Correcting Errors: The relative efficacy of different forms of error feedback in Second Language writing ENGLISH REVIEW: Journal of English Education ISSN 2301-7554 Vol. 1, Issue 2, June 2013 http://journal.uniku.ac.id/index.php/ERJEE CORRECTING ERRORS: THE RELATIVE EFFICACY OF DIFFERENT FORMS OF ERROR FEEDBACK IN SECOND LANGUAGE WRITING Chitra Jayathilake University of Sri Jayewardenepura, Nugegoda, Colombo, Sri Lanka E-mail: chitj1995@yahoo.com APA Citation: Jayathilake, C. (2013). Correcting errors: The relative efficacy of different forms of error feedback in second language writing. English Review: Journal of English Education, 1(2),139-150 Received: 30-01-2013 Accepted: 3-02-2013 Published: 01-06-2013 Abstract: Error correction in ESL (English as a Second Language) classes has been a focal phenomenon in SLA (Second Language Acquisition) research due to some controversial research results and diverse feedback practices. This paper presents a study which explored the relative efficacy of three forms of error correction employed in ESL writing classes: focusing on the acquisition of one grammar element both for immediate and delayed language contexts, and collecting data from university undergraduates, this study employed an experimental research design with a pretest-treatment-posttests structure. The research revealed that the degree of success in acquiring L2 (Second Language) grammar through error correction differs according to the form of the correction and to learning contexts. While the findings are discussed in relation to the previous literature, this paper concludes creating a cline of error correction forms to be promoted in Sri Lankan L2 writing contexts, particularly in ESL contexts in Universities. Keywords: error correction, written corrective feedback (WCF), English as a second language (ESL), metalinguistic information, second language acquisition (SLA) INTRODUCTION This research investigates the role of Written Corrective Feedback (WCF) on the acquisition of one feature (present continuous form) of English Language by adult intermediate ESL learners intending to provide some insight into the aspects of correction of learner errors. This paper begins reviewing briefly how WCF has been addressed in Second Language Acquisition (SLA): It will examine some implications presented in the previous literature while forming the research questions for the current study. Next, the paper reveals the results of the current experimental study, concluding with a few implications for error correction in writing pedagogy for adult ESL learners. Error correction is one of the significant and frequent phenomena in second language writing as it is a crucial aspect experienced by both teachers and researchers in classrooms. Initially, there has been controversy (with Truscott (1999) and Ferris (1999) as two key figures) as to whether written corrective feedback facilitates any accuracy in L2 learners. Truscott (1999) arguing that all forms of error correction in students’ L2 writing is ineffective, held a strong view against error correction. He expressed that error correction of L2 student writing is harmful, therefore should be mailto:chitj1995@yahoo.com CHITRA JAYATHILAKE Correcting Errors: The Relative Efficacy Of Different Forms Of Error Feedback In Second Language Writing abandoned. He continued saying that irrespective of L2 students’ clear desire for grammar correction, teachers should not correct learner errors in their writing. However, in Ferris’ (1999) equally strong rebuttal, she counter argued that Truscott had failed to notice some positive research evidence on the effects of grammar correction. Even with the other existing research data (e.g. Chandler, 2003; Sheen, 2007 and Lee, 2008 ) it is still too premature to have a conclusive answer to the questions - whether error correction is effective in improving the accuracy of L2 writing in the long term for learners of all levels, and what types of WCF are beneficial to learners. Two controversial views over the issue of WCF, initially forwarded by Truscott (1999) and Ferris (1999) continued for over a decade counter- arguing whether corrective feedback given to L2 writers helps to improve their written accuracy. As mentioned, Truscott (1999), strongly rejected error correction as useless and harmful to the accuracy in students’ writing. However, with the follow up research findings by other researchers (e.g. Chandler, 2003; Sheen, 2007 and Lee, 2008), the strong effect of Truscott’s claim was minimized: many of the follow-up studies on error correction have shown that students receiving error feedback for their written texts improve in accuracy over time. Ferris (1997) examined over 1600 marginal and end comments written on 110 first drafts of papers by 47 advanced university ESL learners to investigate what characteristics of teacher commentary appear to influence student revision and whether revisions influenced by teacher feedback lead to effective changes in learners’ writing. The results indicated two