A preference analysis and justification of Arabic written corrective feedback among instructors and undergraduates

There has been extensive discussion on the need to use corrective feedback in writing within foreign language learning. Essentially, corrective feedback is one of the important tools in improving students’ skills in learning a language. This study aims to find out the preference and justification of written corrective feedback (WCF) through the use of Google Docs among instructors and students in a higher learning institute. The effects of the direct and indirect feedback with metalinguistic comments were also studied to determine their suitability in teaching and learning the Arabic language. Quantitative and qualitative data were collected to (1) identify the preferred type of feedback among instructors and students, (2) identify justification of the preferred feedback type, and (3) examine post-test score differences between types of written correction feedback. Two questionnaires were adapted and distributed to 93 first-year students and four instructors of Arabic language for Academic Writing. Two instructors and five students were interviewed to find out their justification of the preferred types of WCF. A total of 50 respondents were divided into two groups according to the type of WCF provided, and post-test scores between the types of feedback were compared to determine if there was any significant difference between the types of feedback. The findings show that instructors prefer indirect WCF with metalinguistic comments while students prefer direct corrective feedback with metalinguistic comments. Post-test scores indicate that higher scores were achieved by students who received indirect feedback with metalinguistic comments. This indicates that students are able to process indirect feedback that is supplemented with metalinguistic comments. Moreover, an online learning environment provides more opportunities for instructors to highlight the students’ errors more clearly.


INTRODUCTION
Written correction feedback refers to the teacher's reaction to the students' errors by informing them the error so that it can be corrected and not repeated in subsequent writing (Van Beuningen, 2010). Most teachers and students agree that producing a good medium for delivering corrective feedback is still relatively new in Arabic language teaching and learning. The idea of using online technology in the delivery of corrective feedback is an active initiative to draw students' attention to their mistakes in writing and that the role of feedback is to overcome these weaknesses.
In general, WCF plays an important role in the formation of students' metalinguistic awareness through their attentiveness to restricted information (Bitchener & Ferris, 2012;Sato & Loewen, 2018). According to Heift and Hegelheimer (2017) and AbuSeileek and Abualsha'r (2014), students who receive corrective feedback through the use of computers while writing receive better results than those who do not receive it, as well as to learn from any mistakes they make. However, Bodnar, et al. (2017) found that not all types of computergenerated corrective feedback had positive effect on students' writing development. Hence, pedagogical use of technology in the delivery of corrective feedback is one of the issues that educators and teaching designers need to address in order to build meaningful student communicative interactions (Heift & Hegelheimer, 2017).
The use of technology alone does not guarantee that every learning outcome planned would be achieved, but thorough planning is a must to ensure that students gain benefit from the feedback provided. Therefore, there is a need to study the technique of delivering computermediated corrective feedback that can help to improve students' writing skills. Although many studies have examined the effectiveness of computer-assisted corrective feedback (Tafazoli et al., 2014), not many studies have looked at the role of online learning as a platform for delivering corrective feedback in foreign language classrooms.

Written corrective feedback in foreign language learning
There is a long discussion about the need to use corrective feedback in learning foreign language writing. Truscott (1999), for example, criticized the ability of feedback in improving students' writing in which he described it as wasting teachers' and students' time. In addition, it has been claimed that corrective feedback does not help but hindering the development of students' writing skills (Daneshvar & Rahimi, 2014;Laurel & Mostafa, 2017). These statements have received negative criticism. Many studies have shown positive effects of corrective feedback in foreign language learning (Afitska, 2015;Van Beuningen et al., 2012). Ferris (1999), in a study, found that WCF is related to students' motivation that they have become independent to correct their own errors in their writing. Ferris and Roberts (2001) also claimed that corrective feedback has a positive effect on second language learners' writing and this is supported by Biber et al. (2011) where they found that students' accuracy, content and form of writing are improved as the results of corrective feedback given by teachers. Therefore, WCF is essential for second and foreign language learners to become proficient in the language they are learning.
These conflicting findings has led to several other studies that seek more certainty on the role of WCF in second or foreign language learning. Among them is a study conducted by Amrhein and Nassaji (2010) which found that students prefer direct feedback while teachers prefer indirect feedback. One of the reasons students prefer direct feedback is because they have no knowledge of the principle of error correction used by teachers (Norouzian & Farahani, 2012). Based on these studies, there are several factors that lead to differences in findings regarding the effectiveness of WCF. Among others, students do not fully understand the feedback given (Razali, 2014), students only pay attention to the type of feedback they like (Schulz, 2001), student limited language skills and the scope of feedback given (Kang & Han, 2015).

