The Effect of Feedback on Student’s Writing Accuracy
Asep Suarman
Asf_suarman@yahoo.co.id

Abstract
This paper is concerned with the effect of direct, uncoded oral and written feedback to the writing accuracy. Nine secondary school students, who were chosen purposively from high, mid and low achiever, involved in this study. Observation and document analysis were used to collect data to which errors of grammatical - covering subject-verb agreement and clause construction -, lexical and punctuation were analyzed. The result shows that the direct, uncoded oral and written feedback gave significant effect only to the high achievers. But, such feedback did not give much effect to the texts made by mid achievers and low achievers. Overall, the feedback did not give significant effect to the student’s writing accuracy. Considering limitations of the study, some suggestions are also put forward.

Introduction
This paper deals with the investigation of direct, uncoded feedback to writing accuracy. This subtitle discusses the background of the study, theoretical support, previous research, research question, the significance of this study.
Among the four language skills (listening, speaking, reading and writing), writing is the latest and the most difficult skill to acquire. It is naturally acquired after listening, speaking and reading. It is considered difficult as it requires complex competent about the rule of language.
Related to writing, many research shows that Indonesians, not only ordinary people who does not deal with books or science every day, but also intellectuals lack of writing (Alwasilah, 2001). It is probably because writing is often neglected in school lessons. In EFL context, many teachers might emphasize on other skills.
In recent English curriculum of Junior high school, school based curriculum, writing is more emphasized than other skill (Emilia, 2010) since it applies genre based approach (GBA). In GBA, teaching process is applied in two cycles of teaching, oral and written ones, wherein four stages are employed. The stages are building knowledge of field (BKoF), modelling of text (MoT), joint construction of text (JCoT) and independent construction of text (ICoT) (Depdiknas, 2004). In the last two stages, teachers’ job was to give feedback to better students’ performance.
Feedback is generally defined as the return of information about the result of a process or activity; an evaluative response: asked the students for feedback on the new curriculum (http://www.thefreedictionary.com). It is included into correction or praise (Ellis, 1994; Harmer, 2007a). Giving feedback should suit with the skill taught. Feedback on speaking and writing is different in nature. So, a teacher should carefully design and conduct it.
Writing skill is often neglected in school (Alwasilah, 2001). Many teachers probably only ask their students to write a text then give mark on it without giving more attention how to better students’ writing. It is perhaps simply because it needs more skill, time and energy for teachers to give feedback to their students writing.
In giving feedback to students’ written work, Williams (2003) identified two common categories: feedback on form and feedback on content. The common feedback in the former, William argues, refers the outright teacher correction of surface errors; which require students to copy the correction and the teacher indicate the place and type of error by giving only markings but without correction, only underlining to indicate only the presence of errors; which requires students to correct the errors on their own. The feedback in the latter, to William, consists mainly of comments written by teachers on drafts that usually point out problems and offer suggestions for improvements on future rewrites, in which students are usually expected to incorporate information from the comments into other versions of their papers.
Nevertheless, William (2003) argues, giving feedback sometimes works and sometimes not. Citing Fathman and Walley (1990) William argues, when the grammar feedback is given, that indicated the place but not the type of errors, students can improve their compositions in their subsequent rewrites them. Indirect written feedback is also more useful than direct correction (Frodesen, 2001, in William 2003) and will be more effective if being coupled with student-teacher conference (citing Brender, 1998; Fregeau, 1999), wherein teacher can ask more questions and students can express their ideas. In giving indirect feedback, some teachers usually tend to code mistakes to indicate the precise location and type of error, while others provide uncoded feedback that simply locates the error without disclosing the error type. Usually with uncoded feedback, it becomes the student’s task to diagnose and correct the mistake (Hartshorn, 2008).
However, the feedback on surface errors does not work when it is inconsistent, unclear and overemphasizes the negative feedback (Fregeau, 1999; Cohen, & Cavalcanti, 1990 in William 2003). The students mostly only copy the corrections on their subsequent drafts or final copies without recording or studying the mistakes noted in the feedback as they do not understand why the words or phrase are indicated as errors.