contrasting views: students sometimes pay much attention to teacher commentary to revise their writing, at other times students ignore the teacher commentary given. However, the revision students made was influenced by teacher feedback and facilitated their learning to a great extent. The results of Ferris’ research seem to suggest the importance of revision made based on WCF. Ferris (2004) has again presented that indirect error correction is more beneficial than direct correction as it pushes learners to engage in hypothesis testing: Ferris’ justification is that when learners are engaged in testing learners’ assumption on language elements, it helps to internalize the language components. Chandler (2003), while supporting Ferris’ overall view that WCF facilitates in SLA, challenged some detailed findings of Ferris (2004). That is, Chandler (2003) revealed that direct correction was superior to other types of indirect correction in producing more accurate writing. Chandler justifies her argument saying that although indirect correction draws much cognitive attention from learners it delays confirmation of students’ hypothesis testing acts. Accordingly, although both Ferris and Chandler agree that WCF helps learners in writing in SLA, they hold contradictory views on the relative efficacy of different types of WCF. Research studies further show the distinction between direct and indirect feedback. Ferris & Roberts (2001) compared these two types and revealed that the students receiving feedback of both underlining and coding of their erroneous forms did slightly better in revising their grammatical errors than the ones receiving only underlining as the feedback (underlining indicates only ENGLISH REVIEW: Journal of English Education ISSN 2301-7554 Vol. 1, Issue 2, June 2013 http://journal.uniku.ac.id/index.php/ERJEE the location of the error whereas coding provides clues to the type of error such as spelling, vocabulary and so on). Both groups were significantly more successful in revising errors than the control group receiving no feedback. In the meantime, Erel & Bulut’s (2007) study investigated the possible effects of direct (the correct form is written on students text), and indirect coded error feedback( a symbol representing a specific kind of error) in a Turkish university context to examine the accuracy in writing. The final results showed that the indirect coded feedback group committed fewer errors than the direct feedback group. In the meantime, Ferris & Roberts’ (2001) study included learners who received no error correction at all in their research. Three kinds of feedback conditions were operationalized in their study - (i) errors marked with codes (ii) errors underlined but not marked or labeled (iii) no feedback at all. The findings revealed that although, there were no significant differences between the groups’ ability to edit their papers, the students who were given corrective feedback outperformed the group with no feedback in editing their papers. Sheen (2007) examines the differential effect of two types of WCF and investigates the relationship between language analytic ability and the effects of WCF on the acquisition articles (one linguistic feature) in English Language, employing a pretest- treatment–posttest design with 91 learners of various L1 (first Language) backgrounds. The three groups formed in Sheen’s research are a direct-only correction group, a direct metalinguistic correction group and a control group. The results indicated that the WCF has a positive effect on the learning of articles of English language: the results also indicated that direct correction with metalinguistic feedback is more effective than WCF without metalinguistic feedback: learners with a high level of language analytic ability benefited more irrespective of the type of CF. Meanwhile, Lee (2008) contributed to the issue of WCF investigating teachers’ WCF practices in Hong Kong secondary classrooms. This study examined the WCF provided by 26 Hong Kong secondary English teachers to 174 student texts, followed up by interviews with 06 teachers. The results indicated that teacher feedback is primarily error-focused and exam- cultured. Lee suggests that feedback practices are influenced by contextual factors such as teachers’ beliefs, values and socio-political issues pertaining to power and teacher autonomy. Thus it leads to another significant concern for future research as practices may relate to the efficacy of WCF. Liu (2008), engaging in a quasi- experimental classroom study investigated 12(two groups, 6 each) university ESL students’ abilities to self- revise their writing across two feedback conditions. Two types of error correction Liu employed in the study were (i) direct correction (correct form provided by the teacher), (ii) indirect correction (only an indication that an error exists). Liu classified the instances of errors as (i) morphological errors, (ii) semantic errors, and (iii) syntactic errors. Results show that both - direct and indirect correction - helped