The Addition of an Affect test and Self-assessment into ESL Writing Assessment: Process and effect. A Case Study in a Non-English Major Postgraduate ESL Writing Class

| May 5, 2007
The Addition of an Affect test and Self-assessment into ESL Writing Assessment: Process and effect. A Case Study in a Non-English Major Postgraduate ESL Writing Class

Keywords: ESL writing; affect test; self-assessment

Huili Wang Yucui Wang

Bio Data
Wang Yucui has just completed her MA from Dalian University of Technology. This paper is part of her Masters thesis. Wang Huili is an Associate Professor and a Masters Advisor in the Foreign Languages School at Dalian University of Technology in P.R. China. She has been teaching since 1989. She earned a Bachelor of Arts from Shanghai International Studies University and then a Master of Education from Dalian University of Technology. She is now doing her Ph.D. Her research interests include English Education, Education Technology, and psycholinguistics.

This paper introduces both affect test and self-assessment into the traditional assessment scheme of non-English major post-graduates ESL writing classes. Learning results after the actual application of a whole semester are analyzed and compared to see whether such addition can improve or encourage ESL learner s writing performance. Participants are also interviewed to reflect on their attitude towards such addition. Results show that both affect test and self-assessment are welcomed by the majority of learners and play a relatively positive role in the whole ESL writing process. As reported by participants, affect test can promote and facilitate the process of English learning by stimulating learners to have a deeper thinking about their learning states and make favorable adjustments accordingly. Self-assessment, similarly, offers learners an opportunity to examine their own writing and find progress by themselves. The comparison of writing scores suggests that self-assessment is more of help in self-revising than in timed-essay writing while affect test leads learners to make more progress in the final writing. However, results in the second round of affect test do not show the similar significance in progress in final writings.
[private] Abstract

This paper introduces both affect test and self-assessment into the traditional assessment scheme of non-English major post-graduates ESL writing classes. Learning results after the actual application of a whole semester are analyzed and compared to see whether such addition can improve or encourage ESL learner s writing performance. Participants are also interviewed to reflect on their attitude towards such addition. Results show that both affect test and self-assessment are welcomed by the majority of learners and play a relatively positive role in the whole ESL writing process. As reported by participants, affect test can promote and facilitate the process of English learning by stimulating learners to have a deeper thinking about their learning states and make favorable adjustments accordingly. Self-assessment, similarly, offers learners an opportunity to examine their own writing and find progress by themselves. The comparison of writing scores suggests that self-assessment is more of help in self-revising than in timed-essay writing while affect test leads learners to make more progress in the final writing. However, results in the second round of affect test do not show the similar significance in progress in final writings.

Keywords: ESL writing; affect test; self-assessment

1. Introduction
The issue of arriving at an effective means of measuring students learning has always been a major concern to not only teachers, but also students themselves as well as education administrators even, given the complex nature of learning process, the educational and political context, and the widespread and growing use of portfolios and web-based instructions in higher education. When it comes to ESL writing, the matter of how to reach an effective assessment turns out to be even more complex.

The term assessment is derived from ad sedere to sit down beside. The implication of its etymology is that it is primarily concerned with providing guidance and feedback to the learner. Arguably, this is still its most important function. (Brown et al, 1997, p. 11) As iterated by many researchers (Blumenfeld and Marx, 1997; MacCombs and Marzano, 1990), learning requires both will and skill. For this reason, assessment, as an indispensable part of education, should help students to be aware of their own thinking, to be strategic and to direct their motivation toward valuable goals, one of which is for students to learn to be their own teachers (Schunk and Zimmerman, 1998). Thus, assessing learners performance only through paper-and-pen task can hardly fulfill the multi-functions of assessment to both evaluate and more importantly encourage and stimulate learners learning. Consequently, in this paper, affect test and self-assessment are introduced into the ESL writing class aiming to diversify the content of assessment and better facilitate learners English study.