Related studies on technology-assisted corrective feedback
Studies have been conducted on technology-assisted corrective feedback. The findings of these studies provide a positive indication of the effectiveness of corrective feedback that is delivered online. These include helping to develop students' writing skills (Duff & Li, 2009), improving communication skills through writing (Lee, 2005), reducing the psychological stress of students who do not like to receive face-to-face feedback (Vinagre & Munoz, 2011). There are also studies done in comparing technology-assisted corrective feedback to the traditional corrective feedback practices. It has been found that technology-assisted corrective feedback is more effective in helping students to identify mistakes in writing, and it encourages the habit of reviewing writing and improving their writing skills collaboratively (Fuente, 2016;Hosseini, 2012).
The use of Google Doc is also seen as a potential platform to provide collaborative WCF. Various functions available in Google Doc, such as chat and word editing, can systematically aid the development of student writing skills (Diez-Bedmar & Perez-Paredes, 2012). A study was conducted by Hosseini (2012) through experiments on the use of written feedback using online annotators among English as a Foreign Language learners. The purpose of this experiment was to find out the effectiveness of technology-assisted correction feedback and feedback provided on paper. The results of this experiment showed that groups using the online system could identify more writing errors than groups that did not use the system. AbuSeileek and Abualsha'r (2014) conducted a study that focused on the use of functions found in Microsoft Word 2010 to give feedback on EFL learners, and found that the use of computer-assisted corrective feedback has a positive effect on students' achievement in the written test. However, the types of feedback preferred by instructors and students, and the impact of its use on foreign language teaching and learning through Google Doc require further study.
The studies mentioned above examined the effect of technology-assisted corrective feedback on ESL and EFL learners, while there have been very few studies done on technology-assisted WCF among Arabic as Foreign Language learners. Among them is a study conducted by Abd Hamid et al. (2014) that studied the extent to which peer feedback through LMS can be used to support the pedagogical approach used by instructors. They found that there was an increase in the quality of students' writing, as well as a correlation between the number of words in the feedback given to the quality of subsequent writing. However, this study only examined feedback provided by peers, and not the teachers' corrective feedback on students' writing. While the above studies have highlighted the importance of corrective feedback, either traditionally or technology-assisted, in developing students' writing skills when learning a second or foreign language, very few studies have been conducted to identify the types of online corrective feedback among teachers and students, as well as their justification for such choices when teaching and learning a foreign language such as Arabic. Knowledge of their choices of feedback and its justifications is essential to designing the best approach to address students' weaknesses in Arabic writing. In addition, the knowledge of the effect of online corrective feedback on writing test scores is also important to determine the best pedagogy in learning Arabic writing. Thus, the objectives of this study are as follows: 1. To identify the types of written corrective feedback that instructors prefer in teaching Arabic writing using Google Doc. 2. To find out the instructors' justification for the preferred type of written corrective feedback used in teaching Arabic writing through Google Doc.
3. To identify the preferred types of written corrective feedback among students who learn Arabic writing using Google Doc. 4. To understand students' justifications for the preferred type of corrective feedback using Google Doc. 5. To examine post-test score differences between types of written correction feedback (direct corrective feedback with metalinguistic comments and indirect corrective feedback with metalinguistic comments).

METHODS Participants
Questionnaires were distributed to all instructors who teach the Arabic language through a blended learning approach (face-to-face and online) at an international Islamic university in Malaysia. It was also distributed to students who specialized in Arabic language to learn their views on the use of online WCF. The questionnaires were distributed to 93 registered first year students and four (4) instructors who taught the students in that particular semester. All instructors filled out the questionnaire with a response rate of 100%. Meanwhile, the students' response rate was 94.6% where 88 out of 93 students completed the questionnaire. All students are between the ages of 21 to 24 years old of which 70% are female and 30% are male. Of the four instructors who participated in this study, two are native speakers of Arabic while two are non-native speakers of Arabic. Two instructors and five students were selected to be interviewed in regard to justifications of the preferred types of WCF. The instructors selected taught Arabic language for Academic Writing during the study period, while the students were the first-year students, with three female students, and two male students. In order to determine whether there were differences in posttest scores based on the type of written correction feedback that students and instructors prefer, two groups were randomly selected and divided as shown in Table 1.
Both types of feedback were selected based on the type of correction feedback that instructors and students preferred the most. Google Doc is used to provide feedback to the students. The procedure using post-test is depicted in Figure 1 and Figure 2.

