Languaging, Corrective Feedback, and Writing Accuracy among Low Proficiency L2 Learners

Low proficiency L2 learners make frequent grammatical errors and apply inappropriate English language use and writing conventions despite learning the language for years. To reduce the grammatical and non-grammatical errors in the low proficiency L2 learners’ writing, teachers provide Written Corrective Feedback (WCF). However, a strong claim on WCF's adverse effects on L2 learners’ language development has prompted mixed responses on the WCF’s effectiveness. Numerous studies on WCF's effectiveness were conducted, and the findings showed inconclusive results. The present study aims to examine the ‘Written Languaging’ (WL) effect of Written Corrective Feedback (WCF) on low proficiency L2 learners’ writing accuracy. WL is a mediational tool used to enhance the WCF’s effectiveness. The study was conducted on 47 low proficiency L2 university students for five weeks to elicit WL of WCF's effects on low proficiency L2 learners’ writing accuracy, where a multiple-case research approach was used to collect and analyze the data of the study. An error ratio was applied to measure the overall accuracy of writing. The Wilcoxon signed-rank test was used to determine whether WL's provision on the WCF affects the writing accuracy of the low proficiency L2 learners' new writing text. The finding shows that WL's provision on the WCF did not facilitate the low proficiency of L2 learners' writing accuracy because the L2 learners' writing accuracy could improve only with teachers' WCF. The study is significant to the pedagogical teaching of writing development for lower proficiency L2 learners, particularly in applying the mediational tool in the provision and processing of WCF and the complementary pedagogical strategy to WCF.