2. A brief review of literature
Previous research on affect test, mostly about learning strategies, concentrate mainly on two directions. One is on the relationship between EFL learners’ learning strategy use and other emotional factors, such as learners belief (e.g. Nae-Dong Yang, 1999), motivation (e.g. Oxford and Nyikos, 1989), and meta-cognition (Abraham and Vann, 1987; Horwitz, 1987, 1988; Wenden, 1986, 1987). The other is to assess the use of learning strategy with different inventories, like Strategy Inventory for Language Learning (SILL)(Oxford; Burry-Stock,1995), Beliefs about Language Learning Inventory (BALLI) (Horwitz, 1996) or in different contexts (Sheorey, 1999). Less work was done to assess whether the addition of such affect test can help to improve ESL learner s performance both quantitatively and qualitatively.

Research on self-assessment roughly follow three lines. Most studies have focused on researches into issues related to the reliability and validity of self-assessment (for example, Bachman and Palmer, 1989; Stefani, 1994; Orsmond et al., 2000). Others have proposed ways and techniques to increase their reliability and validity (Li, 2001; Smith et al., 2002; Taras, 2002). Yet another line of research has focused on the use and advantages of self-assessment (for example, Oskarsson, 1989). However, most of the researches were done in oral and reading discourses and fewer were done in writing context. In addition, the quantitative analysis on how much, if any, the self-assessment can improve ESL learner s writing performance was not so much done as the three main domains mentioned above. And learners attitude towards such self-assessment was not fully explored.

Therefore, this study tends to invite both affect test and self-assessment into the traditional assessment scheme and compare the learning results of learners after the actual application of a whole semester to see whether such addition can improve or encourage ESL learner s writing performance.

3. Research questions

Can affect test help increase students ESL writing performance? If yes, to what extent?
What is students attitude towards affect test in ESL writing?
To what extent, if there is any, can self-assessment improve students ESL writing?
What is students attitude towards self-assessment in ESL writing?
4. Methodology
Unlike most previous researches on affect test and self-assessment, the target participants of the present study are post-graduate students who are thought to be more mature in psychology, more aware of their affect condition and more self-regulated in their ESL study.

Five classes, two classes in the first semester and three in the second semester, comprising 122 non-English major postgraduates, taking part in ESL Writing Class for 18 weeks, participated in this study. The two classes in the first semester were randomly assigned to one of the two conditions: an affect test condition (ATG), and a control group (CG). The three classes in the second semester were randomly assigned to one of the three conditions: an affect test condition (ATG), a self-assessment condition (SSG) and a control group (CG). All participants had to attend five face-to-face instructions and 8 on-line instructions on Self-access English Learning System (SELS). They had to hand in three compositions and finish all the 6 compulsive tasks on SELS (the other 2 were selective tasks whose marks were not included in the final grades) to get their final grades for this class and receive their credit in English study.

4.2 Samples and procedure
The flow charts illustrate the procedure of research on the effect of affect test and that of self-assessment (see Chart 1and Chart 2).

The compositions handed in by participants in face-to-face instructions, three from each individual, were collected as the samples of the study. That the samples consisted of only compositions from face-to-face instructions is because on-line instructions required little writing task of a complete passage. All the participants, whether in ATG, SSG or CG, attended the same face-to-face and on-line instructions given by the same teacher in each respective semester. They were required to finish the same assignments on line before the same due time and to hand in the same amount of timed-essays with the same topics in face-to-face classes. In such a way, all the participants enjoyed the same learning environment, no matter in class or out of class. The only difference between ATG and CG or SSG and CG is the presence or absence of affect test or self-assessment. Since the three essays were written at the beginning, in the middle and at the end of a semester, they were expected to show the development and progress of participants in ESL writing and thus to tell whether the addition of affect test and self-assessment can better encourage or improve students ESL writing than traditional teaching only. The participants of ATG and SSG were also interviewed by the end of the semester to express their own feelings and opinions on such a change.

The effect of the addition of affect test and that of self-assessment were tested separately in different participants in order not to load too heavy burden on learners at one time and thus win more cooperation from them. In addition, separate testing can also show more clearly the individual effect and function of each part by reducing the inter-influence they may have on each other.