Research design
This study employed mixed-method design which allows qualitative and quantitative data to be collected in order to answer the research questions set for this study. There were two sets of questionnaires developed for the purpose of data collection for Objective 1 and Objective 3. The first set of questionnaire was developed for instructors and the other set was developed for the students. The items in both questionnaires are the same, with the exceptions that the items were written to suit the respondents' role (instructors vs. students). Likert scale was used in the questionnaires. The questionnaires were adapted from studies done by Amrhein and Nassaji (2010), and Sayyar and Zamanian (2015) with several additional questions to answer the objectives set. Prior to being distributed to the respondents of the study, the questionnaires were distributed to three Arabic language experts and design instructors to validate the items contained therein. Some corrections were made based on the feedback from the expert, and a pilot study was conducted where 25 students who took Arabic language (excluding the actual respondents) answered the questionnaire. The pilot study was to obtain Cronbach's Alpha Reliability coefficient, where it came back as satisfactory with a value of 0.79.
To answer Objective 2 and Objective 4, oneon-one interviews were conducted with two instructors and five students. The purpose of the interviews is to find out their justifications for type of WCF preferred and the role of Google Doc in connecting communication between instructors and students. The knowledge gained could help to design appropriate and effective pedagogical approaches in the delivery of WCF. As for Objective 5, a post-test was developed on Arabic language writing skills to assess the students' achievement. This writing test is divided into four sections: 1) content and sequence, 2) vocabulary, 3) grammar, 4) spelling. The test was reviewed and validated by three Arabic language experts. A scoring rubric was adapted from Jacobs et al. (1981) to determine the writing test score. 25 Arabic language learners involved with the pilot study, where 10 of them were interviewed to find out the clarity of the test instructions, and they said that the test was easy to understand and not confusing. At the end of the study period, the test was taken by the participants of this study. The correlation coefficients of post-test reliability for the elements of originality, consistency, flexibility and reliability of the instrument were 0.82, 0.79, 0.80 and 0.77 respectively at p < .005 levels. Based on the results of the above study, the Arabic language writing skill test is deemed appropriate and reliable to obtain a stable score from the respondents of the study.

Data analysis
In answering Objective 1 and Objective 3, the data obtained through Likert scale questions in which respondents were required to rate the type of corrective feedback on a scale of 1 to 5 depending on the benefit of its use. In this context, '1' means that corrective feedback is least beneficial to the students' writing skills, whereas '5' means that respondents think corrective feedback is very helpful for their writing. The mean and standard deviation were calculated to determine the value of each item.
The answers to Objective 2 and Objective 4 were obtained through qualitative thematic analysis using justification provided by instructors and students. For the first coding stage, the independent responses from the participants were used as the initial code. The purpose of this is to develop the researcher's understanding of the justification given by the participants. As for research question 5, the data was analyzed using one-way ANOVA to see the difference in post-test scores based on the type of WCF provided to the students.

RQ1 -The type of written corrective feedback (WCF) that instructors prefer in teaching Arabic language writing using Google Doc
The findings show that instructors prefer indirect WCF over direct WCF. Table 2 explains that instructors are more likely to choose indirect WCF with metalinguistic comments when providing feedback to students. RQ2 -Justification for the preferred WCF type among the instructors who teach Arabic language using Google Doc One of the justifications given in selecting type of WCF is the students' metacognition. The instructors interviewed mentioned that WCF "help students think and understand" and "help students to talk about mistakes with friends" hence WCF has been one of the approaches they practise to help students to improve their writing skills in Arabic language. Instructors in this study assert that student-centred learning is their primary focus, and their experience makes them believe that "indirect feedback is more effective" in giving feedback. The workloads and assignments are also the reasons why instructors prefer indirect feedback. This is because direct feedback requires a lot of time and high focus on students' writing. Comments such as "time constraints", "busy with administrative tasks" and "large numbers of students" are among the reasons why direct feedback is not preferred by the instructors. Nevertheless, instructors' preference of feedback also being influenced by the level of Arabic language proficiency among the students. The instructors commented that "choosing how to give feedback depends on students' proficiency" and "not all students can understand indirect feedback" give the impression that the choice of feedback type depends on the students' level of Arabic language proficiency. It also shows that the level of confidence among the instructors towards the level of students' proficiency affects the type of feedback used. The instructors' justification for choosing the type of WCF is as shown in Table 3. Table 4 shows students prefer direct WCF combined with metalinguistic comments. The findings suggest that students perceive direct feedback with metalinguistic comments help them to improve their writing skills in Arabic language.

