Teacher-student collaborative assessment ( TSCA ) in integrated language classrooms

Assessing students’ productive performance is challenging in China because large class sizes inevitably lead to heavy workloads for teachers. To address this problem, a new method of assessment, teacher-student collaborative assessment (TSCA)—was proposed in 2016 to organize and balance different modes of teacher assessment, self-assessment, peer assessment, and computer-mediated assessment. The present study took one intact class as a case, aiming to explore how TSCA could be carried out efficiently and systematically in the classroom and how students perceived TSCA. Qualitative data obtained include students’ writing drafts and revision, interview, and reflective journals of the students and the teacher. Interview data indicated that the students responded to this type of assessment positively and thought they benefited greatly from the teacher’s instruction and peer discussion. This was triangulated by the students’ reflections in which all the students spoke highly of TSCA and agreed that this method was a good way to pinpoint their weaknesses and help them learn how to revise their essay better. The students reported that they formed a new perception of self-assessment and self-revision and felt that a lot was gained.


INTRODUCTION
Assessment, a natural part of the teaching and learning process, is administered to support learning (Berry, 2008); therefore, it has a huge impact upon teaching and learning (Baird, Andrich, Hopfenbeck, & Stobart, 2017;Wiliam 2017).This notion, which emphasizes the role of a learner and teacher in assessment, results from a paradigm shift from input-oriented norm referenced assessment to outcome-based and standard-referenced assessment (Davison & Cummins, 2007).Recent studies on this line of research, especially formative assessment, stress the importance of this learning and teaching potential (e.g., Black, 2015;McDonald, 2018;Shepard, Penuel, & Pellegrino, 2018;Taras, 2008).
Different from its traditional sense of evaluating or grading, assessment here mainly means providing helpful feedback and revision to improve learning.
In language education, second language (L2) teachers spend a significant amount of time responding to students' written work believing that problems in students' writing need to be dealt with, and errors need to be corrected.L2 learners also value and want more written comments from teachers (Lee, 2008).Two meta-analyses showed that there was a medium overall effect of oral corrective feedback (CF); the effect was maintained over time (Li, 2010); and written CF could lead to greater grammatical accuracy in L2 writing (Kang & Han, 2015).All these studies have proven that Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN:2301-9468 providing CF as a form of assessment is widely agreed both in practice and in research.
In the EFL context, what teachers in mainland China have been doing is similar to what Lee, Mak & Burns (2016) described: Mark single drafts, focus predominantly on errors, and respond to errors comprehensively.This 'teacher-only' approach is timeconsuming and discouraging, it may not come to any fruition.For one, it is no easy feat for teachers to keep up with assessing large classes with 50 to 60 students in particular.For another, students might not even read teachers' comments or revisions carefully since what they care most is the score on the paper.The effectiveness is compromised even for those who do read the teacher-written feedback, since students might not understand why it should be corrected this way, and it cannot be guaranteed that they will not make the same mistakes again.
In order to overcome the limitations of teacher assessment and lessen the workload, some teachers seek to adopt alternative forms of assessment: selfassessment and peer assessment (e.g., Hanjani, 2016;Yu & Hu, 2017;Yu & Lee, 2014).Students' involvement in the assessment process can enhance their learning autonomy and help them process linguistic features deeper which lead to 'internalization.'However, peer assessment and self-assessment studies have indicated that students need training and experience in order to perform these tasks effectively (e.g., Freeman, 1995;Jafarpour, 1991).Often times, students do not know exactly how to revise their own or peer's written work, or they may not trust their own or peer's revisions.This inability to find errors (Sun, 2017a(Sun, , 2017b)), distrust in the peers' correction (Tsui & Ng, 2000), and habitual focus on surface-level mistakes (Hanjani, 2016;Khonbi & Sadeghi 2012) threaten reliability and validity of such assessment (Blanche, 1988).
Still, many teachers have resorted to technology in the form of automated scoring systems to use it alone or to supplement teacher assessment (e.g., Burstein, Chodorow, & Leacock, 2014;Chen & Cheng, 2008).These computer-mediated assessment tools are fast and saves teacher time: with a click of a button the scores are shown; grammatical mistakes are revealed; or alternative expressions are suggested.However, these conveniences have to be weighed against their shortcomings.For one, a sentence may be grammatically correct, but the content could be totally off (Attali, 2004).Likewise, the structure of sentences may not make sense in relation to each other.In addition, the automated assessment tools cannot tell whether the objectives are achieved or not (Sun, 2017b).In a word, machines cannot address the aspects of writing that require human evaluation such as communicative effectiveness, styles, relevant content, and audience awareness.
Given the aforementioned weaknesses of teacher, self-, peer and computer-mediated assessment in EFL writing contexts, the current study introduces a new method of assessing students' work-teacher-student collaborative assessment (TSCA) (Wen, 2016a(Wen, , 2016b))-to tackle the challenges of assessing students' production: low efficiency and poor effectiveness without diminishing the production or compromising the feedback.TSCA is mainly carried out in class with the aim to assess a selective few of students' written compositions finished after class.
Compared to other forms of assessment (See Table 1), TSCA has three prominent features (Wen, 2016b).The first feature of TSCA is collaboration of students and teachers.TSCA is not a simple combination of teacher assessment and self-assessment or peer assessment and machine assessment (e.g., students selfedit the first draft, and teacher edits the second draft), but rather, it is a joint assessment where the teacher selects a sample of students' typicalwritten products on a certain task which in turn is then assessed by students and the teacher collaboratively in class.Secondly, in TSCA, both a teacher and students check both learning outcomes and the quality of the work.TSCA is not confined to assessing students' language products, but it also includes examining whether the students have achieved the objectives of the learned unit.Specifically, this form of assessment attaches great importance to evaluate how well the learning objectives are achieved, rather than the quality of products alone.Another feature of TSCA is multiple assessments.After in-class TSCA, students have a better understanding of how to assess, which serves as their starting point to carry out effective self-assessment or peer assessment after class.Afterwards, the revised version can be assessed by an automated scoring system where students can see their scores and read some general comments.The records of their improvement can also be saved.
The current study aims to probe whether this form of assessment can serve as an effective alternative to existing assessments and students' perceptions of it.Specifically, the following research questions were put forth: 1. How can TSCA be carried out in the classroom?2. Is it effective?3. What are students' perceptions of TSCA?