4.3 Instrument
Self-report questionnaires (all in Likert 5-point scale, see Appendix 1) & interview, Software Package of Social Sciences (SPSS) and Excel 2003 were employed in the research.

Although most of the affective assessment researches involve some type of learner s self-report, the questionnaire as a research method has been questioned because of its possible problems: social desirability biases in responses and over-subjectivity (Oxford, 1995) to name just a few. Nevertheless, researchers have also discovered, through conducting repeated studies with clear instructions in situations where no grades or sanctions are involved, that many or most learners are capable of accomplishing the questionnaire in a relatively objective manner (Oxford, 1995).

In order to reduce the biases on responses, participants in this study were not required to reveal their names on the questionnaires. Instead, they were just required to write down their Student s ID number for the convenience of matching and handing out of the results afterwards. They were also told that not a single result of any individual would be revealed to teachers and, there would be no connection between the results of their questionnaires and their final grades. Besides, in the instruction of the questionnaire, an emphasis was made by a written sentence in black and italic form, There is no rights or wrongs between choices, so please choose what I believe rather than what I should do or what others think is right . It is expected that the concealment of their names and the emphatic instruction can relieve the participants worries or psychological burdens and thus invite more true thought from them.

All the questionnaires were administered in the face-to-face classes and collected immediately the participants had finished them in case that they might lose or leave them nowhere and couldn t hand them in, and thus guarantee the reliability of the results.

It should be noted that taking into account the mother-tongue preference and the convenience of comprehension and acceptance of the participants, all the questionnaires were written in Chinese. The interview to both teacher and students were also carried out in Chinese. It is for the consistence of language of this paper that part of the questionnaires and the results and feedback represented here are translated into English.

5. Detailed description of the research procedure on the effect of affect test
5.1 Participants
Twenty-two non-English major postgraduates, 8 female and 14 male, in Writing Class 2, and another 21 non-English major postgraduates, 9 female and 12 male in Writing Class 3 participated in the ATG to test the effect of addition of affect test.

5.2 Instrument
Four questionnaires (The Questionnaire on Learner s Learning Concepts, TheLearning and Study Strategies Inventory (LASSI), The Questionnaire on Learner s Learning Style, Learner s Review on Affect Test) and an interview. All the questionnaires have been tested and utilized more than once by Chinese researchers (Liu Runqing, 2003; Wen Qiufang, 2003)

5.3 Questionnaire Administering & Collecting
The writing class for non-English major postgraduates lasted for 16 weeks, five of which were face-to-face instructions when students were required to get together in a classroom for a lecture as in a traditional teaching class while the others were on-line tasks. The 4 questionnaires were administered in face-to-face instruction classes.

On average, each questionnaire took a participant about 10 minutes to finish, and was not a heavy burden. The questionnaires were administered during the break or after the instruction finishes so that it would not influence the classroom teaching. Every time the questionnaires were collected, the results were analyzed and then returned to both teachers and students in the form of diagnosis and advice (see Appendix 2) in the next class.

The interview was carried out after the questionnaires on participants attitude to the addition of affect test were collected and analyzed. Each individual was interviewed respectively in a quiet room with no other people, students or teachers, present in order to prevent the interviewee from aimlessly echoing with others or purposely favoring the teachers. The interview with each participant lasted for about 15 minutes on average. The whole process was recorded and then transcribed and summarized if necessary.

6. Detailed description of the research procedure on the effect of self-assessment
6.1 Participants
All the 24 non-English major postgraduates in writing class 1 agreed to participate in the entire self-assessment activities, but many of them were absent for at least one class during the daily instructions and therefore did not finish all the self-assessment procedures. Thus their data were excluded and only 15 participants complete data were taken into account as the samples of SSG.