William (2003; see also Stitt-Bergh, 2007) suggests teachers choose an effective method of feedback considering the goal of the course, the shortcomings of common methods of feedback, the positive aspects of them and what students want about the feedback; and to lessen student confusion, teachers should consistently use a standard set of symbols or markings to indicate place and type of error and train the students in what kinds of corrections to make based on each symbol that is available on most writing textbooks, or teacher-made on their own which should be familiarized by their students.
Some functions of the teachers’ feedback in correcting students’ written work, according to Iseni (2011) are to help the teachers improve their work in this direction; to focus attention on accuracy and content; to help both teacher and students together correct written work and oral work and to help teachers not only correct spelling, grammar, lexical and other mistakes but also these corrections might be accompanied by certain comments on the content of the written work, showing the student where the work was effective and where it was not.
Further, there are different procedures to help teachers to correct written work namely, according to Iseni (2011; see also Brown (2001, pp. 355-6) for feedback guidelines; Harmer, 2007 p. 149 for the feedback symbols), a teacher can underline the mistakes and write the right symbol in the margin on the same line; or do not write the symbol to helps the students find the type of the mistake themselves; write the symbol on the margin but do not show exactly where the mistake is, and show only the line. This helps the student find the mistake, helped by the symbol put on the margin; or put a cross (X) on the margin, put as many crosses as mistakes and let students find the mistake and the type in one line or put a cross next to the line in the margin, but do not show how many mistakes are and let the students think about mistakes, knowing that there is something wrong in a particular line. Harmer (2007) adds that a teacher can correct students’ work by putting ticks on good points or other symbols or might write summarizing comment at the end of students’ work.
On the top of that, some studies have investigated feedback with their own various focus, participants and results with also various recommendations.
The contradictive findings denying the effect of feedback on students’ writing are pun forward by Truscot (1996, 2007 in Harshorn, 2008; in Pan, 2010) that the feedback does not have any impact on their accuracy. This is also supported by Fregeau (1999 in William 2003); Cohen, & Cavalcanti (1990 in William 2003) who suggested that the feedback on surface errors does not work when it is inconsistent, unclear and overemphasizes the negative feedback. Fleck and Ledder, (1998 in Pan (2010); Sheppard, (1992 in Pan, 2010) and Pan (2010) also found that the teacher error feedback alone may not facilitate the learning of linguistic information or even do not have significant effect on the writing accuracy. Tsao (2010) also partly supported that error correction cannot improve the accuracy of student writing except for high proficiency students.
Other studies concluded that the feedback did have effect on the students’ writing ability. Usaha (1998) revealed that a highly significant improvement in the students’ holistic writing and reduction of errors, but there was no significant difference in the length of writing in comparing first and last writing and the highest error rate was found in wrong word followed by sentence structure, verb, article, and noun ending respectively. Hartshorn (2008) found out that the manageable, meaningful, timely, and constant feedback treatment improved mechanical accuracy, lexical accuracy, and certain categories of grammatical accuracy, but had a negligible effect on rhetorical competence and writing fluency. Bitchener et al. (2005) found a significant effect for the combination of written and conference feedback on accuracy levels in the use of the past simple tense and the definite article in new pieces of writing. Grami (2005) supported that ESL students do want and appreciate the teacher feedback in any kind of it. Finally, Tsao (2010) found that error correction is effective in promoting the overall quality of student essays, particularly for intermediate to high proficiency students.
Based on the contrastive result above, this study is intended to shed light on the following research question: Is there any effect of direct uncoded oral and written feedback on students’ writing accuracy?
Hopefully, this study might have at least three significance: theoretically, it may add empirical support to existing theories and findings ; practically, its results may help to clarify the benefits and detriments of giving feedback to students’ writing; and professionally, the teachers in the research site may learn how to give appropriate feedback to students ‘ draft texts.
Methodology
This study involved only nine subjects of grade nine of junior high school students. They have learned English for about 2,5 years and were chosen purposively based on their daily performance; the achievement in previous reading tests was used as the basis. Three students were categorized as the high achievers, three from the medium achievers and the rest from the low achievers. They studied English twice a week, 80 minutes each, every Tuesday and Saturday.
To collect data, document analysis, writing tests and text analysis were used. At the time of the study, none of the participants joined any English course.
As this study was qualitative, the researcher himself acted as a research instrument besides writing test and text analysis. The feedback investigated was direct and uncoded oral and written one since it was simple, direct and does not need any coding to agree with. The material on the lesson plan was writing report text, which was one of essay text to be taught in odd term of grade 9 of junior high school.
