國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 4.5
Bob‘s Distribution of the Researchers‘
Evaluation of Writing Coherence drafts were both very close to the full score: 5.
Moreover, Bob got 0.23 more points in the second draft of the description type. In his one-picture narration type, Bob made the most progress: he got 1.28 more grades in the last draft. As to the coherence of Peter‘s comic-strip narration type of writing, he got 0.25 more grades in the second draft and 0.08 more grades in the third draft. Peter also scored 0.08 more grades in the second draft in the description type of writing.
Besides, he also made the most progress in the one-picture narration type of writing:
he got 0.48 more grades in the last draft.
In the researchers‘ rating of coherence, Peter‘s coherence averagely scored a little higher than Bob‘s, but Bob‘s progress was more significant than Peter‘s. As a result, the participants were proved to make progress in coherence, which is also corroborated by the GEPT raters‘ scoring. This suggests the instruction of CD and TP should have some positive effects on the coherence of the participants‘ writing. Their scores also indicate their great understanding and application of CD and TP.
Effect 3: the Frequency of Certain Categories in CD and TP
Then it is essential to examine how often the participants adopted CD and TP through counting the number of cohesive words and TP types in each draft. The results were gathered from the participants‘ written products, the researchers‘
Figure 4.6
Peter‘s Distribution of the Researchers‘
Evaluation of Writing Coherence
‧
coherence analyses and the raters‘ comments on coherence. Through analyzing the frequency of CD and TP categories in the participants‘ writing, how the instruction of CD and TP affected the participants‘ adoption of certain kinds of CD and TP can be clearly identified.
The Analysis of CD
The categories of CD appearing in the participants‘ writing were checked to know if there was any connection between the CD categories used by the participants and their progress of coherence. Table 4.4 presents the frequency of the cohesive words applied in the participants‘ written texts. ―B‖ represents Bob and ―P‖ is Peter.
Table 4.4
The Frequency of Cohesive Words in the Participants‘ Drafts
%
Due to little use of substitution, ellipsis and collocation in the participants‘
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
essays, these three categories were excluded in this section. The following is the process of how the results of CD frequency generated. First of all, the number of certain cohesive devices applied in one of the participants‘ drafts was counted. The first sentence would have one cohesive device, and the next sentence would have one or two cohesive devices to connect the previous sentence and the following sentence.
Then the percent of the cohesive devices in one draft was calculated based on the total number of the sentences in one draft and so was the sum of Bob‘s and Peter‘s each CD category.
According to Table 4.4, the CD category which was applied most frequently in all of the two participants‘ drafts was lexical reiteration. This category appeared in each of the students‘ drafts. Pronominal reference was the second most that both of them adopted. The third category which appeared often in the students‘ writing was temporal conjunction. Then Figure 4.7 and 4.8 illustrate the distribution of the use of reiteration in Bob‘s and Peter‘s writing based on the three different writing types.
The above two figures show that, except for Peter‘s comic-strip narration type, the participants both applied less and less reiteration categories in their other drafts. In other words, they attempted to avoid repeating the same words or phrases in their later drafts. The following figures are the distribution of the participants‘ application of reference.
Figure 4.7
Bob‘s Distribution of Lexical Reiteration
Figure 4.8
Peter‘s Distribution of Lexical Reiteration
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
In the above two figures, both Bob and Peter used more reference categories in their third drafts of comic-strip narration type. However, in the description and one-picture narration types of writing, Bob applied less reference in his later drafts, while Peter used more. Then Figure 4.11 and 4.12 show the distribution of the participants‘ application of conjunction.
Contrary to the application of reiteration, both the participants could write down more conjunctive words to make their sentences more connected in their later drafts.
Most of all, Peter seemed to cultivate the habit of adding conjunctive words to his sentences because he never used this kind of cohesive DEVICE in the beginning of the study.
To sum up, the similarity between the cohesive words used by Bob and Peter Figure 4.9
Bob‘s Distribution of Reference
Figure 4.10
Peter‘s Distribution of Reference
Figure 4.11
Bob‘s Distribution of Conjunction
Figure 4.12
Peter‘s Distribution of Conjunction
‧
was the significantly increasing frequency of applying conjunctive adverbs; on the contrary, the frequency of reiteration decreased by the end of the study. Besides, the two participants showed different development of using some cohesive words. While Peter employed more reference but less reiteration in his later drafts, Bob used less reference but more reiteration. Without the same or similar words written by these participants, they learned to use proper cohesive words to make their writing more coherent. In either case, they both could make progress in their writing coherence. In general, the instruction of CD, to some extent, had its effect on students‘ performance in writing cohesion.
The Analysis of TP
Besides the exploration of CD, this section focuses on the coherence in the participants‘ written works. The frequency of the participants‘ application of TP types is recorded in Table 4.5. Three of the TP types—TP with a hyper-theme, split rheme and a thematic jump—were excluded in the analysis as the participants never used them.