Introduction
To be competent in a language, one must master writing skills (Buyukyavuz & Cakir, 2014) because writing enhances the learners' ability to comprehend ideas and concepts (Voon Foo, 2007).However, for some learners, especially second language (L2) learners, writing is a skill that is difficult to master (Mastan & Maarof, 2014).The difficulties to write have been shown in studies on problems and difficulties learners face while writing the target language (e.g., Busch, 2010;Chen, 2002;Daud, 2012;Erkan & Saban, 2011;Hisham, 2008).Writing with accuracy is a daunting task for many L2 learners, especially the low proficiency L2 learners, who always need their teacher to correct them since they have limited ability to perform in L2.Therefore, teachers must give feedback in the writing classroom because it guides students in their writing development (Ene & Kosobucki, 2016;Ferris & Robert, 2001;Gower, Philips & Walters, 1995;Hyland & Hyland, 2006).Teachers also need to ensure that students are engaged with the feedback given by deliberating the feedback either through oral or writing expression so that they understand and can internalize the feedback given so that it can be applied in the revision or new piece of writing (Bitchener & Storch, 2016).For the low proficiency L2 learners, Written Corrective Feedback (WCF) is necessary.However, a strong claim on WCF's adverse effects on L2 learners' language development (Truscott, 1996) has prompted mixed responses on the WCF's effectiveness.Although the claim was made years ago, studies on the role of WCF have continued to interest L2 Acquisition (SLA) and L2 Writing (SLW) scholars since then.There is growing evidence of the positive relationship between WCF and writing accuracy over time.However, the findings have been inconclusive, and this has prompted further investigations into how WCF's effectiveness can be improved for low proficiency L2 learners to interpret teachers' WCF accurately.
For low English proficiency learners, applying similar errors in the revision writing despite getting WCF from their teachers is unavoidable (Silver & Lee, 2007).Their ineffective use of writing strategies and lack of metacognitive knowledge in English are among the reasons they have writing problems (Hyland, 1998).Despite getting the WCF, these learners still have errors in their writing.Thus, they need to engage with the WCF provided by their teachers to enhance uptake and retention (Lee, 2013) to positively affect their writings (Jerry, Mohd Jan & Samuel, 2013;Kassim & Ng, 2014a, 2014b).However, for the low proficiency L2 learners to successfully uptake and retain the corrections provided for subsequent written tasks, they need to notice and process the WCF.This area of how L2 learners notice and process WCF receives little attention in WCF research (Jonsson, 2012;Kim, 2013).
Nevertheless, it has been studied a lot in the field of Oral Corrective Feedback (OCF), where studies have shown that L2 learners' understanding of the corrective feedback provided is often different from teachers' objective (Egi, 2007;Kim & Han, 2007;Mackey, Gass & McDonough, 2000).Therefore, noticing and understanding teachers' corrective feedback does not necessarily show that L2 learners have accurately interpreted teachers' corrective feedback.Other pedagogical writing tools or strategies promoting the need to notice and understand the WCF should be applied so that learners' engagement in noticing and understanding the WCF can be enhanced to improve L2 learners' writing accuracy.Thus, we suggest that L2 learners need to language or 'languaging' the WCF provided by their teachers to improve their writing accuracy.
The low proficiency L2 learners generally make frequent grammatical and non-grammatical errors and apply inappropriate use of the English language despite learning the language for years (Normazidah, Koo & Hazina, 2012;Singh, Singh, Nur Qistina & Ravinthar, 2017).Despite getting constant corrective feedback from teachers in oral or written form for the language and writing errors they make in the writing tasks, the writing accuracy of the low proficiency L2 learners in an English course at a local university in East Malaysia, is relatively low.This group of L2 learners was given a 'Written Languaging' (WL) to encourage them to engage more with the given corrective feedback, which in this study is the WCF.The WL applied in this study would function as a tool to enhance the WCF's effectiveness.It is recommended that L2 students be allowed to 'languaging' the corrective feedback given to them (Suzuki, 2012;Moradian et al., 2017) to allow them to identify differences and discrepancies between their interlanguage and the target language (English) and learn about English writing features that they are not aware of or forget (Suzuki, 2009b(Suzuki, , 2012;;Swain, Lapkin, Knouzi, Suzuki & Brooks, 2009).WL prompts L2 learners to think and reason thoroughly about the errors found in their writing (Suzuki, 2012), which influences them to do the task efficiently and increase their attention to achieve self-regulatory abilities.Studies have shown that the WCF received by L2 learners becomes more effective when L2 learners apply WL, resulting in improved writing accuracy.Therefore, this notion of applying WL of WCF to improve writing accuracy underpins this study's rationale.
Furthermore, studies on WL of the WCF of L2 learners writing in the Malaysian context are still not advanced.Thus, this study's findings will benefit the pedagogical teaching of writing development in Malaysia, particularly in applying the mediational tool in the provision and processing of WCF and the complementary pedagogical strategy to WCF.The study aims to elicit whether the WL of the WCF affects the low proficiency L2 learners' new writing text's accuracy.It attempts to answer this research question: Does the WL of the WCF of low proficiency L2 learners' writing text affect their new writing text's accuracy?

Sociocultural Theory
The theoretical framework of the study is Sociocultural Theory (SCT).According to SCT, learners' L2 development is enhanced when a social interaction between a novice (L2 learner) and an expert (L2 teacher).This interaction acts as a form of assistance that an expert offers to a novice; it can be in a physical (as in dictionary) or symbolic (as in language) form (Bitchener & Ferris, 2012).In SCT, this form of assistance is called mediation and tools.WCF is considered a physical form of assistance or a tool in SCT (Bitchener & Storch, 2016) that teachers (the experts) provide for the errors found in the writing of L2 learners (the novices).From the perspective of SCT, learners play an active role in learning.
Therefore, they will respond and take advantage of the assistance offered, which is the WCF, by applying the correct forms of the errors found in their writing.It will eventually lead to the learners' L2 development.SCT also highlights the role of mediation in language learning, which can be achieved via material or symbolic tools.The material tools, like the WCF, enable actions to take place.L2 learners edit the errors in their writing following the correction given by their teachers (WCF), which shows that the errors' editing has occurred.The symbolic tool, on the contrary, enables and shapes the action of editing the errors.'Languaging' is an example of a symbolic tool that L2 learners use to communicate with the experts (the teachers) and organize their actions (Wells, 1999).As a mediational tool, 'languaging' facilitates the formation of ideas that enable high-level cognitive processes, such as self-regulation (Bitchener & Storch, 2016).The theory also asserts that a novice will get appropriate assistance from an expert if they collaborate.However, not all assistance forms are helpful and supportive of L2 development (Lantolf & Thorne, 2006).It is assumed that excessive assistance will result in L2 development inhibition (Bitchener & Storch, 2016).Thus, it is essential to provide an 'appropriate' amount of assistance that is just enough for learners to perform beyond their current capabilities.

