non-understanding routines which were not as readily successful
in bringing about comprehension also appear in the data. An example of
the teacher's elaboration which led to the student's confusion (indicated
by asterisks [**]) can be seen in Excerpt 4.
Excerpt 4.
-----------------------------------------------------------
Teacher: please find the avocado and
give it to the roofer/
Subject: what it's look like?/
Teacher: avocado is a big green oval + and it's rough/ it has=
Subject: =oval/=
Teacher: =bumpy skin/ it's shaped like an oval/ +
Subject: it's/ + it think it's color/ what's color look like?/
Teacher: dark green very dark green/=
Subject: =very=
Teacher: =almost like brown or black/ it's very dark green/
Subject: it's brown/ [**]
Teacher: it's GREEN/
Subject: ohh/
Teacher: but DARK green/
Subject: dark green
Teacher: and it has bumpy skin/ rough skin/
Subject: ah green skin/ green bump/
Teacher: yeah/
Subject: okay/
-----------------------------------------------------------
Although the student did temporarily misunderstand the teacher, this
non-understanding routine was successful in that the subject did demonstrate
comprehension of this particular instruction. There are instances in which
the subjects never achieved successful comprehension of an instruction
and simply gave up. In these instances, after long and arduous negotiations,
subjects did "tune out" as Chaudron warned (1982, p. 178); please see Excerpt
5 for an example.
Excerpt 5.
-----------------------------------------------------------
Teacher: if you want to say you're
finished it's okay/ but i want
to tell you that the veterinarian wants the coconut/
Subject: i'm finished/
Teacher: alright/ are you sure?/
Subject: yeah/
Teacher: do you think you can find where the coconut is?/
Subject: no/
Teacher: okay/
-----------------------------------------------------------
In another such example, the student seems to vent frustration in a
different way. After negotiating several sets of directions, there is a
knock at the classroom door resulting in an interruption during which the
student apparently moves cards, which had previously been placed, to different
locations on the board. The teacher discovers this in Excerpt 6.
Excerpt 6.
----------------------------------------------------------
Teacher: did you move a lot while
i wasn't looking?/
Subject: i start all over/
Teacher: you started all over?/ (laughs)
-----------------------------------------------------------
Excerpts 4 through 6 are taken from the individual task in Round 2 (Task 2A) which is the task that yielded a negative correlation between interaction and comprehension. It should be pointed out again, however, that the subjects received unsolicited feedback from the teacher in Round 2 while in Round 1 they did not. In Task 2A, the students who engaged in negotiation with the teacher occasionally became confused, as shown by the examples in Excerpts 4, 5, and 6, while those who simply listened to the teacher's feedback without engaging in negotiation achieved very high levels of comprehension. This resulted in the negative correlation reported above. It should also be noted that Task 2A yielded the highest comprehension scores of all the tasks, with 8 of the 10 subjects scoring 100% comprehension. Furthermore, the lowest comprehension scores in Task 2A would have been considered rather high in Task 1A (please see Table 3 and Table 4). This suggests that for the most part, with respect to comprehension, making these interactional adjustments for learners may be worth the trade off in occasional instances of confusion.
The role of learner participation in the group setting will be addressed
below in the discussion of Research Question 4. However, it should be pointed
out briefly here that with respect to comprehension, the group setting
proved to be the great equalizer in both rounds of the study. The factors
that led to both the positive correlation between interaction and comprehension
for the individual task in Round 1 and the negative correlation between
interaction and comprehension for the individual task in Round 2 seemed
to be neutralized by the group setting as there is no significant correlation
between interaction and comprehension for the group task in either round.
Before addressing the relationship between interaction and acquisition, a brief summary of the findings of the quantification of these variables presented above is included here. Also, as stated above, although the two rounds of the study are similar, there are substantial differences (i.e., the teacher's guidelines for directing the tasks, and the number of new words introduced); therefore, comparisons across rounds are not made.
Concerning interaction, as displayed in Table 2, the mean interaction scores are as follows: Task1A=1.9, Task1B=2.8, Task2A=7.4, Task2B=6.5. Tasks 1A and 1B introduced six new lexical items to the subjects while Tasks 2A and 2B introduced nine. In comparing the individual task to the group task in each round, as displayed in Tables 9 and 10, there is no significant difference between the amount of negotiated interaction that occurs in these two treatment settings.
Concerning lexical acquisition, as discussed above, productive acquisition scores are very low, or nonexistent, and therefore do not merit further investigation. However, receptive lexical acquisition scores are much higher. At least one fourth of the new lexical items presented during the tasks were identified correctly during the receptive posttest. As displayed in Table 5, the mean receptive acquisition scores for new lexical items presented in the tasks are as follows: Task1A=1.7, Task1B=1.5, Task2A=3.1, and Task2B=2.4. As displayed in Tables 9 and 10, the acquisition scores for the new lexical items presented in the individual tasks are not significantly different from the acquisition scores for the new lexical items presented in the group tasks.