6.2 Instrument
Rubric, questionnaire and interview

6.3 The co-establishment of the rubric
Stage 1: The students and lecturers were invited to offer their own opinions on most 5 weighty evaluation criteria, which were then collected and classified as a basic set of assessment criteria (see Table 1).
Stage 2: Each individual criterion in the basic set of assessment criteria is then rearranged in a descending order according to the repeated times it appeared in participants contributions. And the first 6 were selected in order to enable students to focus on the most important aspects without being detracted by too many criteria.
Stage 3: A list of criteria was administered to both teachers and students who were invited to give weight to each individual criterion by rearranging them in a descending order. They were also invited to add more criteria which were not included in the list (see appendix 3).
Stage 4: The feedback from both students and teachers were then analyzed according to the following formation.
Students votes + teachers votes *3
Weight of each individual criterion = ———————————————–
Sum of votes
The results were then compared with diverse evaluation criteria in ESL writing tests such as TOEFL, CET4 &6, IELTS, and each individual criterion was expanded and expatiated accordingly to make it more understandable and more applicable. The rubric was then formed (see Table 2).

6.4 The insertion of self-assessment into daily teaching and learning
Students were asked to write three compositions altogether, two in class and one after class. The first one was free writing. The second one was the revision of the first draft. The third one was a timed essay writing task with given topics. In the first face-to-face class, students were asked to write a composition with whatever topic they were interested in or good at to test their real writing ability. In the second face-to-face class, the first drafts were given back to students with teacher-assessed marks on. The co-established rubric was also handed out to students who were then asked to revise and edit their first draft according to the handed out rubric in the following 3-4 weeks. They were also asked to grade their first drafts and second drafts according to the rubric and write down their self-reflection on these two drafts. The self-reflection is mainly about their greatest improvement in the second draft with supportive details. The third one is a timed essay undergone in the second period of the last face-to-face instruction. Students were given 45 minutes to write an argumentative essay with a given topic. They were also asked to do the self-assessment again by grading their writing based on the rubric after the writing. Both the composition and the graded rubric were asked to be handed in by the end of the class. In this way, participants had done self-assessment for three times in the whole semester, twice after class and once in class.

6.5 The collection of feedback from participants at the end of the semester
Like what is done in the testing of affect test, a questionnaire was administered to the participants in SSG followed by an interview with each individual participant to collect their attitude towards the addition of self-assessment, including the co-establishment of the rubric, the self-grading of the composition and the self-reflection on the revised draft and the initial draft. Questions in questionnaire and interview were mainly to test whether self-assessment, in participants opinion, can promote their interest and confidence in English learning and increase their self-awareness on their learning state by doing a self-reflection.

7 Results and discussion
7.1 Affect test
7.1.1 Comparison of writing scores
The statistics and the graph (see Figure 1) illustrate the trend of means in ATG and CG. Mean in ATG raises from 12.91 – the full score is 20 – in the first writing to 15.5 in the revised writing and remains there in the final writing. Mean in CG raises from 13.64 in the first writing to 16.05 in the revised ones and falls down a little to 14.82 in the final ones. The scores of the second drafts rank the highest of the three, both in ATG and in CG. That s not surprising because the second draft is the revised version of the first draft and the students were given sufficient time to do the revising. They could also refer to any references or consult anyone for help. The improvement in the final writing is fairly reasonable given the efforts from both instructors and students devoted to writing class.

The comparison of means of ATG and CG indicates that participants in ATG achieved more progress in scores than those in CG. The mean of ATG was almost one point lower than that of CG in the first writing (12.909vs13.636), but the difference reduced to half a point in the second writing (15.500vs16.045). The mean of ATG even surpassed that of CG in the final writing (15.500 vs. 14.818), which indicates a bigger progress in ATG.

The results of Independent T-test (see Figure 2) also prove the significance of such an improvement in ATG. The difference of means between ATG and CG were not significant in the first and the second writings or the progress of the second writing, but showed significance at the 0.05 level in the progress of means in the final writing (with the 2-tailed significance value of 0.029).