Writing Accuracy
To write accurately means to write with no errors (Arnold, 2008;Khorasani & Sadzadeh, 2015), and the language produced in writing should conform to the norms of the L2 (Skehan, 1996).L2 learners, who produce fewer writing errors, write more accurately as they become more proficient in the language (Arnold, 2008;Wolfe-Quintero, Inagaki, & Kim, 1998).However, for low proficiency L2 learners, writing accurately can be a daunting task.Therefore, getting WCF from teachers is crucial because it enhances L2 learners' language learning development and motivation (Hyland & Hyland, 2006).There is growing evidence showing the positive relationship between WCF and writing accuracy improvement over time (Bitchener & Ferris, 2012).With WCF, L2 learners notice the errors found in their writing (Long, 1996).However, noticing alone does not guarantee learners' accurate interpretation of teachers' WCF (Kim, 2012).
To interpret teachers' WCF accurately, L2 learners need to 'languaging' the WCF to uptake and retain the corrective feedback for future writing tasks.In this study, they need to 'Written Languaging' (WL) the WCF received.Suzuki (2012) proposed that WL provides the platform for L2 learners to think and reason thoroughly about the errors they make in their writing.WL influences learners to do the task efficiently and heightens their attention to achieve their self-regulatory abilities.In his study on WL's effect in response to indirect WCF on enhancing learners' writing accuracy over revision, the finding shows that WL generated by indirect WCF affects learners' writing accuracy significantly (Suzuki, 2009a).He obtained a similar result when exploring WL's effectiveness in response to direct WCF on L2 Japanese students' writing accuracy over revision tasks (Suzuki, 2012).
Later, Moradian et al. (2017) conducted a similar study on two groups of low-intermediate Iranian EFL students, and the study also revealed that WL enhances the efficiency of WCF, resulting in significant effects on the students' writing accuracy.All these show that WCF received by L2 learners become more effective when L2 learners apply WL.
L2 learners need to engage with the WCF provided by their teachers to enhance uptake and retention (Lee, 2013) to positively affect their writings (Jerry, Mohd Jan & Samuel, 2013;Kassim & Ng, 2014a, 2014b).However, for L2 learners to successfully uptake and retain the corrections provided for subsequent written tasks, they need to notice and process the WCF.This area of how L2 learners notice and process WCF receives little attention in WCF research (Jonsson, 2012;Kim, 2013).In addition, studies have shown that L2 learners' understanding of the corrective feedback provided is often different from teachers' objectives (Egi, 2007;Kim & Han, 2007;Mackey, Gass & McDonough, 2000).Therefore, noticing and understanding teachers' corrective feedback does not necessarily show that L2 learners have accurately interpreted teachers' WCF.Other pedagogical writing tools or strategies promoting the need to notice and understand the WCF provided should be applied so that learners' engagement to notice and understand the WCF can be enhanced---to improve L2 learners' writing accuracy.In relation to this, L2 learners need to 'Written Languaging' (WL) the WCF provided by their teachers to improve their writing accuracy.