As the curriculum applied Genre Based Approach (GBA) wherein BKoF, MoT, JCot and ICoT stages were done, the feedback were given by the teacher on JCot and ICot stages. In JCoT, feedback was given to groups of students, in ICoT, to individuals. The latter was the one studied. Different from JCoT stage wherein the groups of students were asked to write report text about domestic animals, in ICoT, the students were asked to write a report text individually about wild animals.
In this study, the students were required to gather information about the animals in their house, and then they drafted their composition in their classroom with guidance of the teacher. Since the study aimed to investigate the effect of teacher feedback to student writing accuracy, the teacher’s feedback on the draft text was analyzed. After accepting the feedback, the student revised their composition. The revision was also analyzed. Then, about two months after the feedback was given, a writing test was held. The text from the test was also error-analyzed.
Error analysis, as Corder (1967 in Ellis, 1994 p. 48) argued, could be signified in three ways: firstly the errors provide the teacher with information about how much the learner had learnt; secondly, they provide researchers with evidence how language was learnt; and thirdly, they served as devices by which the learners discover the rules of target language. The reason why the error analysis is used in this study is to find the evidence about how the language has been learnt by students.
In analyzing errors in students’ text, a clause was the basis (Halliday, 1985; Eggins, 2004). The student text was synthesized in clauses, and then was analyzed for the errors. No marks were given to the text, but error frequency was tallied. Finally, the errors on first draft text, revision text and test version were compared.
 
Finding and Discussion
Since the errors were various, the direct, uncoded oral and written feedback was given only to certain errors as the target i.e. grammatical errors including subject-verb agreement, clause/phrasal construction, lexical errors and punctuation error. The grammatical errors on subject-verb agreement refers to the forms of verb (inflexion, tense) which, according to Huddleston and Pullum (2005, p. 88 see also Derewianka, 1998 p. 61), involve persons as well as number to all verbs in present tense – since a report text mostly uses this tense. Clause construction errors refer to the error in clause structure which is made up of a number of phrases or groups; typically consist of a nominal group (Subject) followed by verbal group (V) and another nominal group (O); the nominal group can consist of a number of words of which the main word is the last word of the group and is a noun; verbal group may consist of a single word where this word is a verb (Eggins, 2004 p. 128). Lexical or phrasal construction errors refer to the errors in the construction of nominal, verbal, adjectival, adverbial or prepositional phrases (Eggins, 2004, p 124). And, finally, punctuation refers to writing mechanics like capital letters, commas, full stops, sentence and paragraph boundaries etc. (Harmer, 2004, p 49 see also Haliday, 1985).

The result of error analysis of all student texts is displayed on tables and described more in following paragraphs including three student texts: draft, revision and test versions. The analysis consecutively discusses texts made by high achievers (student 1-3), medium ones (student 4-6) and low ones (student 7-9).
The following table shows the result of error analysis of texts written by high achievers.


The table shows that, in draft text, student 1 (high achiever) made 10 clauses with 71 words with a number of errors related to grammar, lexical and punctuation errors. Student 1 wrote more words in draft text (will be abbreviated ‘DT’ on the rest of the paper) but some of them were unnecessary and ungrammatical, like in ‘… Monkey food they is banana and bean…’, and ‘Kind skin colour monkey is colours brown’. Those words might be the first language effect. In term of errors, student 1 committed four errors of subject-verb agreement (written ‘SVAE’ in the rest of the paper) e.g. ‘Monkey are mamals. They is animals’; two clause construction errors (abbreviated ‘CCE’ in the rest of the paper) e.g. ‘Features front they like human being and they also be able go along …’; eight lexical errors (written ‘LE’ in the rest of the paper) like in ‘… mamal’, and ‘Kind skin colour monkey’, and two punctuation errors (shortened ‘PE’ in the rest of the paper) i.e. They have body hair, and…’ and ‘ they like. human.’
To the errors, the teacher gave direct feedback on the student composition by explaining, underlining and pointing errors, with arrows or circles, the words, phrases or clauses to correct.