Table 4.5
The Thematic Patterning in the Participants‘ Writing
% simple
The following is the process of how the results of TP frequency generated. In
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
the beginning, the number of TP types used in one of the participants‘ drafts was counted. Every two sentences could be identified one TP type. Next, the percent of the TP type in one draft was calculated based on the total number of the sentences in one draft and so was the sum of Bob‘s and Peter‘s each TP type. The distribution of each TP type in the participants‘ written works is illustrated in the following figures.
The illustration is also on the basis of the three writing types that the participants wrote. First of all, Figure 4.13 and 4.14 show the distribution of Bob‘s and Peter‘s applying the simple linear TP and TP with a constant theme at the time of the study.
Figure 4.13
The Distribution of Bob‘s Use of Simple Linear TP and TP with a Constant Theme
From the above figure, it can be found that Bob was able to adopt theses two TP types in turn. Moreover, Bob preferred to use simple linear TP type in his later drafts (D3, D5 and D7) of each writing type, while the TP with a constant theme accounted much higher proportion than the simple linear TP type except for D2 and D4.
Figure 4.14
The Distribution of Peter‘s Use of Simple Linear TP and TP with a Constant Theme
In Figure 4.14, we can find that these two TP types also took turns to appear in
‧
Peter‘s written drafts. Besides, the proportion of Peter‘s applying simple linear TP and TP with a constant theme was far higher than Bob‘s. In addition, except for the
description type (D4 and D5), the difference of the frequency of these two TP types in Peter‘s writing was not as significant as Bob‘s.
Then it is also necessary to examine how the participants applied the other TP types they learned in the study. Figure 4.15 and 4.16 present the distribution of the other TP types used by the participants.
Figure 4.15
The Distribution of Bob‘s Use of the Other TP Types
0
Percent of the use of the other TP types %
TP with a subsequent rheme TP with interrelated rhemes no TP
In Figure 4.15, it seemed that, compared to Figure 4.13, except for the
description type, Bob tended to averagely apply simple linear TP, TP with a constant theme, TP with a subsequent rheme and TP with interrelated rhemes in most of his drafts. However, there were not any TP types among some sentences in each of his drafts, but the proportion of no TP type decreased in each writing type. Therefore, Bob‘s topical progression became more and more evident in his later drafts of these three different writing types.
Figure 4.16
The Distribution of Peter‘s Use of the Other TP Types
0
Percent of the use of the other TP types %
TP with a subsequent rheme TP with interrelated rhemes no TP
‧
As shown in Figure 4.16, Peter applied few of these two TP types, TP with a subsequent rheme and TP with interrelated rhemes, in his later drafts. His application of TP types focused on the simple linear TP and TP with a constant theme.
Additionally, only two of his drafts (D4 and D6) had no TP types. Therefore, Peter seemed to be able to make his sentences more connected than Bob.
In conclusion, though the frequency of Bob‘s and Peter‘s TP types was not exactly identical, the results presented that both the participants could apply well the TP types that they learned throughout the study. Moreover, with the combination of CD and TP, the participants could identify the TP types easily and correctly no matter what CD categories were employed. As a result, the participants were able to
compose coherent texts with CD and TP carefully applied.
The Raters‟ Comments on the Writing Coherence
To explore if the above-mentioned distribution of CD and TP types was associated with the progression of the participants‘ writing coherence, the raters‘
comments on coherence were listed to be compared with the frequency of specific CD and TP types. The raters‘ comments are presented in Table 4.6.
Table 4.6
The Raters‘ Comments on Coherence
Points Beginning
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Peter F F F W Fair coherence
Fight (D7)
Bob F G F G With proper transition
words, coherence
Peter F F F G With details, fertile
imagination G: good F: fair W: weak
The comments on Bob‘s transition words and cohesiveness are almost ‗Fair‘, which reassured the steady development of Bob‘s coherence in writing. The good application of some related cohesive words helped him remain steady in his performance of writing coherence. On the other hand, Peter got less ‗Weak‘ in his textual coherence and some of his drafts were even ranked ‗Good‘ in terms of the use of transition words. Peter‘s significant progress in writing and coherence
corresponded with his application of more conjunction words and clear reference in his later writing. In line with the frequency of CD and TP in Bob‘s and Peter‘s writing, the comments they received all include proper transition words, good coherence and many details, especially in the later drafts in each writing types.
Therefore, besides the frequency of CD and TP types used in the participants‘
written drafts, the raters‘ comments on the coherence and organization also reflected the improvement in the participants‘ writing. With the instruction of CD and TP, the participants both could choose and apply the proper CD and TP categories to make their writing more cohesive and coherent.
The Development of the Participants‘ Writing
The process which the participants went through after learning CD and TP is the subsequent concern in this section. Certain types of writing development were
categorized to answer the second research question from the process and perception data. The patterns of the data are classified as follows: the development of writing coherence and of writing skills.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
The Development of Writing Coherence
The first concern was that the participants were able to get themselves quite familiar with the application of CD and TP by examining the process and perception data collected. Based on the data, the participants‘ few problems with coherence identification demonstrated that it was indeed feasible for low-proficiency students to acquire the ability to be good at CD and TP. The development would be described based on the participants‘ familiarity with CD and TP.