Context and participants
As mentioned before, university English teachers in mainland China adopted the teacher-only approach when assessing writing.Tsui and Ng (2000, p. 149) found that Hong Kong students viewed the teacher as a 'figure of authority that guaranteed quality.'This is also the case for Chinese mainlanders.The expectation of guaranteed quality placed on the teacher is exacerbated at the tertiary level with larger class sizes.Assessing students' written work becomes a real chore and a headache when the number of students reaches 80 or even 140 per class.Keeping up with assessing and giving feedback in such large classes has become an Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN:2301-9468 unmanageable undertaking for teachers.No wonder some scholars conclude that assessment has become our enemy and is considered a curse (Lee, 2018).Teachers have tried alternative assessments (Alderson & Banerjee, 2001), such as self-assessment or peer assessment, and more recently have resorted to automated scoring systems (Sun, 2017a, Online Survey) to alleviate teachers' workloads.However, all these types of assessment are not effective and efficient at the same time (see the introduction of article), making a new form of assessing highly urgent.TSCA is designed to draw on the strengths of these other assessments, hoping to tackle the challenges faced by Chinese teachers (see Table 1 for a comparison between TSCA and other types of assessment).The present study lasted two academic semesters from September 2016 to June 2017.It was conducted in the naturalistic settings of an EFL integrated course designed for second-year English majors in a university in mainland China.The teacher, also the researcher, is a non-native English speaker who has been teaching at the university for 13 years.The 24 students (6 male, 18 female) were 18-19 years old with Chinese as their mother tongue.Their English language proficiency was at an intermediate level.
The integrated English course is a required course for first-and second-year English majors.It is composed of a series of theme-based units to enhance students' comprehensive English competency.The course is taught twice a week for 16 weeks per semester, with each class session lasting 90 minutes.Students took listening, reading, speaking and writing courses in their first year and continued with these courses in their sophomore year.The cohort had no prior experience of TSCA before this study.