In the RT, student wrote 10 clauses 58 words similar to the draft one with some corrections. She could correct most of the errors except one SVAE e.g. ‘Monkey are’ and one LE e.g. ‘Their face like human’. Some errors to which the teacher gave only oral feedback still existed. All errors to which the feedback was given in written form could be corrected.
In the test version, student 1 managed to write 10 clauses with 57 words with only one error. The text is similar to the revised with one CCE still existed e.g. ‘*They body fur…’ (their)
For student 2 of high achievers, the table shows that she made 11 clauses with 68 words and some errors in DT. Compared with RT, Student 1 wrote less words in since there are some prepositions left like ‘ …diet usually (on) grass’ and ‘in order (that) the rhinocerous ...’ In term of errors, student 2 made two SVAES e.g. ‘Rhinocerous sondaicus belong to …’and ‘… he taste …’; one clause construction error e.g. ‘we should hidden …’; four LEs like in ‘… they can kiss (detect) people or enemy …’ and ‘… can be mezbah hard …’; and five PEs which are mostly wrong in writing capital letter e.g.’ … mammals. they live on forest. they often avoid ….’
On the errors, the teacher similarly gave direct feedback on the student composition by explaining, underlining and pointing errors, with arrows or circles, the words, phrases or clauses to correct.
In the RT, student 2 composed 12 clauses with 58 words in RT and only two errors made. She could correct almost the errors except two SVAES and one PE. The former happened because student 2 wrote incorrectly the feedback from the teachers who put inflection ‘-s’ after the word ‘belong’ and ‘diet’ rather than she wrote the words with apostrophe e.g. ‘belong’s to mammals’. The
PE was still related to capital letter e.g. ‘the rhinocerous Sondaicus not go’. This might because of the student’s carelessness since in the DT she managed to write the capital letter to the word.
In TT, student 2 could compose similar text to the revised one with consisting of 12 clauses with 57 words but still conducted some errors. Two SVAES or inflection errors as in ‘belong’s’ and ‘diet’s’ and one PE remained, which yet was related to capitalization but in different word, namely, ‘… far distance. and if we will see it …’
For student 3, it can be seen on the table that in DT, she made 12 clauses with 72 words with some errors. She wrote more words than in RT and TT, because, like student 1, she wrote some words which were unnecessary as the word could be understood implicitly like in ‘They have feature, they can …’ or exaggerate the use of determiner like in the ‘… have mustache the luxuriant and long…’ and ‘… have toot the long’. Such use of determiner indicates that student 3 did not master the use of determiner or mistranslate it from her first language. In term of errors, student 3 made one SVAE error e.g. ‘Rabbit are mamal’ and four CCEs e.g. ‘they habitat forest. They skin white and black’. There were also six LEs like the word ‘ … mamal.’ and ‘… carot and kangkung …’; and five PEs consisting of wrong capitalization e.g.’ … kangkung). they habitat…’ and ‘… and long. then they…’ or exaggerating full stop e.g. ‘… with fast. possible. with body flaty.’
On the errors, the teacher similarly gave direct feedback to the one five given to previous students by explaining, underlining or pointing errors - with an arrow, a cross or a circle - the words, phrases or clauses to correct.
In the RT, student composed 12 clauses with 63 words and only few errors. She could correct all the errors except two LEs e.g. ‘..mamal’ (mammal) and ‘flaty’ (flat). This happened possibly because the teacher probably did not give feedback on those words or student 3 forgot the oral feedback from the teacher.
Meanwhile, in the TT, student 3 composed 10 clauses with 55 words and only two errors existed. Her TT is similar to the RT with one clause missed and one subject agreement error as in RT and one clause construction error e.g. ‘They a long toot.’ where she missed writing the word ‘have’ which probably because she was careless or in a hurry. Two LEs also remained as in the RT.
In short, the students of high achievers could maintain the length of the text (number of clauses and words are relatively similar) and lessen errors in the TT. They made good progress in their writing accuracy which possibly because of the teacher feedback. It is seen from the errors in DT which decrease in revised and TTs, especially related to clause construction, lexical and PEs. This indicates that the feedback seems to have good effect on their writing accuracy.

The following table shows the summary of the result of error analysis on text made by mid achievers which is then explained in following paragraphs.
