ESL Teachers of Young Learners: What are their Perceptions and Practice of Written Corrective Feedback?

Written corrective feedback (WCF) has been used by teachers to help improve learners’ writing competency (Veren et al., 2020). While past studies focused on the effectiveness of WCF on learners, research from the teachers’ perspective is limited (Lee, 2020). This study fills in this gap by investigating the beliefs and practice of 11 primary school teachers’ WCF on grammatical mistakes using questionnaires and analysis of learners’ essays. The findings reveal congruence in the teachers’ beliefs and practice in providing direct feedback. However, discrepancies are found in the amount of feedback, whereby learners’ essays are mark ed comprehensively though most teachers view selective feedback as useful. The findings imply that there are underlying factors influencing teachers’ WCF decisions and practice which may be different from their beliefs. This study hopes to prompt teachers to reflect on their feedback provision while stakeholders are hoped to provide teachers with more autonomy in their teaching.


Introduction
According to Handayani (2017), writing is the most difficult but essential skill in language learning. One of the reasons is that writing skill requires learners to accurately put forward ideas using the knowledge of vocabulary, grammar and mechanics (Wulandari et al., 2019). Therefore, when it comes to young learners such as those in primary schools, they are bound to find writing a considerably challenging task (Imaniar, 2018) due to their low proficiency level (Gultekin & Nystrom, 2019).
Corrective feedback (CF) is commonly employed by language teachers to indicate the learners' mistakes and to further improve the learners' language competency (Sakanlai & Sukseemuang, 2021). CF can be given either orally or in the written form, but past research has found that the written form of CF, i.e., written corrective feedback (WCF) to be more useful particularly for low proficiency learners such as young learners, owing to the nature of the feedback being permanent and more noticeable (Aoyama, 2020). WCF can be categorised according to its scopes (comprehensive and selective) and its type (direct and indirect). The scopes of feedback refer to the amount of feedback which is given to learners (Rahimi, 2019). Comprehensive WCF gives feedback for all the mistakes in learners' writing, while selective WCF provides feedback on only a few selected types of mistakes (generally one to five types) in learners' writing (Razali et al., 2021). On the other hand, in WCF types, teacher may provide direct feedback by giving the correct answers to the learners' writing, or indirect feedback by only indicating where the mistakes are without providing the correct linguistic forms (Wong, 2021).
Since past studies discovered that the majority of young Malaysian learners' mistakes were grammatical mistakes (Harun & Abdullah, 2020;Liong et al., 2019), and grammar structures are crucial to help organise ideas into comprehensible sentences in writing (Fareed et al., 2016), there is a need for teachers to improve young learners' grammatical competency, be it through the teaching and learning process or the WCF given to indicate their grammatical mistakes (hereby referred to as grammar-focused written corrective feedback).
In past related studies on grammar-focused written corrective feedback (GWCF) in the primary school setting, more emphasis was given to the perceptions of young learners and the effects of feedback on them (Ogawa, 2017;Tursina et al., 2019), as compared to the teachers' perceptions of feedback (Gultekin & Nystrom, 2019). Nevertheless, Lee (2020) argues that it is equally important to focus on the teachers as they are the ones making the GWCF decisions for their learners. Moreover, learners at this stage are still highly dependent on their teachers for their learning. Thus, it is imperative to understand primary school teachers' perceptions and practice of WCF because they determine the kind of feedback received by the learners which have important consequences. For example, teachers who do not believe in giving feedback may ignore or tolerate their learner's errors in writing, and this could give the impression that the language structures used are acceptable thereby causing fossilisation of errors at secondary and tertiary education levels (Plaza, 2020;Prayatni, 2019;Shoaei & Kafipour, 2016).
As such, to address the research gap mentioned, the researchers conducted this pilot study to investigate teacher beliefs and practice of GWCF in the Malaysian primary ESL context. The research objectives were to: From the Activity Theory perspective, Storch explains that teacher GWCF provision could result from a network of connections between six major components (tool, subject, object, rule, community and division of labour). For instance, the feedback (tool) employed by the teachers (subjects) to improve the learners' grammatical accuracy (outcome) could be affected by other contributing factors such as past experiences, rules, community expectations, power relationships in schools, knowledge and beliefs (Lee, 2014). Borg (2001) defines belief as "a proposition that is accepted as true by an individual which serves as a guide to his thinking and actions" (p. 186). However, when looking into past research, most of these studies investigated teacher WCF beliefs in general at the secondary and tertiary education levels. The studies found that teachers generally believed that WCF should be given according to the learner proficiency level (Abdullah & Aziz, 2020;Hidayah et al., 2021). For instance, more proficient learners could be given indirect WCF, while lowproficient learners who need more guidance, could benefit from direct WCF. In terms of WCF scopes, some teachers perceived WCF could be given based on the predetermined learning goals (Gultekin & Nystrom, 2019). For example, teachers could give comprehensive WCF if they wish to inform the learners on the overall progress in writing, while selective WCF could be used if the teachers would like to improve their learners' accuracy in a particular grammar item (Mulati et al., 2020).
However, WCF studies that looked at teacher beliefs of young learners are more limited. To the best of the researchers' knowledge, there was only one study that looked at WCF in the primary school setting, and the context was in Sweden. Gultekin and Nystrom (2019) interviewed seven primary school teachers in Sweden to investigate how their WCF beliefs could promote ESL learners' writing development. It was found that most of the teachers agreed different WCF should be given based on the writing purposes (whether to achieve fluency or accuracy) and learner proficiency level, while positive reinforcement could be used to further motivate the learners to self-correct their errors in the future. However, the focus of the mentioned study was on general feedback, and there was no attempt to find out the teachers' perceptions on whether WCF worked best if it was accompanied by teachers' correction of the mistake or otherwise, or whether WCF should be given to all or only selected mistakes made by the learners.
As teacher beliefs can impact their classroom practice (Wei & Cao, 2020), some studies looked at the alignment between teachers' WCF perceptions and their WCF practice. Though there are some past WCF studies which showed good alignment between teacher beliefs and teacher practice (Yunus, 2020), the majority of the studies reported otherwise (Mahmud, 2016;Şakrak-Ekin & Balçıkanlı, 2019). Some of the possible reasons for the misalignment were class size (Abdullah & Aziz, 2020), time constraints (Aquino & Cuello, 2020) and heavy clerical workload in schools (Mahmud, 2016). Nonetheless, these past WCF studies were about teacher beliefs at secondary and tertiary education levels. Little is known about the alignment between primary school teacher WCF beliefs and practice (Rajagopal, 2015). Thus far, there is only a study which probed into primary school teachers' beliefs and practice of feedback. The study was conducted by Dessie and Sewagegn (2019) looking at the primary school teachers' beliefs and practice of feedback in Ethiopia. Through questionnaires, interviews and essay analysis, majority of the teachers were found to view feedback as essential in helping their learners, and that most of their feedback consisted of marks and general evaluative comments. However, the analysis of feedback practice was rather general, and there was no attempt to compare the different feedback practice among the teachers.
Owing to the dearth of studies on feedback in the primary school setting, it is apt that present and future research on feedback should focus more on the primary school setting. Teachers' beliefs and practice of feedback do not impact learners' learning only at secondary and tertiary levels, but also at the primary level. Furthermore, there are bound to be differences in the demands and expectations across educational levels, including teachers' perceptions and practice of feedback (Lira-Gonzales & Nassaji, 2020). As such, findings on feedback at the secondary and tertiary educational levels cannot be generalised to the primary school context. Hence, more research is needed to fill in this knowledge gap so that we are able to understand more about primary school teachers' beliefs and practice of feedback in greater detail (Irwin, 2017;Prawiro & Kholisna, 2020).