Students' justification in selecting WCF via Google Doc
The preference of feedback type among the students differs from the type of feedback that the instructors prefer. Some of the justifications given by the students why they prefer direct feedback with metalinguistic comments are "I am informed of all my mistakes in writing", "I want to know my mistakes and the type of mistakes I made", "We need the instructor's guidance to correct all mistakes", "I am weak in Arabic writing". These statements show that students are lacking of selfconfidence when it comes to learning the Arabic language. One of the student-participant mentioned that "I don't know what I did wrong in writing" and this proves the fact that the students' limited language ability in Arabic language made them prefer teacher-centered lessons, as they value the instructors' feedback to improve their assignments. Moreover, indirect feedback with metalinguistic comments encourages them to reflect on the errors that they have committed, although this may require higher level of language ability, hence increasing the amount of work they have to complete. The student's justification for the preferred WCF type is as shown in Table 5. Differences in post-test scores between types of written correction feedback between direct WCF with metalinguistic comments and indirect WCF with metalinguistic comments Table 6 shows the difference in mean scores between Group 1 (indirect feedback with metalinguistic comments) and Group 2 (direct feedback with metalinguistic comments). Mean score for Group 1 is 81.92, while Group 2's mean score is 71.84. Meanwhile, the One-way ANOVA test results in Table 7 showed that the difference in writing scores between the two types of feedback was statistically significant [F = 46.353, p <.05].
The post-test scores indicate that students who received indirect feedback with metalinguistic comments achieved higher scores than students who received direct feedback with metalinguistic comments.

DISCUSSION
Among the objectives of this study are to study the preferred type of WCF among instructors, and the preferred type of WCF among students in the context of Arabic language writing classroom. Although this study was conducted on non-Arabic native speakers who study writing in Arabic language, the findings are consistent with the findings of the previous studies conducted on students learning English Language. Many previous studies have found that teachers give high value to indirect feedback that includes metalinguistic feedback (Eslami, 2014;Simard et al., 2015). However, students have different perceptions of indirect feedback. They felt that indirect feedback does not help them in improving the quality of their writing in Arabic language. This part will discuss the findings of this study in line with past studies on WCF.

Direct feedback vs indirect feedback
The findings confirm that instructors' and students' perceptions and justifications for direct and indirect feedback are different. Instructors prefer indirect feedback with metalinguistic comments while students prefer direct feedback with metalinguistic comments. Based on the justification of instructors and students, it is believed that they have different reasons for choosing different types of feedback.

Students' preference
Although students prefer feedback being provided electronically rather than face-to-face (Chen, 2016), their level of Arabic language proficiency makes them prefer direct feedback with metalinguistic comments. The use of technology enhances students' motivation to learn independently and actively (Helen, 2013), however, their limited language ability causes them to think that indirect WCF does not help them in completing the tasks given by their instructors. Online learning is also considered as fun learning compensation; however, the abundance of tasks will increase their workloads (Nur Agung et al., 2020). As a result, they value the accuracy and speed of the feedback given. Razali (2014) found that students are very concerned with the grammatical accuracy of their assignments, and that their work should be error-free. The studentparticipants of this present study are very concerned that their assignments are full of errors, and this would cause them receive low grades for the assignments. This is the reason why they would want feedback that is fast and easy to understand so that they can rectify the errors easily. This is in line with Hyland's (1998) claim that students prefer the easier option of relying on their teachers' feedback in achieving better grades. On the other hand, indirect feedback that gives clues without the correction does not help them to improve their writing. Moreover, indirect feedback with metalinguistic comments also requires students to be more active in their learning and encourages reflection on the mistakes made (Hamel et al., 2016). This results in increased workload and demands for higher level of Arabic Language ability for such reflection. This situation causes students to prefer direct feedback over indirect feedback.

Instructors' preference
Compared to the students, instructors have different perceptions about the types of feedback that they need to provide. Most instructors find that direct feedback with metalinguistic comments takes up longer time. This indicates that the strategy for selecting the type of written feedback depends on the instructors' workload. Therefore, the findings show that instructors value students' autonomy and expect them to play an active role in correcting their own mistakes. Instructors also view indirect WCF leads to self-correction that can benefit and help students to remember mistakes made (Amrhein & Nassaji, 2010). Likewise, metalinguistic approaches contribute to long-term metacognitive development and language acquisition (Ebadi, 2014). As a result, students' preference of direct WCF contradicts instructors' preference of indirect WCF, which requires students to work harder, and also promotes students' learning autonomy.
Moreover, all of the instructors mentioned that the best form of feedback depends on the context in which the feedback is given. Not only do they strive for student-centred learning, they would also need to consider students' motivation and the students' level of Arabic language proficiency, which would determine how far the feedback given could benefit the students in their learning. Because of these reasons, some instructors give feedback based on what they think the students would want, although this is not always the case. Furthermore, the instructors need to ensure that the type of errors made by students be stated clearly although they do not prefer direct WCF.