Method
We chose a qualitative, multiple-case research approach for this study because it provides a holistic and in-depth explanation of the issues concerned by allowing us to explore beyond the quantitative statistical results and investigate current real-life phenomena through detailed contextual analysis of a limited number of conditions or settings, and their relationships (Zainal, 2007).Also, qualitative, multiple-case research examines the problems studied within the context in which the problems occur (Yin, 2014;Zainal, 2007).This research approach explores or describes the problems in a real-life setting and explains their complexities in real-life situations that may not be shown through experimental or survey research (Starman, 2013).Two groups were studied for this multiple-case study research study: the treatment and control groups.The Treatment Group (TG) was asked to WL or write comments/opinions on the WCF done by the teachers.The Control Group (CG), on the other hand, only did the revision essay by making any changes or corrections in response to the WCF provided by the teachers.They did not have to WL the WCF provided.The CG acted as the baseline measure of the study.This stage was necessary for the study because it provided findings that could be compared between the two groups to determine whether WL's provision on the WCF (as the treatment) affected the participants' writing accuracy in a subsequent essay, namely writing tasks 2 and 3.The cycle was repeated for each of the timed essays because it helps us to examine in-depth and understand better the problem studied in this study: will the participants' writing accuracy in the subsequent writing tasks 2 and 3 be improved by the provision of WL on the WCF provided by the teachers, or their writing accuracy can solely be improved with revision.To make sure the participants understood the teachers' WCF, they were provided with a list of error codes and brief descriptions (see Table 3.1) to refer to during the revision and intervention sessions---an intervention session where the participants were asked to WL or give comments/opinions on the WCF provided by their teachers (for the TG), and a revision session where the participants were asked to revise their essay based on the WCF provided by their teachers (for the CG).
The study was conducted for five weeks.During the five weeks of study, the participants needed to complete the writing tasks, where they were asked to do three in-class writing tasks with different prompts.We implemented the intervention examined, namely the WL task application on the teachers' WCF in Weeks 2 and 4. Table 3.1 shows the time of the data collected and treatment.In-class writing task 3 Storch (2010) and Duff (2006) argue that studies done in natural and authentic contexts have more significant pedagogical relevance for language teachers.Therefore, the study used the prompts adapted from the textbook used for the course taken by the students, i.e., Q: Skills for Success-Reading and Writing 3 (Ward & Gramer, 2015).Based on the prompts, the participants were asked to compose a five-paragraph essay in each writing session.Table 3.2 shows the prompts for each writing task.s The researchers conducted a pilot study on the writing tasks and the WL task before the actual inclass writing task 1 to examine the feasibility of the prompts used in the writing tasks and the WL task instructions.For the actual study, the researchers asked the participants to WL the WCF provided to see whether the intervention had caused the change in the writing accuracy, where the writing became more accurate (grammatically and non-grammatically) in the new essay writing of the participants' writing overtime.

The Participant
We collected the study data from 47 low proficiency L2 learners who took the Reading and Writing in English course at the time of the study at the Centre of Promotion of Knowledge and Language Learning (PPIB), Universiti Malaysia Sabah (UMS).A purposive opportunistic sampling, a non-probability sample, was applied to select the study's participants.According to Creswell and Clark (2011), purposive sampling involves identifying and selecting individuals or groups of knowledgeable and experienced individuals about the issues or problems studied.They should also be available and can communicate their experiences and opinions coherently in a deliberative manner (Bernard, 2011).
The researchers purposely selected these participants for this study because they were directly involved with the experience of getting corrective feedback for their writing task since they took the Reading and Writing in English course for that semester.They could be the best sample that can demonstrate the WL's effect on the WCF's efficacy on the writing accuracy of low proficiency L2 students.Thus, the study participants' selection is opportunistic, that the familiarity and convenient access to information and participants had become an advantage to complete the study's data collection on time.The participants have also learned and were exposed to English grammar, such as the parts of speech, subject-verb agreement, and tenses, in the Communicative English Grammar (Level 1) and Oral Communication in English (Level 2) in their first year of the study in UMS.Reading and Writing in English is the Level 3 of the foundation course that is compulsory for the participants to take in the second year of their study at UMS.The English foundation courses (i.e., Communicative English Grammar, Oral Communication in English and Reading and Writing in English) are offered to students with MUET Bands 1, 2 and 3. Malaysian University English Test (MUET) is an English proficiency test administered by the Malaysian Examination Council (MEC) for Malaysian students who wish to pursue a first-degree program at any local university.In MUET, students are tested on the four skills, namely speaking, writing, listening, and reading.The total score obtained for MUET will determine the students' MUET band, either Band 1 (the lowest proficiency) or Band 6 (the highest proficiency).In this study, the participants had obtained MUET Bands 2 and 3, and are described as a not fluent and modest user of the English language because they can hardly use the language, make frequent grammatical errors, have limited understanding of the language, and have limited ability to function in the language (MUET, 2015).MUET Bands 2 and 3 are equivalent to CEFR (Common European Framework of Reference) A1 and CEFR A2 or IELTS (International English Language Testing System) Bands 2 (intermittent user) and 3 (extremely limited user).