 
In the DT, Student 4 of mid achievers made 10 clauses with 64 words and a lot of errors. He wrote more words in DT but some of them were ungrammatical and unnecessary e.g. in ‘…Fruit *toheir. Food is papaya …’, and ‘Parrot also to *proposes (possess)’ where the word ‘*toheir’ and ‘to’ are not needed since they make the clause ungrammatical and incomprehensible. In term of errors, student 4 committed three SVAEs e.g. ‘Parrot *are bird. They *is animals’; eight CCEs e.g. ‘Parrot is can imitation voice man’ and ‘Colour parrot various there this one colour Red, …’ ; four LEs like in the words ‘…toheir’ , ‘bedides’, ’body he beautiful and pretty’, and three PEs related to capitalization e.g. ‘this one colour Red, Blue and Black.’
On the errors, the teacher gave similar direct uncoded oral and written feedback to correct them.
In the RT, student 4 made 7 clauses with 43 words and some errors. She could correct most of the errors except one SVAE e.g. ‘parrot live in…’ and one CCEs e.g.‘ parrots can imitation man voice’; two LEs e.g. ‘to heir food’ and ‘… also can behav’ and also four PEs e.g. ‘Parrots are bird, they eat. Fruit …’. Some errors to which the teacher gave only oral feedback keep existed. Most errors to which the feedback was given in written form could be corrected, but other new mistakes appear.
In the TT, student 4 managed to write 7 clauses with 37 words and some errors. The text is like in revised one with one clause missed. He also made three CCEs, e.g. ‘Eat parrots sesame and food.’ and ‘Parrots habit stable,’ ; three LEs like the words ‘ … include…( included)’,’ hou (have)’ and ‘prety’.
11
He also committed two PEs in ‘… habit stable, colour parrots. Sort …’, where the coma should be full stop, the full stop should be omitted.
In the meantime, student 5 succeeded in making 8 clauses in her text with 44 words in draft but many errors are present. She made three SVAES as in ‘Horse are mammal. Most horse live in the jungle.’; four CCEs e.g. ‘they consume is grass …’ , ‘their skin colors white and black.’; and six PEs as in ‘… in the jungle. they consume is grassand vegetables. they have …’
To those errors, the teacher gave similar feedback to the one given to previous students.
In revised version, student 5 made 8 clauses with 44 words and one error which all are almost the same as in draft version. She could correct almost all errors except one clause construction error e.g. ‘They consume is grass .…’; two LEs as in ‘Horse’s are mammals. Most horse’s live….’where the inflection ‘s probable mean the inflection of plural: ‘horses’.
In the TT, student 5 wrote 8 clauses with 41 words whose clauses and words are similar to those in the DT. She made one subject agreement error as in ‘Horse are mammal’; two CCEs e.g. ‘Their skin color black and white’ and ‘they able to ride on’.
Student 6 could make 10 clauses with 67 words with some errors in draft version. She made one CCEs e.g. ‘chameleon form of small insecests such as …’; LEs e.g. ‘…insecests,…’ …dragonflies…’, etc., and three PEs as in ‘…slowly. food chameleon …’, ‘and cicadas. the special’ etc.
On the errors, the teacher gave similar feedback to those done to previous students.
In RT, student 6 made 11 clauses with 66 words. He could correct some errors to which the teacher gave feedback. But, some new LEs still come up. For example, the words, ‘mofies’ and ‘invorment ’ were written correctly in the DT, but they were incorrect in the RT. This might because of the student’s carelessness. In punctuation, she still made errors as in ’… animal. that mofies …’and ‘… dragonflies, and cicados.’
Meanwhile in the TT, she wrote 7 clauses with 43 words in her text. But, the number of errors increased compared to the RT. She made five CCEs as in ‘Chamelon to stay is tree’ and ‘ Chamelon a mimicry…'; ten LE as ‘Chamelon’, Thet, tangue etc..
In short, from the error analysis to the texts made by mid achiever students, it can be seen that the students could correct most errors in RT but most of the errors turned up in TT. In the TT, they mostly missed one clause with nine words on the average. Unfortunately, most errors came up again in the TT, and even those related to LEs are worse. Only those PEs decrease mostly. They only
made half errors on the average. This indicates that although in very little amount, there is an effect of teacher feedback to writing accuracy for those who are categorized as mid achiever students. And, to see the effect to the low achiever, look at the table below.