On one hand, such a significance of difference may be a natural result coming from the difference of writing ability and competence existing before the addition of the affect test. In other words, the scores of the first writing served partially as a predictor of the corresponding final scores. On the other hand, however, such a significant difference also probably resulted from the presence or absence of the affect test. The Test of Between-Subjects Effects (see table 3) proved such a deduction. Although the correlation value between progress of the final writing and the first writing shows greater significance, (significant at the 0.01 level), the correlation value between progress of the final writing and the addition of affect test shows significance at the 0.05 level, which, to some extent, indicates that the affect test also contributes to the significant progress in the final writing in ATG.
Surprisingly enough, the results of the second semester didn t follow the same trend of the first semester. No significance in means of progress, whether in the revised writings or the final ones, was shown (see Table 4).

7.1.2 Students review on the addition of affect test
The feedback from the participants review on affect test (see Table 5) shows that the overwhelming majority of the participants hold a positive attitude towards the addition of affect test into ESL writing class. Seven aspects out of the nine positive items received a mean value of more than 3.5 with the highest mean as 3.96. According to participants feedback, the biggest help of affect test lies in the fact that it helps learners to think more about their English study. Most participants agreed that it was necessary to bring such affect test into the whole learning process. The results also show that affect test is of little help in promoting learners motivation and interest in English learning. Although some participants may consider affect test as no contribution to change their present state of English study, no one regards it as a waste of time or detraction from their normal study.
Participants feedback from the interview can be mainly divided into two kinds, supportive or neutral. Most participants reported that affect test was helpful to their English study while only two participants reported it was of little help. Their feedback consolidated the positive role the affect test has played in the semester-long instructions.

I like such affect testing, because it helps me find my shortcomings in English study, increases my interest and also functions as a supervisor. My previous experience in English study proved to be somewhat blind or aimless. But now I seem to have found the direction . Those items with positive feedback strengthened my confidence while those with negative feedback stimulated me to devote more efforts (excerpt from participants conversation transcription: supportive).

It helps me realize how to learn English more effectively and efficiently. It also improves my learning methods. Take vocabulary-memorizing for example. What I did before was to learn each new word in a vocabulary book by heart alphabetically. Now I try enlarging my vocabulary through reading and writing. In the past, all I considered as English learning was to attend instructions given by teachers. That s all. This semester, I ve changed my idea and was able to insist on learning English by myself day by day. Reading, listening, in whatever way, I made sure myself to get in touch with English every day (excerpt from participants conversation transcription: supportive).

I have already formed my own habits and thoughts on English learning and would not like to be influenced by outer factors. Therefore, I didn t care too much about the results of the affect test. However, I also appreciate it because it is very interesting. Anyway, it may provide some information to teachers about what students are thinking about and then help their teaching.
(excerpt from participants conversation transcription: neutral)

The students review can easily lead us to a conclusion that affect test is welcomed by students, especially those who are not so confident in English learning because they often feel at a loss in English learning and need more guidance and instruction from others.

Many students also suggested that such affect test should be involved in the assessment system even earlier, upon their entrance to the university for example. It is true that postgraduate students are more aware about their own thought and affect feelings and can yield more reliable results in affect testing. However, with the maturation of their way of thinking and behaving, their learning habits (including learning strategies and learning style and many other things) have also been fixed and become really hard to make any change by just a few questionnaires. In addition, undergraduates have more time of English learning and naturally more time to make adjustment of their learning habits according to the results of affect testing. Another reason to begin affect test earlier is that the changes in the way of teaching as well as the way of learning in colleges require undergraduate students to realize their responsibility as the master of learning as soon as possible. In that sense, affect test can help them to change their concepts of English learning and stimulate their motivation as well.

All in all, all the interviewees consider it necessary to add affect test into students assessment system for it can give them more information on their learning results. Even the few who thought it no use changing their way of learning also support the addition of such an affect test, for they also found it interesting and they could at least get some information beyond their own thinking. The results of affect test are returned to participants in the form of a diagnosis, which adds more humane care into the cold assessment system. It is of greater significance in a web-based learning environment where students interact more with the computer than with the real persons, which may cause feeling of isolation and loneliness. The non-quantitative way of expressing the results can make them feel more like interacting with a psychologist.