Further statistical analysis allows the investigation of the relationship between the interaction scores obtained by the subjects during the tasks and the receptive acquisition scores for the new lexical items presented in the tasks. Correlations between the interaction scores and the receptive acquisition scores using Pearson product moment correlation coefficient were calculated. Results of these calculations are presented below in Table 12. This analysis reveals that there is no statistically significant correlation between interaction scores and receptive acquisition scores in any of the tasks conducted during the treatment phase of the study.
-----------------------------------------------------------
Table 12. Correlation Between Interaction and Receptive
-----------------------------------------------------------
Excerpt 7.
-----------------------------------------------------------
|
|
|
|
|
|
|
| please find the cricket= | |||||
| =cricket= | |||||
| =cricket= | |||||
| what does cricket look like? | |||||
| =i don't got no cricket/= | |||||
| =cricket is like a black kind of bug= | |||||
| =black?/= | |||||
| =and it sits on the grass/ |
Færch and Kasper list factors which determine whether or not learning will actually take place:
2. the ease or difficulty with which it can be integrated into the associative network of the learner's long-term memory;
3. the communicative value attributed to it by the learner;
4. occasions for rehearsal, in order to improve memory strength (1986, p. 270).
The fact that the experimental group did perform significantly better
on the posttest than the control group for overall receptive lexical acquisition
in Round 2 (and tended to do better in Round 1), as describe above, suggests
that participating in the tasks did have some effect on acquisition; therefore,
further analysis is warranted. Data concerning the subjects' occasions
for rehearsal beyond the treatment phase of the study or outside of class
are not available. However, a decision was made post hoc to make a frequency
count of the occasions for rehearsal in each task. Each time a new lexical
item appeared in a subject's utterance, that subject received one rehearsal
point. For the results of this frequency count from all four tasks and
their descriptive statistics, please see Table 13.
-----------------------------------------------------------
Table 13. Rehearsal Scores for All Information-gap Tasks
-----------------------------------------------------------
Further statistical analysis allows the investigation of the relationship
between the rehearsal scores obtained by the subjects during the tasks
and the receptive acquisition scores for the new lexical items presented
in the tasks. Correlations between the rehearsal scores and the receptive
acquisition scores using Pearson product moment correlation coefficient
were calculated. Results of these calculations are presented below in Table
14. This analysis reveals that there is no statistically significant correlation
between rehearsal scores and receptive acquisition scores in any of the
tasks conducted during the treatment phase of the study. It appears that
the number of opportunities for rehearsal of the new target items during
the tasks is not associated with receptive acquisition of these items.
-----------------------------------------------------------
Table 14. Correlation Between Rehearsal and Receptive Acquisition Scores
-----------------------------------------------------------
Considering the fact that the above analysis attempts to associate productive rehearsals with receptive acquisition, another decision was made post hoc to make a frequency count of students' occasions for receptive rehearsals or more appropriately, teacher repetitions. This analysis is possible only in the individual task in each round (Task 1A & Task 2A) since during the group tasks the same input was available to all subjects. Each time a new lexical item appeared in the teacher's utterances one teacher repetition point was counted. For the results of this frequency count from Task 1A and Task 2A and their descriptive statistics, please see Table 15.
Further statistical analysis allows the investigation of the relationship
between the teacher repetition scores and the subjects' receptive acquisition
scores for the new lexical items presented in the tasks. Correlations between
the teacher repetition scores and the subjects' receptive acquisition scores
using Pearson product moment correlation coefficient were calculated. Results
of these calculations are presented below in Table 16. This analysis reveals
that there is no statistically significant correlation between teacher
repetition scores and the subjects' receptive acquisition scores in either
of the individual tasks conducted during the treatment phase of the study.
It appears that the number of teacher repetitions of the new
-----------------------------------------------------------
Table 15. Teacher Repetition Scores for Tasks 1A and 2A
-----------------------------------------------------------
-----------------------------------------------------------
Table 16. Correlation Between Teacher Repetition Scores and Subjects' Receptive Acquisition Scores
-----------------------------------------------------------
Before addressing the relationship between comprehension and acquisition, a brief summary of the findings of the quantification of these variables presented above is included here. Also, as stated above, although the two rounds of the study are similar, there are substantial differences (i.e., the teacher's guidelines for directing the tasks, and the number of new words introduced); therefore, comparisons across rounds are not made.