Instructional procedures
TSCA was composed of three phases: pre-class, inclass, and post-class (Wen, 2016b).The pre-class phase was the preparation stage in which the teacher went through students' work and selected and graded a 'typical' sample of written products to be assessed by students and the teacher collectively in class.The inclass phase was the assessment stage where the teacher presented the selected work (SW), and students worked in groups of four to discuss and revise it, and then proceeded to share the revision with the whole class under the teacher's guidance.This stage incorporated 'teaching" in 'assessing' where students knew not only how to assess and but why assessing had to be done in this way.The post-class phase was the revision stage where all the students revised their products after class.These three phases completed a TSCA cycle.
Post-class revisions were a crucial phase of TSCA.After the collaborative assessment, self-assessment or peer assessment after class (guided by the TSCA focus in class) could be of great value for the students to reinforce what they learned.Afterwards, the revised version was assessed by the automated scoring system where students could see their scores, read some general comments, and saved a record of their improvement.The teacher could also recommend quality examples for students to read, which could serve as a great incentive for them to write more.
The 8 TSCA sessions are briefly described in Tables 2 and 3.In the first semester, four theme-related after-class writing activities were assigned (one every 4 weeks).Three of them were related to themes covered in class, and one was related to a movie (See Table 2).For instance, Task 1 was an extension to the text "What Makes a Global Language."The TSCA cycle started with the submission of the first draft and ended with the submission of the revised version (see Figure 1).The TSCA was carried out after students' first written drafts were collected.Each TSCA in-class session lasted between 20-40 minutes, depending on the focus of assessment and schedule of the whole course.
In the second semester, four content-related writing assignments were given within 2 weeks under the same theme of Kindness and Indifference (See Table 3).Each assignment was related to the content discussed in class without requiring any further extension.Though students were not asked to revise their drafts, since they needed to hand in a new composition after each class, they had a chance to revise the first 3 pieces of written work in the last writing which included all the previous three (See Figure 2).
Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN:2301-9468  First, in classroom observations, the TSCA sessions were audio-recorded (totaled 120 minutes) and video-recorded (totaled 125 minutes) in the first and second semesters respectively.Students' group discussions were also recorded (approximately 200 minutes) using their own mobile phones.The teacher observed the class while teaching, watched the videos, and listened to the audios after class in order to see how well the students were engaged in the TSCA.
Second, in the first semester, two semi-structured interviews (approximately 100 minutes) were conducted in Chinese and recorded after the first and second TSCA cycles.These interviews were designed to probe the students' experience of engaging in the TSCA in class and revision after class.After the first writing task, 6 students were purposefully selected to be interviewed based on their revisions: 2 were from the 'well-revised' group, 2 from the 'ill-revised' group, and 2 from the 'mediocre-revised' group.After submission of the second writing task, the second interview was conducted.Eight students (other than the 6 who participated in the first interview) were randomly chosen and individually interviewed for approximately 7 minutes each.
The interviewees had the opportunity to reflect on several aspects of the TSCA, including: (1) whether they could identify the problem in the selected sample, (2) their revision, and (3) their perception about the TSCA.Students were assured that the interviews were for the teacher to diagnose and improve her teaching only and that their responses would have no effect on their final grade.
Last, after each cycle of TSCA, students wrote reflections (in Chinese) on their writing, TSCA and revision from four aspects: overall evaluation, gains, problems and suggestions.The journals were tagged as student 1, 2, 3 etc., based on their student ID.The teacher also reflected upon the classroom practice of TSCA, mainly from two aspects: the effectiveness and problems arising from it.