The Instrument
The instrument used in this study was writing tasks.The writing tasks of the study were analyzed for writing accuracy.An error ratio was applied to measure the overall accuracy of writing.This assessment measure was used to study the effectiveness of corrective feedback (see Chandler, 2003;Karim, 2013;Truscott & Hsu, 2008;van Beuningen et al., 2012).The error ratio used to measure the writing texts' overall accuracy is the total number of errors divided by the total number of words written.The result is then multiplied by 100.A 100word ratio is used to interpret the total number of errors in percentages (it provides error rates in students' writings per 100 words).The errors in this study refer to grammatical and non-grammatical errors.The errors' accuracy rate of the writing tasks was calculated and compared based on other studies on the effects of WCF on the grammatical and nongrammatical accuracy of L2 writings (see Karim, 2013;van Beuningen et al., 2011).Table 3 shows the grammatical and non-grammatical errors.
The Wilcoxon signed-rank test of SPPS was also applied to measure the differences between the writing tasks' error ratios of the CG and TG.The Wilcoxon signed-rank test is a non-parametric test used to analyze repeated measures data where the participants are assessed for more than two occasions or conditions.The test was used to elicit the significant differences in the error ratio of the participants' writing tasks across occasions, which in this case, the writing tasks 1, 2, and 3 of each group (CG and TG) that were conducted in a different time, namely Writing Task 1 (WT1) was conducted in Week 1, Writing Task 2 (WT2) in Week 3, and Writing Task 3 (WT3) in Week 5.The differences in the participants' writing error ratio will determine whether the intervention (the application of WL of the WCF) affects the participants' writing accuracy over time.In addition, the participants' total errors (the minimum and the maximum number of errors) and the total number of words (the minimum and the maximum number of words) of each writing task were analyzed descriptively using SPSS and then compared manually in this study.To simplify each error category's counting, we applied (with permission) a coding system used in Ferris, Liu, Sinha, and Senna's (2013) study.Table 3.3 shows the coding system.The study's intra-rater and inter-rater were also calculated to ensure consistency in implementing the assessment measures employed in analyzing the participants' writing texts.A rater checked some of the participants' writings a second time to check the scoring's reliability after the first rater, a native English speaker, had done the initial scoring and analysis.Then another rater would check some of the participants' writing individually, and this was to ensure excellent inter-rater reliability.The raters have a degree and Master's degree in Teaching English as a Second Language (TESL) and have also been teaching English language skills for more than five years.