Concerning comprehension, as displayed in Tables 3 and Table 4, mean comprehension scores are as follows: Task1A=7.4, Task1B=9.6, Task2A=11.6, and Task2B=10.5 (a score of 12 represents 100% comprehension in each task). The group task in Round 1 (Task 1B) yielded comprehension scores which appear to be significantly higher than the individual task in Round 1 (Task 1A) based on the univariate test (Table 10). However, since the overall multivariate tests were not significant (Table 9), this finding is interpreted as spurious. In Round 2, the difference between the comprehension scores attained in these two settings is not statistically significant (Table 9 & Table 10).
Concerning lexical acquisition, as discussed above, productive acquisition scores are very low, or nonexistent, and therefore do not merit further investigation. However, receptive lexical acquisition scores are much higher. At least one fourth of the new lexical items presented during the tasks were identified correctly during the receptive posttest. As displayed in Table 5, the mean receptive acquisition scores for new lexical items presented in the tasks are as follows: Task1A=1.7, Task1B=1.5, Task2A=3.1, and Task2B=2.4. As displayed in Tables 9 and 10, the acquisition scores for the new lexical items presented in the individual tasks are not significantly different from the acquisition scores for the new lexical items presented in the group tasks.
Further statistical analysis allows the investigation of the relationship between the comprehension scores achieved by the subjects during the tasks and the acquisition scores for the new lexical items presented in the tasks. Correlations between the comprehension scores and the receptive acquisition scores using Pearson product moment correlation coefficient were calculated. Results of these calculations are presented below in Table 17. This analysis reveals that there is no statistically significant correlation between comprehension scores and receptive acquisition scores in any of the tasks conducted during the treatment phase of the study.
As stated above, there is no statistically significant correlation between comprehension scores and receptive acquisition scores; however, the negative correlation between comprehension and receptive lexical acquisition noted in Task 2B is disturbing if not significant. Therefore, this discussion addresses a possible explanation for the occurrence of this negative correlation. Scores achieved by a single subject (S7) seem to be a major contributing factor as seen in Table 5 which shows that while S7 had the highest receptive acquisition score (55.5%), her comprehension score of 33.3% stands out as being extraordinarily low. Examining the transcription of Task 2B reveals that S7's participation in the task was extremely limited and that she completed the entire task (which lasted over 17 minutes) without uttering one syllable. Furthermore, after successfully carrying out the first two instructions (which led to her receiving 4 comprehension points), she completely withdrew from active participation and never touched another flash-card despite continued encouragement from the teacher as well as other subjects. In spite of her lack of active participation, she was exposed to the same input as the other subjects. Moreover, some of the other subjects attempted to help her by providing additional information about the new lexical items. Please see Excerpt 8 for an example in which S10 offers new information about the target word fig, and S9 offers new information about the target word custodian in addition to the teacher's encouragement to participate.
It is true that S7 did not successfully carry out the task instructions, however this does not necessarily lead to the conclusion that no comprehension took place on her part. Interpreting the negative correlation between comprehension scores and receptive acquisition scores to conclude that comprehension hindered acquisition would be shortsighted. However, the results of the present study offer no evidence at all that comprehension facilitated acquisition. Needless to say, the analysis of the data collected in the present study offer no evidence for the existence of any sort of a three-way relationship for the variables in which negotiated interaction facilitates comprehension which, in turn, facilitates acquisition.
Excerpt 8.
--------------------------------------------------------------
|
|
|
|
|
|
|
| anybody have anything to say to help S7?/ she's looking for the fig= | |||||
| =YEAH/= | |||||
| =but she can't find it/ go ahead S10/ | |||||
| um the little one is um sorta green/ | |||||
| it's a little one and it's sorta green he said/= | |||||
| =and= | |||||
| =can you find it now S7?/ anybody else have anything to say?/ | |||||
| me/ | |||||
| stay down S9/ go ahead S9/ | |||||
| put at + the + cleaning place/ | |||||
| she says to put it at the cleaning place/ yeah the custodian/ but first she has to find the fig/ can you find it S7?/ do you want me to help you any more or are you finished?/ + S7/ + well okay i guess S7's finished/ |
During group interaction, do the relationships among interaction,
comprehension, and acquisition differ for those subjects who orally participate
in the interaction as compared to those subjects who are exposed to the
interaction but choose not to orally participate?
As is shown in the findings presented above, no significant correlation exists between interaction and comprehension or interaction and acquisition in the group tasks in either of the two rounds of the present study (Task 1B and Task 2B); please see Tables 11 and 12. Considering these results, it is extremely doubtful that those subjects who orally participate in the interaction during the group tasks perform significantly differently from those subjects who are exposed to the interaction but choose not to orally participate.