Further, in the DT, student 7 who was low achiever made six clauses with 41 words and some errors. He committed one SVAE as e.g. ‘Spider are insects’; three CCEs as in ‘Live they in forest’ and ‘Have colors: yellow, …’; four LEs e.g. ‘fies’,’ blak’,’ ussuly’ and ‘they net to make….’. He also made two PEs as in ‘… height 1-5 cm. and ussuly they net to make and …’.
To the errors, the teacher gave similar direct feedback orally and written to correct the errors.
On revised version, student 7 rewrote 6 clauses with 44 words but the errors still existed. He only managed to correct some errors but he made other new errors. He made one subject- verb agreement as in ‘Spider are insects’ and two CCEs e.g. ‘…and ussuly they to make net’. He also committed five lexical construction because many unnecessary words or strange words coming up in the text like ‘They eat they insect is and fics, blak, heght etc.’ and two PEs e.g. ‘… house. have colors’ and ‘… leight. and ussuly they …’
On the TT, student 7 made a text with six clauses consisting 44 words but he also made errors like in the draft version. He kept the subject agreement error e.g. ‘Spider are insects’ ; two CCEs like ‘… live
13
they in forest’ and ‘ … house. have colors:…’ and seven LEs as ‘inselis, fils, they, blak etc.’ In punctuation, he made three errors as in ‘and fils they. live they in forest and in house. have ‘.
Meanwhile, student 8 composed eight clauses with 43 words in DT and some errors. He made one SVAE e.g. ‘Whale belong to big mamal.’ and two CCEs as in ‘they body big and long.’ and ‘ they skin/ color blank and white’. He also committed four LEs in the word: ‘mamal.’ ,’flankton’ and ‘in habit’; and five PEs which mostly capitalization as in ‘… big mamal. they diet small fish and flankton. they in habit in the sea. they…’
On the errors, the teacher gave similar direct uncoded oral and written feedback by explaining, underlining and pointing out them.
And, in the RT, student 8 succeeded in making a report text with 7 clauses consisting of 43 words and some errors. He made one SVAE in ‘Whale belong to…’ and one CCEs ‘they bodu big and long’. He also kept the LEs as the words: ‘mamal.,’ ’flankton’ , ‘in habit’and ’they (their)habit …’’ and two PEs in the ’They in habit in the sea. they habit is result water beach. they bodu big …’
However, in the TT, student 8 made the text with fewer clauses and more errors than in the RT. He composed 7 clauses with 36 words and many errors. He kept one subject agreement error as draft and RT in ‘Whale belong…’ and three CCEs as in ‘they food small fish and plankton. they color/skin black and white. they habitat to the sea’. He also made five LEs in the words: ‘mamal, they (their), bike ‘ and nine PEs mostly capitalization like in DT e.g. ‘mamal. they food small fish and plankton. they color/skin black and white. they habitat to the sea. they Havebody long and Big. They Have a measurement long and bike. they ….’
Finally, student 10 (low achiever) composed ten short clauses with 47 words and some errors. She made no subject verb agreement but almost all (eight of ten) clauses are ungrammatical as they have no verbs like in ‘Tiger usually . animal. Food meat. Habitat tiger forest. Or I . bushes.etc.’. She also made three LEs e.g. ‘tiger, carnovora, balck’ but made eight PEs related to overusing full stops as in ‘Tiger usually . animal.’, ‘Habitat tiger forest. Or I . bushes.etc.’
As done to previous student, the teacher gave direct uncoded oral and written feedback.
In the revised, she wrote eight clauses with 39 words with few errors existed. She made no mistakes in subject verb agreement but still made three CCEs in ‘Tigers usually mammals.’ (no verb), ‘Habit they Roam.’ and ‘Their color black and yellow.’ (no verb) She composed only one LEs, writing the word ‘and’ in the end of the text but still wrote three PEs e.g. ‘Habit they Roam.’, ‘own region Their color black and yellow.‘ and ‘They have 70 cm height and …’.
However, in a TT, student 9 made seven clauses with 38 words and more errors than in the revised version. She made two SVAE e.g. ‘Tiger see body big and long’ and , ‘tiger eat small animals.’ She also committed five CCEs like in ’ Tigers usually mammals’.(no verb), ‘They in forest or in bushes,…’ (no verb) and three LEs in the words: ‘blak’ and ‘see’ (‘they see color’ and ‘they see body’). She also made six PEs as in ‘… or in bushes, tiger eat small animal’ and ‘they is roam, they see’.