What to assess
There is a slew of varying problems in terms of language, content, and structure in students' writing.Table 4 lists the common problems that teachers might encounter in students' written work.As briefly outlined here, problems could be language-related such as grammatical mistakes or wrong word choices; the content of the writing was not clearly stated or supported; and sentence structures within and between paragraphs could be very weak.As a result, the communicative goals were not achieved in an effective manner.In a reflection, 75% of the students reported that their ideas were not expressed in the way they wanted due to their limited language repertoire.
It was simply not practical to deal with the problems mentioned above all at once, but to focus on one of them at a time.Studies (e.g., Bitchener & Ferris 2012) showed that focused CF is more effective than unfocused CF.Selective focused correction gives students precise feedback which directs them to address specific errors, helping them to not only edit the current essay and but to avoid or reduce such mistakes in the Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN:2301-9468 future (Bitchener & Ferris 2012).Without such a clear focus, attention would scatter on multiple tasks and efficiency in learning would be lowered (Wen, 2016a).
In order to decide what to focus on for assessing, the guiding principle of the 3Ps is suggested as follows.The first is prominence-if more than 50% of the students have the same problem, then it is prominent enough to require attention.In other words, if only 10% of students have the problem, it is not prominent enough to garner consideration in the TSCA.The second is progress.TSCA is not supplementary to learning; instead, it is learning, aiming at the improvement of learners' language proficiency bit by bit.Progress here does not necessarily mean a big leap forward.Choosing a focus that helps tackle a 'small' problem or even raise consciousness is what TSCA aims to fulfill.Assessment will not work well if the focus of attention is so big that it is beyond the control of the teacher.The third is progression.The order of the focus has to be progressive, starting from the basics first and then step-by-step working up to the next level.This can be illustrated by my classroom practice.Tables 1 and 2 outline the foci of assessment covered in an academic year: the predicate verb, introductory paragraph, titles, supporting evidence (1st semester), the setting of a narrative story, lexical variety, topic sentences and transitions (2nd semester).It would not have been logical to work on transitions first if the students had not even mastered the 'topic sentence' phase.Similarly, it would not be right to focus on the body paragraph before the introductory paragraph.Therefore, the order of what to focus on is important because each concept should be built on one another.
The 3Ps can be better illustrated by using the four cycles of TSCA in the first semester.For the first TSCA, linguistic features, specifically the predicate verbs, were chosen to be the focus for 3 reasons.Firstly, these errors frequently occurred in students' writing.The fact that 22 out of 24 compositions (92%) had this problem made it so prominent that it deserved attention.Secondly, grammatical accuracy is the basic requirement of writing and the correct usage of main verbs in a sentence is fundamental to writing.It is the simplest to be handled.Thirdly, this focus is teachable since students are familiar with 'error correction,' and it could be dealt with successfully within a limited time.The following sample is what was worked on in class.
English has become an important communication tool for people all over the world.Everyone are able to understand each other because of the tool.It no doubt that the tool has promoted greatly the communication of economy, policy, and culture.Therefore, the counties in the world have connected more tightly.It also have an important influence on world peace.

How to Assess
This section details how TSCA was carried out under the teacher's guidance in class by using two examples.
To begin with, the teacher presented to the class a sample of students' writing as shown in the left column of Table 5 but did not tell them what the problem was.Students were first asked to make some comments on the good and weak points.One student said, "the author is trying to lead you to the scene."Another student said, "it brings us a bright beginning, so the accident will be more shocking for us."However, they failed to find out the problem until the teacher stepped in and mentioned that this writer used 67 words to approach the scene and 29 words to describe the scene.The objective was to describe the bystander's psychological reaction to an emergency, so the bulk of the writing should be devoted to the description of the incident, and the bystander's inner thought.However, this student spent a lot of effort describing the setting of the story, which was too lengthy.
After the problem was identified, students were given a short time to revise it individually first, and then worked in groups for a better revision.When they were ready, they volunteered to share their ideas with the whole class.The teacher's version was then presented/revealed (see Table 5, right column), with the setting of the story cut down to one sentence (21 words), once again emphasizing that it is not advisable for the setting to take too much space.

The Sample
The Revised (teacher's version) It's a lovely afternoon though nearly sunset.I was walking home like usual after hanging out with my old friends.And I unconsciously glanced a little girl, having fun with herself just besides the street, and it is a narrow street.At that moment, I was curious about what she was doing, so I looked directly at her while I was walking.Before I found out what's she doing, the crash happened.She seems did not notice that van was coming for her, and all of a sudden, she laid on the ground, without any movement… On a late afternoon, I was walking home on a narrow street like usual after hanging out with my old friends.A toddler was waddling in front of me, not realizing that a van was coming, and all of a sudden, she was knocked down by the van.The driver hesitated for a second and drove off ...
The second example of TSCA elaborated here is lexical variety.When explaining the reasons why people do not help in emergencies, students used the words 'people,' 'help,' and 'indifferent' repeatedly to such a degree that the effectiveness was compromised and the alternative expressions learned in class such as "apathetic people; offer assistance" were not used!This problem was brought to the students' attention through the two samples as shown in Table 6 where the words 'indifferent' or 'people' were overused.
Each sample was shown on PowerPoint slides one at a time.After discussion, students shared their revision with the rest of the class.In their try-out revision, they simply deleted the word 'indifferent.'As a teachable moment, the teacher highlighted that to achieve 'lexical variety,' they could either use synonyms like 'apathetic' or 'callous' in this case, or vary the forms of the same word such as 'indifference.'The teacher guided the discussion and the revision along the way.In the revision phase, the focus was not only on how to revise it, but on reasons why it needed to be revised in the way it was.Students needed to know the underlying cause of the problems before they could identify them and proceed to edit their own work all by themselves.
Students played an indispensable role in TSCA.They thought about or edited the selected sample alone first and then engaged in group discussion to figure out the problem in the sample and discussed how to revise it in the later phase.Their engagement was constantly guided by the teacher.After revising the first sample, more practice was provided related to the focus.When dealing with the lexical repetition, the following sample with the overuse of 'people' was selected for them to revise in class.
So why people do not help in emergencies, maybe there are so many reasons for themselves.Some people would say they have no ability to save people, and they also do not know how to rescue people, because they are not professional that they cannot lend a hand easily.And some people would think about what if the person that they helped accuse them on the contrary.