Result
Tables 4.1 and 4.2 show the mean (M) and standard deviation (SD) for the number of words, total errors, and error ratio of the participants of CG (N=22) and TG (N=25), respectively.In Table 4.2, the mean and SD of the number of words produced, the total error found, and the error ratio of WT1 for the participants of the Treatment Group (TG) were 199.48 (SD=69.792),25.60 (SD=10.344),and 2.16 (SD=0.800)respectively.In WT2, the number of words produced, on average, was 201.68 (SD=50.586),and the total error was 25.76 (SD=11.054).The error ratio of WT2, on average, was 2.12 (SD=0.781): a slight increment from the error ratio in WT1.The number of words in WT3, on average, had increased from 201.68 (SD=50.586)(WT2) to 236.56 (SD=82.265).The error ratio of WT3, nevertheless, remained the same, i.e., 2.12 (SD=0.781).Perhaps this was due to the average number of words produced in WT3, i.e., 236.56: the average number of words produced in WT3 was more than the average number of words produced in WT2.The rise in the number of words produced in WT3 might have affected the error ratio.4.3 shows that the minimum total number of words of the CG participants' writing task 1 (WT1) was 128 words, and the maximum total number of words was 322 words.The minimum and the maximum total number of words were then increased in writing task 2 (WT2), i.e., 150 words and 396 words, respectively.However, the total number of words in the CG participants' writing task 3 (WT3) had decreased to a minimum of 117 words and a maximum of 377 words.For the total number of errors found in the CG participants' writings: the minimum total number of errors of the WT1, WT 2, and WT 3 had decreased gradually, i.e., from the minimum of ten errors in the WT1 to seven errors in the WT2, and two errors in the WT3.Nevertheless, the maximum total number of errors in the CG participants' writings shows inconsistent results.The maximum number of errors in the WT2 had increased from 41 in the WT1 to 48 in the WT2.Then the number decreased to 34 errors in the WT3.Table 4.4 illustrates the results for the TG participants' writing tasks: the minimum total number of words in the WT1 was 110 words, and the maximum total words were 448 words, whereas, in WT2, the minimum total number of words was 120, which then increased to 126 words in the WT3.The maximum total number of words in the WT2 was 330 words (the number has decreased compared to the WT1 maximum total number of words, i.e., 448 words).However, the maximum total number of words in the WT3 had increased to 516 words (higher than the maximum number of words produced in the WT3 of the CG participants, i.e., 377 words).The total number of errors in the CG participants' writing tasks and the TG participants' total number of writing task errors show inconsistent results.The total number of errors in the WT2 had increased to a minimum of ten errors and a maximum of 56 errors compared to a minimum of nine errors and a maximum of 55 errors that occurred in the WT1.The total number of errors was then decreased to a minimum of seven errors and a maximum of 46 errors in the WT3.
The Wilcoxon signed-rank test was used to determine whether or not the provision of 'languaging' on the WCF affects the writing accuracy of the TG participants' new writing texts (WT2 and WT3).The test was also used to determine whether the writing accuracy of the CG participants' new writing texts (WT2 and WT3) have improved only with the teacher's WCF and the revision texts of the WT1 and WT2.The Wilcoxon signed-rank test was used to investigate any changes in scores when the same individuals are subjected to more than one condition.To see the changes, the researchers compared the following combinations: a.Total Error: WT1to WT2 WT1 to WT3 WT2 to WT3 b.Total Words: WT1to WT2 WT1 to WT3 WT2 to WT3 c. Error Ratio: WT1to WT2 WT1 to WT3 WT2 to WT3 The Wilcoxon signed-rank tests (see Table 4.5) were conducted with a Bonferroni correction applied, resulting in a significant level set at p < 0.017.Median (IQR) total errors for the WT1, WT2, and WT3 were 21.50 (15 to 27), 24.50 (17 to 32.50), and 15.00 (8.75 to 20.25), respectively.Table 4.5 shows no significant difference between the total errors in WT1 and WT2 running trials (Z=-0.906,p=0.365).However, there was a statistically significant reduction of total errors in the WT1 and WT3 (Z=-2.615,p=0.009) and in the WT2 and WT3 (Z=-3.396,p=0.001) trials.The median (IQR) total words for the WT1, WT2, and WT3 of this group were 216.00 (162 to 279.50), 262.50 (222.25 to 327.00), and 251.00 (196.50 to 298.75), respectively.Table 4.5 illustrates a significant reduction of total words in the participants' WT1 and WT2 (Z=-2.614,p=0.009) trials.However, there was no significant difference between the total number of words in WT1 and WT3 (Z=-2.260,p=0.024) and WT2 and WT3 (Z=-0.860,p=0.390) running trials.Next, the median (IQR) error ratio for the participants' WT1, WT2,and WT3 was 10.70 (8.23 to 13.55),8.95 (6.45 to 14.35),and 6.85 (3.8 to 8.98), respectively.Based on the Wilcoxon signedrank test, there was no significant difference in the error ratio of the CG participants' WT1 and WT2 (Z=-0.467,p=0.641).However, there was a significant difference in the error ratio of WT1 and WT3 (Z=-3.491,p=0.000) and the error ratio of WT2 and WT3 (Z=-3.337,p=0.001) running trials.Table 4.6 shows the Wilcoxon signed-rank test for the TG participants' writing.The Wilcoxon signedrank tests were conducted with a Bonferroni correction, resulting in a significant level set at p < 0.017.Median (IQR) total errors for the WT1, WT2, and WT3 were 23.00 (19 to 30.50), 23.00 (17 to 31.50), and 11.00 (11.00 to 21.50), respectively.There was no significant difference between the total errors in WT1 and WT2 running trials (Z=-0.543,p=0.587).However, there was a statistically significant reduction of total errors in the WT1 and WT3 (Z=-2.901,p=0.004) and WT2 and WT3 (Z=-2.784,p=0.005) trials.For the total number of words in writing, the median (IQR) total words for the WT1,WT2,and WT3 of this group were 194 (156 to 220.50),200 (166 to 231),and 225 (172.50 to 284.50), respectively.Nevertheless, there was no significant reduction in total words in the participants' WT1 and WT2 (Z=-1.129,p=0.259) trials.Also, there was no significant difference between the total number of words in WT1 and WT3 (Z=-2.328,p=0.020) and WT2 and WT3 (Z=-1.957,p=0.050) running trials.Next, the error ratio: the median (IQR) error ratio for the participants' WT1, WT2 and WT3 was 13.30 (10.75 to 15.95),12.70 (9.35 to 14.65),and 6.7 (5.6 to 9.4), respectively.Based on the Wilcoxon signed-rank test, there was no significant difference in the error ratio of the TG participants' WT1 and WT2 (Z=-0.467,p=0.641).However, there was a significant difference in the error ratio of WT1 and WT3 (Z=-3.491,p=0.000) and the error ratio of WT2 and WT3 (Z=-3.337,p=0.001) running trials.
Based on the findings shown in Tables 4.5 and 4.6, it seems that the participants' writing in both groups (TG and CG) has improved in terms of total errors and error ratio.Using the Wilcoxon signed-rank test, the results showed a significant improvement in the number of total errors.The error ratio (in the writings of both TG and CG) was found in the WT3 compared to the total number of errors found in the WT1, and the same result was also yielded in the WT3 when compared to the total errors of the WT2.It seems to indicate that with or without the assistance of the WL task, the TG and CG participants' writing accuracy could improve solely with teachers' WCF.Therefore, does the WL of the WCF of low proficiency L2 learners' writing text affect their new writing text's writing accuracy?The finding shows that the provision of 'languaging' (WL) on the WCF of the writing texts does not facilitate the writing accuracy in the new writing text of low proficiency L2 learners, which is the participants of TG of the study.