students self-edit their texts. Moreover, indirect feedback has been more beneficial in reducing morphological errors than semantic errors. Similar to the findings of Lie (2008), Chandler (2003) also revealed that teachers’ feedback on students’ CHITRA JAYATHILAKE Correcting Errors: The Relative Efficacy Of Different Forms Of Error Feedback In Second Language Writing grammatical and lexical errors helped to improve in both accuracy and fluency of students’ writing. These findings further disprove Truscott’s (1999) claim on the negative effect of error correction. Given the conflicting results on the effects of different WCF, it is difficult and premature to formulate any conclusion without investigating the phenomenon of the efficacy of different categories of WCF further, in different ESL contexts. Therefore, the present study follows this line of research by examining three types of WCF:  Direction only (D Group)  Direction and Correction (DC Group)  Direction, Correction and Metalinguistic Information (DCM Group) In second language writing, focusing only on one grammatical feature, namely present continuous form of English. Hence, the following questions are addressed in this preliminary study: Does WCF have an effect on adult learners’ acquisition of L2, present continuous form in particular? Do different WCF types (Direction only, Direction and Correction, and Direction, Correction and Metalinguistic Information) have different effects on the acquisition of present continuous form in English Language, by adult ESL learners? Which type of WCF facilitates more in delayed L2 context (for long- term acquisition), on ESL learners’ acquisition of present continuous form? METHOD This study employed an experimental research design with a pretest-treatment-posttests structure, using ESL Classroom. One week prior to the start of the WCF treatment, the participating students completed the pretest. The immediate unexpected posttest was completed following the three WCF sessions and the delayed unexpected posttest 5 weeks later. The pretest-posttest design was selected as it is reliable in measuring the language acquisition, quantitatively (Brown 1988). The research sequence, a modified design of Sheen (2007), thus, was carried out over a period of approximately 9 weeks.  Pretest (1st week)  Three treatment sessions(2nd week)  Posttests Posttest 1 (3rd week) Posttest 2 (9th week) The participants of this study were 48 third–year, female undergraduates, aged 23-24 of the Faculty of Arts, University of Sri Jayewardenepura, Colombo, Sri Lanka. They had scored between 20-25 marks for a language test conducted by the university, prior to the research and were from similar linguistic background: their mother tongue was Sinhala (Sinhala is a national language of Sri Lanka and it is the First Language of the majority of Sri Lankans): Tamil is the other national language; they had started learning ESL at the age of 08 years, at state-run schools situated in rural areas of Sri Lanka and learned ESL at school at least for 10 years ; they had obtained Simple Passes (A student can obtain a Simple Pass, if s/he scores approximately between 30- 40 marks (out of 100) for the ESL test paper: students sit for this General Certificate of Education (Ordinary Level) examination, after learning ESL for 08 years at school: they are between the ages of 15- 16 years at this time. The aim of the test is to evaluate students’ knowledge in grammar, vocabulary, reading and writing skill) for English Language at the General Certificate of Education (Ordinary Level) Examination administered by the Ministry of ENGLISH REVIEW: Journal of English Education ISSN 2301-7554 Vol. 1, Issue 2, June 2013 http://journal.uniku.ac.id/index.php/ERJEE Education of the country. At the university, they followed the subjects offered for the degree in the medium of Sinhala. Given the amount of understanding and commitment on the research area, it was decided that the researcher would participate in the study as the teacher (she has 20 years of experience and postgraduate qualifications in teaching ESL) and the corrector of the errors. Four groups, were formed from a class of 48 students: the Direction Only group (n =12), the Direction and Correction group (n 12), the Direction, Correction and Metalinguistic information group (n=12) and the Control group/ C Group (n 12). The Control group received no information of any erroneous forms of the present continuous forms of their language. The Direction only group received only an indication of the location of an error on the student’s text: that is, the location where the error occurred was underlined or circled by the teacher. Direction and Correction group was operated as indicating the location of an error and providing the correct form near the erroneous form on the students’ text. Direction, Correction and Metalinguistic information group was operationalized as indicating the location of the error, providing the correct form and explaining the error with some metalinguistic