The effectiveness of TSCA Students' final products
The last writing activity covered all the previous three tasks assigned before (see Figure 2).By comparing the final product with their first draft, the progress that students made can be clearly seen.In one student's (Student 8) second writing, the word 'people' was used a great deal (left column, Table 7), but this was improved in Task 4, in which 'people' was replaced by 'bystanders,' 'onlookers,' and 'passersby' (see right column of Table 7).
Similar improvements can also be found in student 12's work.After the second and third TSCA covering lexical variety and how to write good topic sentences, he changed his topic sentences by adding more details to make them clearer.His original "to be the first one" was rather vague.However, after TSCA, he revised it to "to be the first one to offer assistance" which was much clearer.It is noteworthy to mention that he avoided the repetitive use of the word help in the same paragraph by using "offer assistance" and "lend a hand" instead.
Students' development in writing can be illustrated in student 20's revision of the fourth task.After the last TSCA-using transitions to link paragraphs-was implemented, students came to realize the necessity and "strategy" to make their arguments flow by summarizing the content of the preceding paragraph before moving on to guide the reader to focus on what to look for next.In the first draft, paragraphs 2 and 3 were two 'unrelated' paragraphs.However, this was corrected in the revised draft by using a subordinate clause led by "although," thus linking the main ideas of the previous paragraph to the next one.It is worthwhile to mention that when the student uploaded the revision online, the automated scoring system commented: "The author achieved fluency by means of some simple cohesive devices." Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN:2301-9468 First of all, to be the first one.As for this part, there is a rule that everybody will do what the others do.So if he doesn't help, they will not too.Then, it is someone to be the first person that will encourage people to give a hand.What's more, to make the laws.It means that the government will stand at the rescuers' side by making laws.

Topic sentence
Task 4 Helping in emergencies (Excerpt-Student 12) First of all, to be the first one to offer assistance.
There is a rule that everybody will do what the others do.So if he doesn't help, they will not too.
Then, it is someone to be the first person that will encourage people to help.What's more, to make the laws to protect those who lend a hand.It means that the government will stand at the rescuers' side by making laws, which is protecting the rights of rescuers.

Students' perceptions
In students' reflections, all the students spoke highly of TSCA and agreed that this method was a good way to pinpoint weak points and improve their writing.They reported that they formed a new perception of "selfassessment" and "self-revision" and felt much was gained from TSCA.
By the end of the first semester, students recognized the weaknesses and problems in their writing.They were conscious of their own mistakes, and it is precisely this realization that is the prerequisite for progress.Eleven students, accounting for 46%, explicitly mentioned their awareness of "prevalent problems which were often overlooked before": This "no-revision" practice was replaced by "multiple-revision" practice by the end of the semester.Six students (25%) expressed the importance of revision: …I revised my writing to the degree that there is no room for me to improve.After 4 TSCAs, I came to know that good writing comes from constant editing and writing ability can be enhanced through revision.(Student 13) Knowing weaknesses in writing and redrafting are a must to enhance learning.The cognition of problems leads to correction, which in turn leads to improvement.Twenty students (83%) submitted their revised draft to the online automated assessing tool after an average of 4 revisions.
After a year of classroom practice, all the students (100%) reported that TSCA is conducive to writing revision.In addition to awareness and the change in their habit, students also acknowledged their gains from group discussion and the teacher's instruction.
1) Gains from peer discussion: A third of the class (8 students) wrote in their reflection that they benefited greatly when discussing with their peers.In the group discussion, students are able to know other view points and different opinions which in turn spark ideas of their own.For instance, student 5 reflected: especially with the ones who are better than me, made me understand how they organize their sentences or even the whole passage.I came to know how they bring their thoughts together to make a logical argument, which is a great help to me.
TSCA created an active and dynamic classroom atmosphere.Listening to their peers' revision, students could reflect on whether their own ideas were correct or not.On top of that, after they were familiar with procedures and objectives of assessment, a growing number of students were stimulated by their peers, switching from "listeners" to "participants."This changed the classroom dynamics as everyone served as "scaffolds" for everyone else.
2) Gains from teacher's instruction: One of the crucial features of TSCA is the teacher's guidance.From determining a focus to selecting typical samples, from guiding the students to realize the weaknesses in the sample to leading them to work collaboratively, the teacher guaranteed the effectiveness of TSCA.Thirteen students (accounting for 54%) mentioned the important role of the teacher in their reflection.
"In class, my teacher pointed out some problems of our compositions.I was being very "lucky" since those problems were also mine.After the teacher's instruction and my own revision in line with her instruction, I felt I improved a lot" (student 4).
"I realized why I made such mistakes as "inappropriate titles" or "illogical sentences" after the teacher's explanation" (student 7).