Discussion and Suggestions
The study investigated the effects of WL as the complementary pedagogical strategy to the WCF on the low proficiency L2 learners' writing accuracy.The writing and WL tasks were used to collect the data, and 47 L2 learners participated in the study.The study found that the provision of WL on the teacher's WCF of the writing tasks does not affect the low proficiency learners' writing accuracy.It does not facilitate the writing accuracy of the low proficiency learners' new writing text.
Studies on the effect of the mediational tool, such as WL, in the provision and processing of WCF, have illustrated that L2 learners' writing accuracy can be improved significantly (Bitchener, 2008;Bitchener & Knoch, 2008;Bitchener, Young & Cameron, 2005;Ellis, Sheen, Murakami & Takashima, 2008;Suzuki, 2009aSuzuki, , 2009bSuzuki, , 2012;;Moradian et al., 2017).However, most of the studies conducted were on learners with intermediate or advanced proficiency levels.For a low proficiency learner, who makes frequent grammatical errors and has limited ability to function in the language, teachers' WCF is essential.The WL of the WCF has supposedly helped the participants internalize the English Language's grammar and writing conventions that will result in minimal or zero errors in the new writing text.However, for this study, the finding seems to suggest that the provision of WL of the WCF on the low proficiency L2 learners' writing texts does not facilitate the writing accuracy of the new writing text.The Wilcoxon signed-rank test results also indicate that the writing accuracy of the low proficiency L2 learners of this study could improve solely with their teachers' WCF.
The study has also shown that despite Truscott's opposition to the application of corrective feedback in an L2 class, L2 learners need CF, especially when the L2 learners' proficiency level is low.Low proficiency L2 learners need teachers' corrective feedback to solve their language-related problems, whether grammatical or non-grammatical.As advocated by the Sociocultural Theory (SCT), learners' L2 development is enhanced if there is a social interaction between a novice (the low proficiency L2 learners) and an expert (the L2 teacher).This social interaction also acts as assistance or the mediational tool that an L2 teacher can offer to low proficiency L2 learners.WL is a form of assistance or the mediational tool used in the study to enhance the efficacy of WCF.However, the study found that the mediational tool, which in this study is WL, has an insignificant role in improving the low proficiency L2 learners' writing accuracy.
Based on the findings, we recommend that L2 teachers provide only an 'appropriate' amount of assistance that is enough for L2 learners to perform beyond their existing capabilities.It is considered that excessive assistance would negatively affect L2 learners' language development (Bitchener & Storch, 2016).To provide an 'appropriate' amount of assistance, we had identified the L2 learners' level of proficiency (as suggested by SCT) as low proficiency L2 learners based on their MUET results.They obtained Bands 2 and 3, described as not fluent and have limited functions in English.Based on their proficiency level, we assumed that these students needed extra assistance throughout their task performance, which is the writing task in this study.Hence, a WL of the WCF provided by class teachers was introduced.This assistance was supposed to make the low proficiency learners better understand the errors pointed out in the teachers' WCF after the implementation of WL of the WCF gradually and to produce the essay with minimal writing errors.However, the finding shows the contrary.Thus, it is suggested that the L2 teachers teaching the low proficiency L2 students provide only the 'appropriate' amount of assistance that is enough for them to perform well in any given task (as proposed by SCT).Also, the practice of WCF should be improved.Rather than only providing the corrective feedback in writing (WCF), the teachers should explain the WCF in detail verbally.The teachers need to have a face-to-face session explaining the WCF because this will give an opportunity for the low proficiency L2 learners to ask for more details if they have problems comprehending the WCF provided.The low proficiency L2 learners' uptake and retention will further improve if they are made aware of the errors found in their writings.