information. That is, linguistic comments were provided which explained the correct form. However, to cover the focus of the study, the researcher corrected a few other errors than those involving present continuous form in the three experiment groups while correcting non-present continuous forms on the texts of the control group. The linguistic structure targeted in the present study was the present continuous form in English. The decision was made after exploring students’ common errors indicated in the proficiency test held prior to the current study. It revealed that many learners frequently, but erroneously, use this form in their written L2 usage: it suggested that they had only partial knowledge of the target form indicating a certain developmental stage of the target form. Thus, focus was given to the correction of present continuous form in this study. The instrument was the three tests administered:  Pretest  Posttest 1  Post test 2 (delayed Posttest) The three experimental groups completed the treatment and tests while the control group completed only the tests. The purpose of the Pretest was to measure the students’ knowledge of the grammar element – present continuous form - prior to the treatment. The Posttests were aimed to determine whether the students had acquired the language elements addressed through treatment given during the period of study. Posttest 2, in particular, was held to ascertain the reliability of the acquisition of language elements presented through treatment. The activities of the test items were randomly changed from one test to another to minimize the test effects. First, in the first week of the study, the pretest was administered to all the participants and the marks obtained at the Pretest were examined by the teacher cum researcher to gain an idea of the students’ knowledge of present continuous form in English. CHITRA JAYATHILAKE Correcting Errors: The Relative Efficacy Of Different Forms Of Error Feedback In Second Language Writing In the second week, the teacher, accordingly, conducted 3 (1 hour each) treatment sessions, (a total of 03 hours) meeting students every other day. Although materials used for treatment were teacher–directed, students’ level of second language proficiency was taken into consideration when selecting these materials. Two unexpected posttests were held, in the 3rd and 9th week after the commencement of treatment sessions, and their results were recorded. As previously mentioned, three experimental groups completed both the treatment sessions and tests. The control group completed the tests only: they did not receive any WCF but instead followed the normal class room activities: some non-present continuous (erroneous) forms were marked, in the texts of control group, to mask the purpose of the research. Three treatment sessions were held with a day break in between the sessions. Each session involved one picture description activity and one clip (clips of a movie) description activity which elicited the present continuous form from the students. The students were involved in writing two compositions based on the two activities, after discussing the images orally in their groups. Each picture/movie elicited about 10/12 statements of present continuous forms. The steps followed during the treatment session were:  first, the researcher displayed the movie-clips to students  she asked them to describe the actions displayed in groups, while she facilitated them ( they were given only about 10 minutes).  she asked each student to write a description of the images using about 40 minutes (she provided two examples).  then the researcher collected the written texts.  second, she gave each student a copy of a picture and followed the same steps as in the previous activity.  she returned students’ compositions to the students the next day, with researchers’ WCF on the students’ text.  students were asked to go through the WCF in class and to examine the corrections and the feedback given carefully. They were given about 10 minutes for that.  no other comments were made on the WCF by the researcher and no student was allowed to revise their writing which allowed for the effect of the WCF treatment by itself to be investigated. The researcher corrected the errors (usually about 10/12 statements) on one text. However, as mentioned previously, to mask the focus of the study, the researcher corrected a few other errors than those involving present continuous form in the three experiment groups while correcting non-present continuous forms on the texts of the control group. Two tests were employed to investigate the acquisition of the target structure in this study: picture description and a few clips of a movie The same test was used for the pretest, the immediate posttest and the delayed posttest but the items in the pictures were randomly changed to minimize the test effects. Picture description test (30 minutes) The picture used for description consisted of about 10-15 stimuli, each of which required at least one sentence ENGLISH