"In class, my teacher pointed out the issue of repetitive use of words, such as people, help. It left me a very deep impression, since it is also my problem. After assessment, I revised my writing and felt that I improved a lot in expressing my ideas. And, I'm also aware of the lengthy setting now" (Student 20).
Although TSCA aims to reduce teachers' workload, it does not decrease their responsibility (Wen, 2016b).In TSCA, teachers are "decision-makers" as well as "scaffolders."On the one hand, the teacher determines the focus of the assessment according to students' proficiency level and their written work.On the other hand, it is the teacher who dynamically adjusts the pace and content of assessment according to the time limits in class and how well students respond to it.Teachers provide the professional assistance in TSCA, which is more advantageous than self-assessment or peer assessment.

CONCLUSION Contributions of the study
This study made an attempt to address the challenges in assessing students' written work in a relatively large class and in doing so has advanced the assessment research in three aspects.First, in its theoretical contribution, it proposed a new form of assessment, TSCA, so as to maximize the effectiveness and efficiency of assessing.In addition, the study strived to apply TSCA to classroom teaching, showing how it could be done in practice.Lastly, there are pedagogical implications for teachers who may want to try TSCA out in their classrooms.The theoretical, practical and pedagogical contributions of this study have hopefully furthered the research in the assessment field and classroom practice.
To date, numerous studies have investigated and compared the effectiveness of different forms of assessment: teacher assessment, self-assessment, peer assessment, and machine assessment, but little attention has been paid to their incorporation with classroom practices, let alone the integration of these forms into one holistic assessment to maximize effectiveness.The present study proposed TSCA to organize and balance the different modes of assessment and was implemented in a classroom setting with peer assessment and automated scoring systems as a supplement after class.The clearly-defined pre-class, in-class, and post-class procedures serve as the basis for its application in natural or authentic classrooms.The initial classroom practice helped further develop the theory by adding guidelines and strategies to guarantee effective assessment.
Based on a one-year reflective practice, this article proposed a set of guidelines for selecting a focus to assess and illustrated how TSCA was carried out to ensure that assessment is for the purpose of learning.Students' progress could be seen in the first and second drafts of the last writing task.One focus at a time proved to be useful in directing students' attention on one point.It was relatively easier and more beneficial for them to pay attention to one issue alone when revising.As to how to carry out a focused in-class assessment, a teacher followed the procedures of problem-identification and sample-revision with necessary instructions.This was done by always encouraging students to figure out the problems and work out the solutions on their own.Assistance was provided only when needed.
Pedagogical implications for teachers also emerge in connection with students' reflection and interview data.An in-depth analysis of these revealed that the Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN:2301-9468 students held predominantly positive attitudes towards TSCA.They confirmed that TSCA in class provided them with an opportunity to discuss and share ideas with each other which facilitated the revision process after class.Moreover, learners started to realize that a good piece of writing took multiple revisions, and they were willing to make the effort to rewrite their essays more than one time.However, learning is not a linear process, so teachers cannot expect students to acquire fully something new or to correct their errors with just a one-time assessment.TSCA raises learners' consciousness of the problem so that they can bear it in mind when completing the next writing task.Moreover, teachers can further integrate assessing with teaching, reinforcing the teaching by assessing and reinforcing assessing by further teaching.In this way, learning and improvement in writing will come about as a result.