In short, from low achiever students, it can be seen that in the RTs, most students could correct some errors of DT but the errors came up again in the TT. And even, the lexical and PEs increased. It indicates that they could partly respond to the teacher’s feedback but could not maintain the correction when they were tested. Their TTs seem not more accurate than their draft ones - even worse. It suggests that the low achiever students do not acquire the feedback from the teacher. The teacher feedback does not affect their writing accuracy. This is in line with what Truscott (1996, 2007) put forward and what Pan (2010) found that the students made progress in the revised version but the success was not repeated in later TT. Its also suggests the teacher error feedback alone may not facilitate the learning of linguistic information (Pan, 2010).
Overall, the analysis indicates that the teacher’s direct, uncoded oral and written feedback had significant effect only on text composed by high achiever students. But, it had little effect on the text written by mid achievers and almost no effect on the text made by low achievers. This result supports Tsao (2010)‘s finding that error correction cannot improve the accuracy of student writing except for high proficiency students; and high proficiency students responded more favorably to the feedback than low to intermediate ones.
Additionally, to summarize, the table below displays average of error decrease in all students’ texts.
Note: RT= revision text, DT=draft text, TT=Test text

The table shows that on the average 80% of SVAE, 67% of CCEs, 81% LEs and 49% of PEs in DTs remain in TT. The highest errors remained were related to LEs (81%), which is in line with what Usaha (1998) found, that the highest error rate was found in wrong word followed other errors.
Based on the data above, we can see that the feedback from the teacher has little impact on the writing accuracy. This probably because the feedback the teacher gave was not so clear for students. This supports what Fregeau (1999 in William 2003; Cohen, & Cavalcanti (1990 in William 2003) put forward that the feedback on surface errors does not work when it is inconsistent, unclear and overemphasizes the negative feedback. Or, the direct, uncoded, oral and written feedback does not have significant effect on the students’ writing accuracy as argued by Truscott (1996, 2007 in Harshorn 2008) reporting that error correction is not helpful for improving the grammatical accuracy of L2 writing.
Conclusion and Suggestions
This paper has sought the effect of direct and uncoded written feedback with conference to writing accuracy of junior high school students. Based on the document analysis of texts made by nine students of high, medium and low achievers, (three of each), it was revealed that all students performed better writing accuracy in revision text - fewer errors are made – but the errors in DT remained existent in their test versions. It also indicates such feedback from the teacher has little impact on the writing accuracy. The finding also suggests that the feedback gave significant effect only to the high achievers. Only few errors were made again in their test version. But, such feedback did not give much effect to the texts made by mid achievers and low achievers. Most errors on their DT remained existent in their TTs. This is in line with the findings of previous experts that error correction is not helpful for improving the grammatical accuracy of L2 writing (Fregeau, 1999; Cohen, & Cavalcanti, 1990 in William 2003; Truscott, 1996, 2007 in Harshorn 2008, Pan, 2010; Tsao, 2010).
However, due to some limitations of the study, some suggestions are necessary to put forward. Firstly, the type of the teacher feedback investigated is direct, uncoded oral and written feedback focusing on form, other types of feedback like coded, indirect or with conference need conducting in future studies which might be related to writing accuracy, fluency or even content. Secondly, since only nine students involved, in the next study, it
16
would be better to investigate a whole class or even two groups of students to get convincing conclusion. Comparing many students or two groups of them would be more objective and reliable result. Thirdly, as the test stage of this study let students choose the topic on their own, so that many of them wrote the same text as they did in revised text and, hence, the test effect may have happened in this situation - possibly the students still remember what they had written before. This, of course, may decrease the conformity of the feedback result. Therefore, in the following research, prompting students to write the same genre of text but the different topic is really recommended. Finally, as no other rater or analyst but researcher analyzed the errors, the result of the analysis might have not very objective, so, the next study are to involve more raters/analysts to get more objective and fair assessment of students’ composition.

BibliographyAlwasilah, A. C. 2001. Language, Culture and Education: A Portrait of Contemporary Indonesia. Bandung: Andira.