7.2 Self-assessment
7.2.1 Comparison of writing scores
(See Table 6 and Figure 3) As stated in 7.1.1, it s quite natural that means of three writings go up in the second writings and then fall down in the final ones because participants have sufficient time to refer to any consultant and make improvement in their second writing while the final one is required to finish within a limited time and with a given topic.

As Figure 4 suggests, there is no obvious difference between SSG and CG in the first, the second and the final writings. But the difference in the progress in the second writing is rather significant (with the significant value of 0.014), which indicates that participants in SSG achieved more progress in revision.

7.2.2 Correlations of SS and TS
The correlation value between students self-assessed scores (SS) and teacher-assessed scores (TS) of the first writings is 0.13, significant at the 0.05 level. It reduces to 0.09 in the second writings which is significant at the 0.01 level, showing that the students self-assessed scores were more approaching to the teacher s. Such an increase in significance is quite reasonable given the more time and more sufficient references the learners enjoyed in the second writing process. The results indicate that students are able to do self-assessment well, especially with sufficient time or being well trained (see Table 7 & Figure 5, Table 8 and Figure 6).

The correlation between SS and TS of the finial writing shows no significance. It may result from the following reasons according to the followed-up interviews. First, participants became more confident in writing after a whole semester s learning. Consequently, they tended to give higher marks to their own writing when doing the self-assessment. Second, the third writing was a little bit more difficult than the first one. Most of the participants reported that they were short of time, to different extent, to make a careful thinking on evaluating their own writing within such a limited time. Another feature may arise if second glance is given to the comparison of TS and SS in the final writing. Although SS are much higher than the TS, the two lines almost follow the similar trend. In other words, the disparities between SS and TS between different participants were more or less the same, which means most of the participants over marked their writing to the same extent. It indicates, to some extent, that students were more confident about their own writing and thus graded more to their composition (see Table 9 and Figure 7).

7.2.3 Students review on the addition of self-assessment
Participants feedbacks from the interviews show their great support to the addition of self-assessment into ESL writing class. The following are some excerpts from the transcription of participants interview conversations on some typical questions.
What is the biggest help of self-assessment for you?
It s my first time to do self-assessment. I found my shortcomings in writing by myself which is more impressive than receiving teacher s assessment.
I have never revised my own writing before. And now I saw more clearly my disadvantages through self-assessment. It points out a direction of my future effort.
It forces me to have a self-reflection. To be frank, I thought in the past that the composition was written for teachers. But now, I suddenly found that I began to think more about my writing and discovered many problems which I neglected before. For example, sentence patterns and tenses and the things like that will all draw my attention.
Is self-assessment of any help to build up your confidence in ESL writing?
After the self-reflection and the self-assessment of the second writing, I found more confidence in myself. At least I have a clear idea of how to develop paragraphs in a logical order and how to develop each paragraph because now I have a rubric as guidance.
How did you find self-assessment contribute to your ESL writing?
I found it of most help in revising my composition. In the past, I only corrected where teachers had marked with a sign and never thought about how to get rid of such mistakes in the next composition. Neither had I ever thought of where else besides teachers suggestions need improvement [sic]. However, now, being requested to do self-assessment, I would have a second thought on the composition before handing it in, such as can I change this sentence into another way of expression and how? For example, I was to hand in my revised version much earlier. But just before handing it in, I found there were too many sentences with the verb make , so I had another few days thinking about how to alternate it into more expressive verbs. Such a practice indeed made me think more and learn more.

All the participants in SSG found self-assessment of some, if not much, help to their ESL writing. They reported that self-assessment would contribute more if sufficient time were given to allow them to have an overall evaluation of their compositions as was done in the second writings. That can also explain why the correlation between SS and TS in the revised writings is much more significant than that of the final timed essays. Participants self-report also suggested that the co-establishment of the rubric made them feel more involved in the learning process and more like a master of the writing. The rubric helped them better understand the requirements of a high-quality composition and promoted them to follow such requirements in their own writing. However, they also reported that the rubric was of more help during the revision period than during the timed-writing process, because sufficient time allows them to think deeply about their writing with the guidance of each individual requirement listed in the rubric. Most participants reported that they gained more confidence in ESL writing from doing self-assessment since they knew better where their shortcomings were and could then make appropriate improvement and more progress in the next writings.

