The researcher mainly seeks to find out whether the participants in the EG can
retell longer stories after the treatment of the GO map instruction. First, the result of
the original word count is presented. Next, the result of the pruned word count is
shown. Third, the result concerning the differences between the original and pruned
word count is reported. Finally, the effects of the GO map instruction on the length
of the story retelling are addressed.
4.1.1 Results of the original word count
The original word count (i.e., the original number of words contained in the
retold story) is briefly summarized in Table 2 to display the average length of the
stories retold by the EG and the CG in the pretest and posttest. The results here
41
show that the mean scores on the original word count for the story retold by the
participants in the EG and CG and are 147.5 and 153 in the pretest and 145.3 and
141.1 in the posttest respectively. In terms of the mean value, the CG outperformed
the EG (CG: 153> EG: 147.5) in the pretest while the results were reversed in the
posttest (EG: 145.3>CG: 141.1). Regarding the changes between the pretest and
posttest, there is an average decrease of 2.2 words in the original word count for the
EG and of 11.9 words for the CG in the posttest. Hence, participants in both groups
retold shorter stories in the posttest, and the decrease is much sharper for the CG.
The between-group t-test shows that there was no significant difference in the pretest
story retelling performances between the EG and the CG (t=-0.242), suggesting that
the average original word count in each group in the pretest was statistically the same.
Since the value of the within-group t-test indicates that the decrease in the original
word count is statistically insignificant for both groups (EG: t=0.236; CG: t=1.568),
the retold stories in the posttest, though of a shorter length, were not significantly
shorter.
Table 2 A T-test of Mean Scores on the Original Word Count
Pretest Posttest Pre-posttest t-test
M SD M SD t-value
The EG (N=27) 147.5 96.2 145.3 71.3 0.236 The CG (N=27) 153.0 68.1 141.1 59.1 1.568
Note. *p< .05, **p< .01, M=Mean, SD=Standard Deviation
42
4.1.2 Results of the pruned word count
The mean scores on the pruned word count (i.e., the number of words in the
retold story with false starts and repetitions excluded) are summarized in Table 3 to
show the average pruned length of the stories retold by the EG and the CG in the
pretest and posttest. As Table 3 shows, the mean scores on the pruned word count in
the story retold by the participants in the EG and CG are 137.8 and 146.3 in the
pretest and 139.8 and 133.4 in the posttest respectively. In terms of the mean value,
the CG outperformed the EG in the pretest (CG: 146.3> EG: 137.8) while the results
were reversed in the posttest (EG: 139.8>CG: 133.4). Regarding the changes
between the pretest and posttest, there is an average increase of 2.0 words for the EG
whereas there is an average decrease of 12.9 words for the CG. Hence, the
participants in the EG retold modestly more content in the posttest with false starts
and repetitions excluded, while those in the CG retold moderately less. The
between-group t-test shows that there was no significant difference in the pretest story
retelling performances between the EG and the CG (t=-0.227), suggesting that the
average pruned word count in each group in the pretest was statistically the same.
Since the value of the within-group t-test indicates that neither groups retold
significantly longer or shorter stories in the posttest (EG: t=-0.217; CG: t=1.647), the
seemingly great decrease between pretest and posttest for the CG remains statistically
43
insignificant (t=1.647).
Table 3 A T-test of Mean Scores on the Pruned Word Count
Pretest Posttest Pre-posttest t-test
M SD M SD t-value
The EG (N=27) 137.8 87.0 139.8 66.49 -0.217 The CG (N=27) 146.3 71.3 133.4 57.9 1.647
Note. *p< .05, **p< .01, M=Mean, SD=Standard Deviation
4.1.3 Results of comparison between original and pruned word count
In the present study, the differences between the original and pruned word count
means the false starts and repetitions the participants produced. Therefore, the
bigger the difference is, the more false starts and repetitions there are.
The mean scores on the word count of false starts and repetitions in both the
pretest and posttest are summarized in Table 4. The results here show that the
average differences in the EG and CG are 9.70 and 6.70 in the pretest and 5.48 and
7.63 in the posttest respectively. In terms of the mean value, the CG outperformed
the EG in the pretest (CG: 6.7 < EG: 9.7) by producing fewer false starts and
repetitions while the results were reversed in the posttest (EG: 5.48 < CG: 7.63).
Regarding the changes between the pretest and posttest, there is an average
decrease of 4.22 words in false starts and repetitions (5.48-9.7=-4.22) for the EG;
however, there is an average increase of 0.93 words (7.63-6.7=0.93) for the CG.
That is, the participants in the EG made more progress than their counterparts from
the pretest to the posttest by producing fewer false starts and repetitions. On the
44
other hand, the story retelling of the CG even contained more false starts and
repetitions in the posttest. In addition, the between-group t-test shows that there was
no significant difference in the pretest story retelling performances between the EG
and the CG (t=0.735), suggesting that the average number of false starts and
repetitions in each group in the pretest was statistically the same. The within-group
t-test shows that the number of false starts and repetitions is significantly lowered
between the pretest and posttest for the EG (t=2.198*, t< .05). Thus, only the
participants in the EG were able to retell the posttest story with significantly fewer
false starts and repetitions.
Table 4 A T-test of Mean Ratio of False Starts and Repetitions
Pretest Posttest Pre-posttest t-test
M % SD M % SD t-value
The EG (N=27) 9.70 6.6% 13.04 5.48 3.8% 11.00 2.198*
The CG(N=27) 6.70 4.4% 16.73 7.63 5.4% 5.88 -0.258
Note. *p< .05, **p< .01, M=Mean, SD=Standard Deviation, %= Ratio
To better discriminate the differences between the original and pruned word
count, the mean ratio of false starts and repetitions in the original word count is
provided as shown in Table 4. It is calculated by dividing the number of false starts
and repetitions by the total number of words and then multiplying the ratio by 100 %.
Take the ratio of the EG in the pretest for example, the mean ratio is gained as follows:
9.7/147.5 × 100% = 6.6%. Accordingly, the mean ratios of the EG and CG are 6.6
% and. 4.4 % in the pretest respectively and 3.8 % and 5.4 % respectively in the
45
posttest. As the mean ratios indicate, there is an average decrease of 2.8 % in the
ratio (from 6.6 % to 3.8 %) for the EG whereas there is an average increase of 1.0 %
(from 4.4 % to 5.4 %) for the CG. In other words, the participants in the EG
generated proportionally a smaller number of false starts and repetitions in the
posttest.
4.1.4 Discussion of effects of GO map instruction on length
The results show that like the traditional Q & A approach, the GO map
instruction was not an effective method in assisting the students to produce a longer
account when retelling the story in English. The GO map instruction, however, is
proved to be effective in reducing false starts and repetitions in students’ retelling. A
plausible explanation may be inferred: The participants in the EG had transferred the
knowledge of the GO map to the posttest story so that when preparing for the retelling,
they could visualize a clearer story structure in mind. This helped guide their
retelling meta-cognitively and in turn reduce the ratio of false starts and repetitions.
Since they had a better organization of what to retell next, they had more control over
avoiding false starts and repetitions from happening.