REVIEW: Journal of English Education ISSN 2301-7554 Vol. 1, Issue 2, June 2013 http://journal.uniku.ac.id/index.php/ERJEE involving the use of present continuous form. Before administering the test, the researcher made two sample sentences from a similar picture displayed on the board, so that the learners could familiarize themselves with the procedure and the grammar form expected. In administering the test, each student was provided with a copy of the picture and asked them to make 10 sentences similar to the examples provided. Clips of the movie (30 minutes) These video clips were extracted from a movie which consisted of 8 sequential images. The students were asked to describe what was happening in the movie- the actions in the clips. Before administering the test, the researcher made certain that the students were familiar with the new vocabulary. They were asked to make 10 sentences, following the examples provided. Each correct use of the present continuous form was given a similar mark for each student and the final marks for each student were calculated by 100. RESULTS AND DISCUSSION All scores were computed to analyze the data. Table 1 presents the summary of the descriptive statistics for total scores for the tests taken together over the three testing periods- Pretest, Posttest 1 and Posttest2. Table 1: Group Means, Variance, Standard Deviation and Coefficient of Variance for total Test Scores Group Pretest Posttest 1 Posttest 2 M V SD C of V M V SD C of V M V SD C of V C 31.5 24.45 4.95 15.7 31.83 24.70 4.97 15.61 31.33 20.42 4.52 14.42 D 31.33 12.97 3.6 11.9 42.58 18.08 4.25 9.99 43.17 20.33 4.5110.45 DC 31.42 24.27 4.93 15.68 54.75 14.57 3.82 6.97 35.33 26.61 5.16 14.6 DCM 32.33 30.24 5.5 17.01 57.58 22.63 4.76 8.26 58.17 21.61 4.65 7.99 C = Control Group D = Direction only Group CD = Correction and Direction Group CDM = Correction, direction and Metalinguistic Information Group M = Mean V = Variance SD = Standard Deviation C of V = Coefficient of Variance Although the mean scores of the control group remains almost the same for three tests, the mean scores of the treatment groups show a relative increase over time. In other words, the three treatment groups’ gains over time were substantial but the control group showed no improvement. This principally suggests, supporting Ferris’ (2004) hypothesis, that the error correction treatment has been beneficial to learners. At Posttest 1, the mean scores of all three treatment groups are considerably higher than that of the control group. However, DC group shows a CHITRA JAYATHILAKE Correcting Errors: The Relative Efficacy Of Different Forms Of Error Feedback In Second Language Writing considerable decrease of mean scores from Posttest 1 to posttest 2. Out of the three treatment groups, the most consistent mean scores can be noticed in DCM group and in D only groups. Thus, as far as the mean scores are considered, it is revealed that the WCF is facilitative and DCM feedback seems the best in facilitating SLA as it outperformed the other two treatment groups both in Posttest 1 and Posttest 2. That is, in test scores, Posttest 1 results favoured all three treatment groups, but Posttest 2 results favoured the D group and DCM group. This initial assumption can be verified with the statistics of variance and coefficient of variance. Table 1 also displays that even the coefficient of variance of DCM group is relatively low, accounting for 8.26 and 7.99 for Posttest 1 and 2 respectively. This comparatively low score of coefficient of variance verifies the results of the mean scores, positing that WCF with metalinguistic information significantly facilitates adult learners in SLA. It is also interesting to note that even the coefficient of variance of D only group was the next highest indicator of the reliability of mean scores as it indicates the next lowest coefficient of variance at posttest 2(10.45). It is interesting to note that although the coefficient of variance of the DC group is relatively low at Posttest 1, it drastically increased at Posttest 2 confirming the assumption of the mean scores. It must also be noted here that relatively low and similar SD and variance scores of all the groups taken for the study indicate the reliability of test scores. Group 01-Control Group Variable Mean Variance Minimum Maximum Pretest Marks 31.50 24.45 25 40 Posttest 1 Marks 31.83 24.70 25 39 Posttest 2 Marks 31.33 20.42 25 39 Figure 1: Test Scores of Control group S t u d e n t M a r k s 1 21 11 0987654321 4 0 . 0 3 7 . 5 3 5 . 0 3 2 . 5 3 0 . 0 2 7 . 5 2 5 . 