Bitchener, J., Young, S., Cameron, D. (2005), The Effect of Different Types of Corrective Feedback on ESL Student Writing. Journal of Second Language Writing 14 (2005) 191–205.Auckland: Elsevier Inc. available on line at pdfszone.com/.../effects-of-different-error-feedback-approaches-in-st... or at http://printfu.org/corrective+feedback retrieved on November 20, 2011.
Brown, H. Douglas. 2001.Teaching by Principles, An Interactive Approach to Language Pedagogy. New York : Addison Wesley Longman.
Chuang, W. (????).The Effects of Four Different Types of Corrective Feedback on EFL Students’ Writing in Taiwan available online at www.dyu.edu.tw/~cd9000/html/publication/files/JoGE04/07.pdf retrieved on Dec 20, 2011.
Edwards, Jason. 2008. Making written feedback effective. Online article. Available at http://www.teachingexpertise.com/articles/making-written-feedback-effective-3012). Accessed on November 24th, 2011.
Emilia, E. 2010. Teaching Writing: Developing Critical Learners. Bandung: Rizqi Press.
Eggins, S. 2004. An Introduction to Systemic Functional Linguistics; Second Edition. London: Continuum International Publishing Company.
Ellis, R. 1994. The Study of Second Language Acquisition. Oxford: Oxford University Press.
Grami, G.M. A. 2005. The Effect Of Teachers.. Written Feedback On Esl Students.. Perception: A Study In A Saudi Esl University-Level Context. E-Journal: Annual.. Review Of Education, Communication And Language Sciences, Volume 2. Available At Http://Research.Ncl.Ac.Uk/Arecls/Vol2_Documents/Grami/Grami.Htm
González, E.F. 2010. Impact of Teacher/Student Conferencing and Teacher Written Feedback on EFL Revision 1. Online Journal: Volume 34, Number 1. Available at mextesol.net/journal/.../42e05a177ec7f41777abe59cbcf97371.pdf retrieved on December 20, 2011.
Halliday, M.A.K. 1985. An Introduction to Functional English Grammar: Second Edition. London: Oxford University Press.
Halliday, M.A.K.1983. Spoken and Written Language. Oxford: Oxford University Press.
Hartshorn, K.J. 2008. The Effects Of Manageable Corrective Feedback On Esl Writing Accuracy. Online published dissertation. Department of Instructional Psychology and Technology. Brigham Young University.
Harmer, J. 2004. How to Teach Writing. Essex: Pearson Education Limited.
Huddleston, R. and Pullum, G.K. 2005. A Student’s Introduction to English Grammar. Cambridge, U.K: Cambridge University Press.
Hyland, F. 1998. The Impact Of Teacher Written Feedback On Individual Writers. Journal of Second Language Writing. Volume: 7, Issue: 3, Pages: 255-286. Available online at http://www.mendeley.com retrieved on December 22th, 2011.
http://www.thefreedictionary.com
Iseni, A. 2011. Assessment, Testing And Correcting Students' Errors And Mistakes. Language Testing in Asia Volume one, Issue three October 2011. (Page 96).
Stitt-Bergh, M. 2007. Giving Effective Feedback on Student Writing 2007. Online article. Available on http://www.mwp.hawaii.edu/resources/workshop-index.htm
Tuttle H. G. . 2007. Written Feedback: Improving Student Learning. Online article. Available on http://eduwithtechn.wordpress.com/2007/10/26/written-feedback-improving-student-learning/
Usaha, S. 1998. The Impact of Different Types of Teacher Written Feedback on EFL University Students’ Writing. KKU Res J (GS) 8 (2) : April – June. Available online at
Kaprista, S. 2008. The teachers’ Written Feedback in Students’ Composition. Online published paper; Sanata Darma University. Available online at http://els.teylcenter.net. Retrieved on November 25th, 2011.
Williams, J. G. Providing Feedback on ESL Students' Written Assignments. The Internet TESL Journal, Vol. IX, No. 10, October 2003 available online at http://iteslj.org/Techniques/Williams-Feedback. html retrieved on November 2011.
Tsao, Chia-Hsiu. The Effect Of Error Correction On EFL Student Writing. Language Education Center, Fooyin University. Online published article. Available at nutnr.lib.nutn.edu.tw/bitstream/987654321/.../..-...-....p... retrieved on December 23rd, 2011.