However, to further investigate this question, the Fisher Exact probability test is utilized as the samples are small in size. The subjects are classified into groups based on the analysis of the transcriptions which served to quantify the negotiated interaction. Although the criteria for classification are different for the present study and earlier studies (e.g., Seliger, 1977; Day, 1984) the terms high input generator (HIG) and low input generator (LIG) are also be used here rather than unnecessarily confusing the issue with new terminology. Similar to the procedures of earlier studies, the subjects in the present study are ranked according to their interaction scores. The top 25% are classified as HIGs, and the bottom 25% are classified as LIGs (the percentages vary due to ties). With this classification complete, the HIGs and LIGs are further subdivided into two classifications, high achievers and low achievers in both comprehension and receptive lexical acquisition. Thus constructing four separate 2 X 2 contingency tables with the distribution of frequencies derived from these nominal data. The first calculation determines whether the two groups (HIGs and LIGs) differ in the proportion with which their members fall into the two classifications (high achiever and low achiever for comprehension) in the group task in Round 1 (Task 1B). The second calculation determines whether the two groups (HIGs and LIGs) differ in the proportion with which their members fall into the two classifications (high achiever and low achiever for comprehension) in the group task in Round 2 (Task 2B). The third calculation determines whether the two groups (HIGs and LIGs) differ in the proportion with which their members fall into the two classifications (high achiever and low achiever for acquisition) in the group task in Round 1 (Task 1B). The fourth calculation determines whether the two groups (HIGs and LIGs) differ in the proportion with which their members fall into the two classifications (high achiever and low achiever for acquisition) in the group task in Round 2 (Task 2B). The contingency tables are presented below in Figure 10, Figure 11, Figure 12, and Figure 13 respectively. Results of these analyses show that there is no significant difference between the HIGs and LIGs for either comprehension or acquisition in the group tasks in either Round 1 or Round 2 of the present study. These findings preclude the existence of any three-way relationship among the variables (i.e., interaction, comprehension, and lexical acquisition).
-----------------------------------------------------------
Figure 10. Task 1B Comprehension and Participation
-----------------------------------------------------------
Figure 11. Task 2B Comprehension and Participation
-----------------------------------------------------------
Figure 12. Task 1B Acquisition and Participation
-----------------------------------------------------------
Figure 13. Task 2B Acquisition and Participation
-----------------------------------------------------------
The results of the present study support the findings of previous research (e.g., Pica, 1991a) which suggest that negotiated interaction has the potential to facilitate comprehension for the students who are exposed to the interaction as well as those who directly participate. Evidence to support this claim is found in both rounds of the present study.
In Round 1 of the present study, the teacher does not supply unsolicited feedback to the learners. There is no significant difference in the total amount of negotiated interaction that occurred between the individual and group settings (as displayed in Tables 9 & 10). However, in the individual task the new information about the meanings of the unknown lexical items elicited from the teacher was available only to those subjects who initiated interaction while those subjects who did not initiate interaction were not supplied with new information, hence the positive correlation between interaction and comprehension in the individual setting (as displayed in Table 11). However, it appears that the deficit in the level of comprehension achieved by those subjects who did not initiate interaction in the individual task was overcome by the group setting, as evidenced by the lack of a significant correlation between interaction and comprehension in the group task.
In Round 2 of the present study, the teacher does supply unsolicited
feedback to the learners. Although there is no significant difference in
the amount of negotiated interaction that occurred between the individual
and group settings (please see Tables 9 & 10), it is apparent that
the teacher supplied considerably more assistance in overcoming comprehension
problems in the individual task in Round 2 than he did in the individual
task in Round 1. Sufficient evidence for this claim can be seen by comparing
the amount of teacher repetitions in both tasks (displayed in Table 15).
However, this unsolicited feedback coupled with negotiation initiated by
certain learners led to extended elaboration by the teacher causing several
instances of confusion which resulted in a negative correlation between
interaction and comprehension in the individual setting (as displayed in
Table 11). This deficit in the level of comprehension achieved by those
subjects who experienced confusion in the individual task was somehow overcome
by the group setting, as evidenced by the lack of a significant correlation
between interaction and comprehension in the group task (also displayed
in Table 11). One possible explanation for this is that the teacher, out
of consideration for the other members of the group, avoided extended,
and possibly confusing, negotiation with any one particular subject. As
displayed in Table 4, both tasks in Round 2 yielded very high comprehension
scores. These high levels of comprehension had no effect on acquisition,
however, as evidenced by the absence of significant correlations between
comprehension and acquisition for any of the tasks (as displayed in Table
17). The results of this correlational analysis notwithstanding, it must
be pointed out again that the experimental group achieved greater gains
in lexical acquisition in both rounds of the study, as measured by the
pretests and posttests (displayed in Table 6), and that the Group by Test
interaction found in the Round 2 results was found to be statistically
significant (illustrated in Figure 9).