Reflection and future directions
We argue that TSCA is an effective approach and another option for language teachers to assess students' written work.TSCA holds the view that assessment is learning where it achieves both efficiency and efficacy.It is efficient because the teacher does not have to grade every individual paper.However, its effect is not compromised since it is effective in targeting students' needs.TSCA draws out the common problems for students to notice the gap between what they wrote and what they are supposed to write.These are the teachable moments that the teacher can take advantage of when students have the readiness and strong incentive to learn.Most importantly, TSCA is a valuable approach to enable students to learn how to write.
TSCA is still in its infancy; therefore, a dynamic research agenda is imminent for some time to come.Three areas would benefit from further exploration on TSCA.First, future empirical research of experimental design is needed to explore the relative effectiveness of TSCA since the present study is only qualitative.Second, a further theoretical building of TSCA would help improve classroom practices.For instance, when carrying out TSCA, teachers often encounter challenges such as how to provide professional guidance.A detailed study on this would empower teachers to use TSCA more effectively.Last, further studies can also be conducted to see whether applying TSCA to oral work would work well inasmuch as this study only focused on the effectiveness of TSCA to written work.
Figure 2 TSCA cycles in the second semesterData collectionClassroom observations, teaching portfolios, semistructured interviews, and reflective journals were used to document the implementation of TSCA and the students' and teacher's views on it.The use of these aimed to give a more comprehensive depiction of TSCA rather than draw conclusions based from one method alone.Data collected included video and audio recordings of the class (teacher's instructions and students' discussions), students' written work (first draft and revised version), interview recordings, and journals written by both the teacher and students.First, in classroom observations, the TSCA sessions were audio-recorded (totaled 120 minutes) and video-recorded (totaled 125 minutes) in the first and second semesters respectively.Students' group discussions were also recorded (approximately 200 minutes) using their own mobile phones.The teacher observed the class while teaching, watched the videos, and listened to the audios after class in order to see how well the students were engaged in the TSCA.Second, in the first semester, two semi-structured interviews (approximately 100 minutes) were conducted in Chinese and recorded after the first and second TSCA cycles.These interviews were designed to probe the students' experience of engaging in the TSCA in class and revision after class.After the first writing task, 6 students were purposefully selected to be interviewed based on their revisions: 2 were from the 'well-revised' group, 2 from the 'ill-revised' group, and 2 from the 'mediocre-revised' group.After submission of the second writing task, the second interview was conducted.Eight students (other than the 6 who participated in the first interview) were randomly chosen and individually interviewed for approximately 7 minutes each.The interviewees had the opportunity to reflect on several aspects of the TSCA, including: (1) whether they could identify the problem in the selected sample, Task 4 Helping in emergencies (Excerpt-Student 20) (Paragraph 2) There are many reasons why bystanders are reluctant to raise eyebrows in emergencies… (Paragraph 3) I think that kindness of strangers still shine brightness in our daily life … Transitions Revised Task 4 (Excerpt-Student 20) (Paragraph 2) There are many reasons why bystanders are reluctant to raise eyebrows in emergencies… (Paragraph 3) Although having sufficient reasons for bystanders hesitating to give a hand, the kindness of strangers still shine brightness in our daily life… …After TSCA, I am aware of these problems.Every time before submission, I would read it over and checked if I avoided making mistakes mentioned by the teacher… (Student 15)This contrasted with data from students' first interview:I have no habit of revising my writing.Most often than not, I would hand it in without reading it for a second time, just like what my other classmates would do…(Student 2) think the discussion in the classroom is also very effective.Everyone has their own point of view.Through the discussion, I have broadened my own thinking.Student 8 echoed: Discussing my writing with my classmates, Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN:2301-9468

Table 1 .
The comparison of TSCA with other types of assessment

Table 2 .
Summary of TSCA carried out in the first semester Doctors have pondered the connection between mental and physical health for centuries.Some researchers have found the healing power of emotion in curing the most serious diseases.Others, including physicians, on the contrary, doubt the effect of emotion in health.What is your opinion on this?Write an article of 200 words on this topic.

Table 3 .
Summary of TSCA carried out in the second semester

Table 4 .
Common problems in students' writing

Table 6 .
A sample of overuse of a certain word

Table 7 .
A comparison of written work before and after TSCA