7.3 Other possible explanations to the results from affect test
The results of the questionnaires can not only provide feedback to students with the aim to raise their self-awareness about their meta-cognitive conditions, but also provide teachers with analysis of students affect, both of each individual and of learners as a whole. Besides, they can also offer some possible explanations to participants attitude toward the new form of assessment.

As the first pie graphs (see Figure 8) show, 48% of the participants in the first semester and 67% in the second believe in the importance of their own role in English study while others are still not clear about their believing. No one holds the completely opposite idea that it is the ulterior environment or others rather than themselves that determine the result of English learning. The distribution of the participants marks shows that even among those who are not clear about their own believing, more than half are with a score that is closer to 24 (the division standard between self-believing and not clear yet ) than to 32 (the division standard between not clear yet and not self-believing ), which gives us a clearer and more persuasive illustration that most participants are willing to rely on themselves to achieve progress in English learning.

The majority of the participants in ATG ( see Figure 9, 84% in the first semester and 96 % in the second semester) hold a position that it is important and necessary to control and manage the learning process in English study, which are precisely part of the functions that affect test may serve.

Participants positive answers to the questions in questionnaires (see Table 10) suggest that they would like to know more about their learning states and to make appropriate adjustments when necessary. Affect test and self-assessment can fulfill their need to understand their own learning and show their progress and shortcomings in time, and thus undoubtedly receive welcome from learners.

Results of affect test on participants motivation (see Figure 10) show that most participants have a deep motivation or both the deep and the surface motivation.

The results above go in line with the mental characteristics of postgraduate students. As a grown-up, a postgraduate student is often more mature in psychology and thought. He is eager to know more about himself and he respects the power to control himself in all aspects including English learning. He is no long an innocent child who blames everyone or everything else but not himself for the failure in language learning. Such maturity is very helpful to him in English learning, for it will facilitate the process of acquiring more knowledge and seek more learning opportunity on his own initiative. He is better at applying his meta-cognition to direct and regulate his own study activities. Naturally he is more ready to accept the analysis on his affect which he brought into the learning.

8. Conclusion
The result of this research suggests that both affect test and self-assessment are welcomed by the majority of learners and play a relatively positive role in the whole ESL writing process.

The addition of affect test can encourage students to be more self-regulated in English study, although not much significance was shown in the scores between ATG and CG. Affect test, according to participants feedback, can promote and facilitate the process of English learning by stimulating learners to have a deeper thinking about their learning state as well as the affect factors they have brought into learning process, whether positive or negative, and then make favorable adjustments accordingly. Such a self-reflection makes them better aware of their learning goal and motivation, their learning strategy and learning style, their effective learning methods and direction for future efforts and, can thus regulate their English study in a more reasonable way and get more satisfactory learning results.

Self-assessment is more of help in self-editing and revising than in composing process itself. The co-establishment of the rubric forces the learners to have a second, if not more, thought on the criteria of writing assessment and consequently have a better understanding about them. The deep consideration of each individual criterion will leave impression on learners mind and urges them to follow the guide in their own writing. However, just as participants have reported, even when they understand the rubric fairly well and try to implement the requirement in writing, it s just easier for them to make considerable improvement in structure and organization as well as to correct simple grammatical mistakes with the direction of rubric. When it goes to sentence patterns and authentic ways of expression, which need long-time accumulation, the help of rubric is rather limited. The revision and comparison of two drafts are of most help to learners, for it provides learners with an opportunity to examine their own writing and experience the joy of making progress.

Both affect test and self-assessment, especially the former, should be introduced into the whole process of teaching and learning even earlier. That s because although postgraduate students are more mature in both psychology and mental control and thus can do a better job in self-reporting, they have already formed a fixed learning habit which is really hard to break or make any significant change by just a few questionnaires or interviews.