0 V a r ia b le P o s t t e s t 2 M a r k s P r e t e s t M a r k s P o s t t e s t 1 M a r k s C o n t r o l G r o u p What is obvious from Graph 1 is that the scores of control group remain the same at three testing sessions: it verifies the assumption that WCF facilitates acquisition of L2 in adult learners. But, the slight discrepancy in the posttest and pretest scores of the control group seems to imply that even mere exposure to L2 activities may correlate – either positively or negatively – with the accuracy of L2 learners’ grammar. However, this implication warrants future research. ENGLISH REVIEW: Journal of English Education ISSN 2301-7554 Vol. 1, Issue 2, June 2013 http://journal.uniku.ac.id/index.php/ERJEE Group -2 Direction only Variable Mean Variance Minimum Maximum Pretest Marks 31.33 12.97 25 35 Posttest 1 Marks 42.58 18.08 38 49 Posttest 2 Marks 43.17 20.33 38 50 Figure 2: Test Scores of Direction only group S t u d e n t s M a r k s 1 21 11 0987654321 5 0 4 5 4 0 3 5 3 0 2 5 V a r ia b le P o s tte s t 2 M a r k s P r e te s t M a r k s P o s tte s t 1 M a r k s G r o u p - 2 ( D i r e c t i o n o n l y ) Graph 2 very clearly demonstrates the efficacy of WCF, Direction only, in particular. The improvement learners have gained is highly remarkable and what is further noteworthy is that both at Post test 1 and 2, the scores remain consistent, indicating the significant efficacy of Direction only type of WCF in L2 acquisition. It highlights the fact Direction only is highly beneficial in retention of grammar acquired through WCF, perhaps for the purpose of long- term learning. The graph, however, indicates that the language accuracy perhaps gained by treatment, is not parallel with the pretest scores. For instance, language accuracy gained at places 3, 8 and 12 are very high whereas the gain at 2 and 10 are relatively low. This inconsistency seems to imply that mere location of an error may result in unpredictable results as well. However, this necessitates further investigation before drawing any conclusion. Group -3 (Direction and Correction Group) Variable Mean Variance Minimum Maximum Pretest Marks 31.42 24.27 25 41 Posttest 1 Marks 54.75 14.57 49 59 Posttest 2 Marks 35.33 26.61 29 45 CHITRA JAYATHILAKE Correcting Errors: The Relative Efficacy Of Different Forms Of Error Feedback In Second Language Writing Figure 3: Test Scores of DC group Students M a rk s 121110987654321 60 50 40 30 20 Variable Po sttest 2 Mark s Pretest Mark s Po sttest 1 Mark s Group -3 (Direction and Correction) Graph 3 demonstrates a highly significant insight into WCF: although the feedback given seems highly beneficial for learners at Posttest 1, the graph indicates a drastic fall of learners’ acquisition at Posttest 2. As indicated in the green line, the quantity of decrease in students’ scores is almost similar to the pretest scores. It is also noteworthy to mention, a high amount of Posttest 1 scores are displayed in this group. This suggests that WCF is of little use over time (for long-term acquisition), although it has been extremely facilitative for immediate or short term learning. Group -4 (Direction, Correction and Meta linguistic Information Group) Variable Mean Variance Minimum Maximum Pretest Marks 32.33 30.24 21 40 Posttest 1 Marks 57.58 22.63 49 65 Posttest 2 Marks 58.17 21.61 51 65 Figure 4: Test Scores of DCM group S t u d e n t s M a r k s 1 21 11 0987654321 7 0 6 0 5 0 4 0 3 0 2 0 V a r ia b le P o stte st 2 M a r k s P r e te st M a r k s P o stte st 1 M a r k s G r o u p -4 ( D i r e c t i o n , C o r r e c t i o n a n d M e t a l i n g u i s t i c I n f o r m a t i o n ) Figure 4 shows the results of the DCM group for the three testing periods. The pattern in the graph reveals that the DCM treatment group gains over time. ENGLISH REVIEW: Journal of English Education ISSN 2301-7554 Vol. 1, Issue 2, June 2013 http://journal.uniku.ac.id/index.php/ERJEE What is outstanding is the consistency in posttest marks: this remarkable consistent gain at both Posttests1 and 2 scores are relatively high and significant. Unlike in the graph 2, this graph shows that the pretest indicator and posttest indicators are greatly parallel implying consistency in language accuracy and the efficacy of metalinguistic information. This underscores the validly in metalinguistic feedback for adult learners’ SLA. CONCLUSION This research sought to investigate, primarily, whether WCF facilitates adult learners in SLA, focusing on one grammar element, namely the present continuous form. The students of the current study were of the same level of proficiency and received the same amount of scores at the Pretest. They were involved in identical writing tasks. During the period of the study, the students were not exposed to any explicit L2 learning except the treatment employed for the study. Therefore, WCF alone could be considered as responsible for the acquisition of the language element – present continuous form. Thus, it is clearly evident that WCF facilitated in improved accuracy in adult learners’ L2 acquisition (present continuous from) in both the short and long term. The central focus of this research paper, however, is the relative efficacy of different WCF types. Ferris’ (2003) research evidence suggests that indirect error feedback is more helpful on students’ long-term writing development than direct error feedback. Chandler’s (2003) research which investigates two ESL groups receiving either direct or indirect error feedback also showed that indirect