Conclusion
So, does the WL of the WCF of low proficiency L2 learners' writing text affect their new writing text's accuracy?The answer is no: the WL of the WCF of the low proficiency L2 learners' writings does not affect the learners' new writing texts' accuracy.The findings also show that despite getting extra assistance, which in this study refers to the WL of the WCF, the low proficiency L2 learners (the TG participants) did not perform in the new essay writing tasks (WT2 and WT3).It confirms the claim proposed by SCT that is L2 teachers should provide only an 'appropriate' amount of assistance that is just enough for the L2 learners to perform beyond their current capabilities.The excessive assistance given to the TG participants of the study did not affect their writing accuracy.Therefore, L2 teachers teaching low proficiency L2 learners should avoid providing excessive assistance to the already provided WCF.Instead, they should maximize the role of WCF by becoming more aware of their learners' needs and the types of WCF that are preferred by these low proficiency L2 learners.As elicited in the study, the low proficiency L2 learners (the CG participants) can perform well solely with the WCF provided by L2 teachers.Thus, by eliciting the learners' needs and the types of WCF preferred, they will become more active and motivated to improve themselves under appropriate guidance from their teachers.

Table 3 .2
In-class Writing Tasks' Prompts

Table 3 .3
Coding System Used in Marking of the Timed Writing Tasks

Table 4 .1
Mean and SD for the Number of Words, Total Errors and Error Ratio (N=22) for the CG Mean and SD for the Number of Words, Total Errors and Error Ratio (N=25) for the TG

Table 4 .3
The Minimum and Maximum Number of Words and Total Errors (N=22) for the CG

Table 4 .4
The Minimum and Maximum Number of Words and Total Errors (N=25) for the TG

Table 4 .
5 and Table4.6 show the results of the Wilcoxon signed-rank test of the CG and the TG, respectively.

Table 4 .5
Wilcoxon Signed-Rank Test for the Total Error, Total Words and Error Ratio of the CG participants' Writing Tasks for the CG

Table 4 .
6. Wilcoxon Signed-Rank Test for the Total Error, Total Words and Error Ratio of the TG participants' Writing Tasks for the TG