The study presented here is limited in a number of ways, including the relatively small size of sample and the occasional absence of participants in SSG and ATG. Besides, more researches with larger sample and more fixed participants need to be done to better uncover the relationship between the addition or absence of affect test and self-assessment and the improvement in ESL writing performance. Attempts should also be made to explore an effective way to turn the positive affect evoked by affect test and self-assessment into real writing ability in practice.

Abraham, R.G., Vann, R.J., (1987). Strategies of two language learners: a case study. In: Wenden, A.L.,Rubin, J. (Eds.), Learner Strategies in Language Learning (pp. 85-102). Englewood, NJ: Prentice-Hall.

Bachman, L., Palmer, A., (1989). The construct validation of self ratings of communicative language ability. Language Testing, 6, 14-29.

Blumenfeld, P.C. & Marx, R.W. (1997). Motivation and cognition. In H.J. Walberg & G.D. Haertel (Eds.), Psychology and educational practice (pp. 79-106). Berkeley, CA: McCutchan.

Brown, G., Bull, J. & Pendlebury, M. (1997). Assessing students learning in higher education. London: Routledge.

Elaine K. Horwitz (1996). Cultural and situational influences on foreign language learners’ beliefs about language learning: a review of BALLI studies. System, 27, 557-576

Horwitz, E.K., (1987).Surveying student beliefs about language learning. In A. Wenden and J. Rubin (Eds.), Learner Strategies in Language Learning (pp. 119-129). Englewood, NJ: Prentice-Hall.

Horwitz, E.K., (1988). The beliefs about language learning of beginning university foreign language students. Modern Language Journal, 72, 283-294.

Li, L., (2001). Some refinement on peer assessment of group projects. Assessment and Evaluation in Higher Education , 26, 5-18.

Liu, R.Q., & Dan, M.Ch. (2003). Studies on Condition and Development Strategy of ESL Teaching Reform in Chinese Universities. Beijing

McCombs, B.J. & Marzano, R.J. (1990). Putting the self in self-regulated learning: The self as agent in integrating will and skill. Educational Psychologist, 25, 51-69.

Nae-Dong Yang, (1999). The relationship between EFL learners’ beliefs and learning strategy use System, 27, 515-535

Orsmond, P., Merry, S., Reiling, K., (2000). The use of student derived marking criteria in peer and self-assessment. Assessment and Evaluation in Higher Education , 25, 23-38.

Oskarsson, M., (1989). Self-assessment of language proficiency: Rationale and applications. Language Testing, 6, 1-13.

Oxford, R.L., Nyikos, M., (1989). Variables affecting choice of language learning strategies by university students. Modern Language Journal, 73, 291-300.

Oxford, R., & Burry-stock, J. (1995). Assessing the use of language learning strategies worldwide with the ESL/EFL version of the strategy inventory for language learning (SILL), System, 23, 1-23.

Schunk, D. H. & Zimmerman, B.J. (1998). Self-regulation and learning: From teaching to self-reflective practice. New York: Guilford.

Sheorey R., (1999), An examination of language learning strategy use in the setting of an indigenized variety of English. System, 27, 173-190.

Smith, H., Cooper, A., Lancaster, L., (2002). Improving the quality of undergraduate peer assessment: a case for student and staff development. Innovations in Education and Teaching International , 39, 71-81.

Stefani, L., (1994). Peer, self, and tutor assessment: Relative reliabilities. Studies in Higher Education, 19, 69-75.

Taras, M., (2002). Using assessment for learning and learning from assessment. Assessment and Evaluation in Higher Education, 27, 501-510.

Wen, Q.F. (2003). A Road Leading to the Success of English Learning. Shanghai

Wenden, A.L., (1986). What do second-language learners know about their language learning? A second look at retrospective accounts. Applied Linguistics, 7, 186-201.

Wenden, A.L., (1987). How to be a successful language learner: insights and prescriptions from L2 learners. In A. Wenden and J. Rubin (Eds.), Learner Strategies in Language Learning (pp. 103-118). Englewood, NJ: Prentice Hall.[/private]

Category: Monthly Editions, Volume 18