The Participants‟ Familiarity with CD and TP
From the transcription of the writing conferences, journals and interviews, it was found that the two participants were able to add something reasonable to make the coherence more complete when they wrote or revised their texts on their own by examining CD and TP. Take Peter‘s fourth conference transcript as an example.
… yeah, I added or crossed out some words when I could not find any CD and TP.
P-WC-D7-011
When discovering no connection in the texts, the participants not only were able to detect the lack of coherence, but they could know how to make up for the absence of the key words or phrases. That is to say, they applied CD and TP to their written texts to confirm if the coherence existed. If not, they would follow the trace brought by CD and TP to specify what they wanted to express in their writing. Furthermore, they did not decide the coherence categories at random. This indicated their ability to discern whether there was sufficient coherence in their texts.
On the other hand, Peter was also able to recognize whether a sentence was coherent after he revised his texts on his own. He even regarded the TP with a continuous theme used in some of his texts as a simple one.
But I used the very easy one, where the subjects in sentences were connected together.
P-WC-D5-135
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
No matter which type he adopted and investigated in his writing, it was apparent that he had a good grasp of identifying TP. He later recorded his finding in his last journal.
The context became better in the third sentence—Joe and Mark have the same habit—when it was revised as „Joe is also a bad student in his class.‟
P-J-D7-001-003
After revising his final text, Peter found the revised sentences could connect the other sentences more closely than the previous one. In other words, with the study approaching to the end, Peter got more proficient in the coherence detection.
Therefore, both Bob and Peter mentioned that they found it uncomplicated to identify CD and TP when they did this by themselves at home.
No, it‟s not difficult to look for CD and draw TP.
B-WC-D3-004 Mm…little difficulty.
P-WC-D2-002
Now that the participants found it straightforward to employ CD and TP, the process of identifying coherence therefore grew fairly rapid. Take Bob‘s transcript of the second writing conference as an example.
R: Would you spend lots of time circling the cohesive words?
B: No… I just find them…
B-WC-D2-011-012
Due to little difficulty in identifying CD and TP, it undoubtedly took Bob less time to build up his textual coherence. That is to say, when Bob reread his written texts again prior to the writing conferences, it was effortless for him to trace the connection that had been made. Peter was on the other hand capable of detecting the wrong categories of CD or TP he had written down when discussing with the researcher in the writing conferences.
R: … The connection between sentences 1 and 2 is okay. But you‟d better check the connection between the next two sentences. Is there any better connection between them?
B: Um… it should be Joe…
P-WC-D7-021-023
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
In spite of the slight errors made by the participants themselves, they found the correct categories very soon with or even without the guide of the researcher in the writing conferences. This also corroborated the preceding assertion that the learning of CD and TP had a certain influence on the progress in the writing coherence.
Apart from the capability of correcting the loose inter-sentential connection, the participants were also trained to perform the task of analyzing CD and TP
independently. For example, Bob found he could complete the process of identifying CD and TP whether the researcher helped him or not.
Yeah, there is no difference when I search for CD and TP in conferences or at home.
B-I-359
Therefore, when the notion of CD and TP was internalized, the participants then were able to operate the coherence identification individually whenever they wrote in English.
Therefore, Bob referred to the timing he would make use of to involve the CD and TP in writing compositions—when taking tests.
Yeah, I will think of CD and TP to use when I take a test.
B-I-285
Despite the time limit in a test, Bob still could think of CD and TP to use, which corresponded with what he previously mentioned: applying CD and TP was as normal as the other basic strategies he used in writing.
All in all, the above qualitative data exhibited the possibility to equip EFL low-proficiency learners with the revision strategies—CD and TP. The analysis demonstrated that low achievers could learn CD and TP well and apply them to their writing properly. In view of these findings, the participants were really familiar with the application of CD and TP.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
The Development of Writing Skills
Another goal of the present study is to help the low achievers develop the ability to check the coherence in their written texts. From the process and perception data, the two participants were found to prefer CD and to focus more on content level in writing than before. Besides, revising also began to play a significant role in their writing process.
The Preference for CD
In comparison with CD, there was no remarkable development in the
identification of TP because the participants had done very well in the analysis of TP since the very beginning of the study. Consequently, this section would only focus on the progression of CD. Then the qualitative data were probed to trace the progression of the participants‘ learning CD. First, the two participants‘ different development in preferring CD was presented. Then they were found to be clearly aware of the benefits brought by CD.
Different development in identifying CD. Because of their different
characteristics, the following section would respectively describe the progression each participant went through.
At the start of the study, the participants decided the CD or TP categories of some words they had written in haste or with little care. Take Bob‘s first writing
At the start of the study, the participants decided the CD or TP categories of some words they had written in haste or with little care. Take Bob‘s first writing