error feedback with student self-editing contributes to accuracy more than direct error feedback. While supporting these hypotheses of Ferris and Chandler, the current study adds another insight into the phenomenon of WCF: this study suggests both indirect feedback where learners’ self-correction is required to a great extent and metalinguistic information where learners can receive awareness to their errors is equally helpful for the accuracy of L2 learning in adult learners. Simultaneously, it also attempted to find the benefit and the relative efficacy of WCF for long-term acquisition of L2 by adult learners. The results in Posttest 1 and Posttest 2 indicate that the WCF had a positive effect on the acquisition of present continuous form by adult L2 learners. In particular, feedback given with metalinguistic information proved to be highly effective in improving learners’ accuracy as it is at the highest level of efficacy and helps for retention of language elements. Direction only group is also significant, as it indicated second highest gain at posttests. This finding can again be explained with Long’s (1988) view where he strongly suggests that if learners focus on form, learners promote L2 learning in a very effective manner. In the current study all three WCF types are likely to promote awareness in L2, but, only direction only and direction, correction and metalinguistic information comments promote learners to focus on form in a considerable way. That is, when correction is not provided, learners will focus on form attempting to self-correct it and when metalinguistic information is provided also, learners will focus on form with an analytical understanding of the form. This relatively high focus on form benefited learners’ long-term use of L2. CHITRA JAYATHILAKE Correcting Errors: The Relative Efficacy Of Different Forms Of Error Feedback In Second Language Writing To round up, it can be argued that the results of the present study, on the whole, gain support from some previous research in the sense that WCF facilitates students improve their accuracy in writing despite the type of feedback (Chandler, 2003; Sheen, 1997). Moreover, these findings suggest that both metalinguistic information and opportunities for self correction serve to improve adult learners’ grammatical accuracy. The relative efficacy of feedback can be demonstrated along the following cline which may be employed in ESL contexts particularly at Sri Lankan university contexts. No WCF Direction Direction only Direction, Correction & & Correction Metalinguistic information The results of this preliminary study are limited by the small sample size and short term treatment. Future research can focus on different types of learner errors developing longitudinal treatment. The results will be significant for anyone engaged in any English langue teaching/learning contexts where English is used as a Second or Foreign language. This is particularly significant dealing with learner errors in such contexts. REFERENCES Brown, J.D. (1988). Understanding research in second language learning: A teacher’s guide to statistics and research design. Cambridge: Cambridge University Press. Chandler, J. (2003). The efficacy of various kinds of error feedback for improvement in the accuracy and fluency of L2 student writing. Journal of Second Language Writing, 12(3), 267-296. Erel, S & Bulut, D. (2007). Error Treatment in L2 Writing: A Comparative Study of Direct and Indirect Coded Feedback in Turkish EFL Context. Journal of Institute of Social Sciences, 23(1), 397-415. Ferris, D. R. (1997). The Influence of Teacher Commentary on Student Revision: TESOL Quarterly, 31(2), 315-339. __________.(1999). The case for grammar correction in L2 writing classes: A response to Truscott (1996). Journal of second Language writing, 8, 1-10. __________. (2003). Response to student writing: Implications for second language students. Mahwah, NJ: Lawrence Erlbaum Associates. __________. (2004). The “Grammar Correction” Debate in L2 Writing: Where are we, and where do we go from here? (and what do we do in the meantime…?).Journal of Second Language Writing .13,49-62. Ferris, D. R. & Roberts, B. (2001). Error feedback in L2 writing classes: How explicit does it need to be? Journal of Second Language Writing, 10(3), 161-184. Lee, I. (2007). Understanding teachers’ written feedback practices in Hong Kong secondary classrooms. Journal of Second Language Writing, 17, 69-85. Liu, Y. (2008). The Effects of Error Feedback in Second Language Writing. Arizona working Papers in SLA & Teaching, 15, 65-79. Long, M.H. (1988). Instructed Interlanguage Development. In L. Beebe (Ed.), Issues in second language acquisition. Multiple Perspectives (pp.115-141). Rowley, MA: Newbury House. Sheen, Y. (2007). The Effect of Focussed Written Corrective Feedback and Language Aptitude on ESL Learners’ Acquisition of Articles. TESOL Quarterly 41(2), 255-283. Truscott, J. (1996). The case against grammar correction in L2 writing classes. Language Learning, 46(2), 327-369.