WCF through Google Doc platform
This study also aimed to examine whether there were differences in post-Arabic writing test scores between direct WCF with metalinguistic comments and indirect WCF with metalinguistic comments. Both types of feedback were provided using the Google Doc application as a learning platform. Students who received indirect written feedback with metalinguistic comments through Google Doc achieved higher scores than students who received direct written feedback with metalinguistic through Google Doc. The results of this study confirm previous studies which identify the effects of WCF through the use of technology (Seyyeedrezaie et al., 2017;Tabasi et al., 2013). Razali (2014) claims that students who received direct feedback may be able to correct the errors in the revised writing, but they may not be able to do self-correction in the new, subsequent writing due to the fact that direct feedback does not help students to think critically of the errors they commit. Razali further assert that students who received direct feedback may not understand the nature of the errors, hence have the tendency to repeat the same errors. This may be the case for the student participants of this present study where the students who received direct feedback were not able to critically analyse the errors they commit earlier, hence preventing them from producing writings that are error free. The results of this study are due to two factors, namely 1) the ability of indirect corrective feedback with metalinguistic comments to improve the quality of students' writing; and 2) an online learning environment that provides an opportunity for instructors to highlight more clearly the students' errors in their writing as well as giving comments to the students' writing. The combination of effective type of feedback, and the utilization of Google Doc application contribute in improving the quality of students' writing (Seyyeedrezaie et al., 2017). Indirect WCF with metalinguistic comments could help students to understand the errors they made (Ferris et al., 2000), while features that are available in writing collaboration applications, such as those in Google Doc, can play a role in facilitating and speeding up feedback, and this could lead to building knowledge on different dimensions (Salomon et al., 2003). It also promotes interactive language learning activities (Al-Olimat & AbuSeileek, 2015) and supports to improve students' achievement (AbuSeileek & Abu Sa'aleek, 2012).
Although this study found that the achievement score of the group receiving indirect feedback was better than the group receiving the direct feedback, it was not in line with the findings of the study done by Varnosfadrani and Basturkmen (2009) who found that the achievement score of the group which received direct feedback was better than the group which received indirect feedback. The difference in the findings of this study is that the teaching methods used may be different. The study done by Varnosfadrani and Basturkmen (2009) used limited face-to-face teaching mode to provide feedback, while this current study used Google Doc application that is accessible anytime and anywhere. The easy-to-use features of Google Doc give instructors an opportunity to interact more with their students outside of the classroom. In addition, the approaches used in the teaching of Arabic language are different in terms of language structure and grammar compared to English which results in different types of feedback given.
This study is not without its limitation. First of all, it must be acknowledged that the number of participants of this study is small, hence the findings of this present study do not reflect all contexts of learning Arabic language via Google Doc. Moreover, there are other factors that cannot be controlled by the researcher, such as social interaction that the students may have during the study. During the study period, the students may have communicated with the students from the controlled group, or other people who were not part of this study. They may have learnt from each other, and this may affect the results of the post-test that was done. Therefore, future research that are looking into the use of Google Doc as a means of delivering WCF within the context of teaching and learning Arabic language as Second and/or Foreign Language need to address these issues so that better results and findings could be yielded. Hence, this study proposes that blended mode learning should be used to ensure that feedback can be given more effectively. It is also suggested that feedback, be it direct or indirect, should be done together with oral feedback or student-teacher conference so that the students would understand the nature of the errors they have commit (Razali, 2014), hence helping them to learn the language better.

CONCLUSION
This study has highlighted the importance of WCF in improving the quality of students' writing in Arabic language. The knowledge gained from this study, i.e. the difference between the types of WCF preferred among instructors and students, provides more ideas to formulate appropriate approaches to teaching Arabic language to non-native speakers. The findings show that technology also plays a role in facilitating feedback. The use of Google Doc is seen as a means of enhancing interaction between instructors and students in improving Arabic writing. The findings also show that there are significant differences in post-test test scores between groups using direct feedback with metalinguistic comments and indirect responses with metalinguistic comments. The findings of this study may have pedagogical implications for Arabic language writing instructors as they choose the types of written corrective feedback (WCF) to be used in their teaching. It is advisable that instructors not use one-size-fits-all approach as different approach to WCF may make a difference between being a provider of the correct form or being an initiator who provides help through the feedback, but not giving out the correct form directly to the students.