Research Design
This pilot study employed a mixed-method research design to collect both quantitative and qualitative data to understand an identified social issue (Johnson & Christensen, 2014). Questionnaire survey research was conducted to gather scaled data about the teachers' beliefs of GWCF while analysis of learner essays was carried out to obtain more information about the teachers' feedback practice.

Participants
For this study, the researchers invited Year 5 English language teachers (n=11) who were teaching in four national primary schools in Kuching, Sarawak, Malaysia through convenience sampling. The selected teachers represented a subset of the teachers of young ESL learners in the Malaysian national schools.

Instruments
Teachers' beliefs of how GWCF should be given were obtained using a questionnaire which was adapted from source questionnaires (Halimi, 2008;Lee, 2004;Sewagegn & Dessie, 2020) employed in past WCF studies. The questionnaire used in this study was observed to have a high degree of internal consistency with a Cronbach's alpha value of .852.
The questionnaire was divided into two parts, namely Section A and B. Section A, consisting of five items, aimed to obtain teachers' demographic data (gender, teaching experiences, highest educational qualification, etc.). In Section B, the teachers were required to indicate the level of agreement to the statements given (1 for strongly disagree to 5 for strongly agree) as well as to choose the option(s) that best describe their beliefs about the scopes and types of GWCF that should be given to their learners. There were seven items in the questionnaire which were targeted at obtaining information about the teachers' beliefs on the amount of feedback on grammatical errors that should be provided to their learners' writing, while 25 items in the questionnaire were designed to seek the teachers' beliefs on the different types of GWCF (e.g., direct feedback, indirect feedback) that should be given to the learners' writing, as well as the factors that influenced the teachers' beliefs on GWCF. There were also spaces provided for teachers to explain their selection of the option(s) in order to obtain an in-depth understanding of the teacher GWCF beliefs.
Information on teachers' practice of GWCF was gathered via a document analysis form in which the researchers recorded whether the teachers have marked learners' essays comprehensively or selectively (the scopes or amount of GWCF given to the learners), as well as whether feedback was given directly or indirectly (the type of GWCF provided to the learners) on the eight parts of speech outlined by Aarts and Haegeman (2021): noun, pronoun, verb, preposition, adverb, determiner and conjunction.

Data Collection
The entire data collection process lasted for a week. Prior to data collection, the researchers explained to the teachers about the research aim, objectives and the tasks that they were expected to do during the research. Then, copies of the teacher consent form were given to the teachers whereby they were required to read and sign the consent forms before participating in this study. After that, the teachers were given a day to complete and submit the filled questionnaire. Then, with the teachers' and the learners' permission, the researchers took pictures of all the essays collected (n=102). The essays collected were the first draft of the essays based on the latest Primary School Achievement Test (UPSR) Paper 2 Section C format.

Data Analysis
The quantitative data obtained from the questionnaire and the document analysis form were keyed into Statistical Package for the Social Sciences (SPSS) Version 22 to obtain mainly frequency statistics (raw counts and percentages). Frequency statistics was chosen as it was suitable to organise the participants' data collected from the discrete variables in this study (Larson, 2006).

Figure 1 Data analysis procedures
The data analysis procedures are shown in Figure 1. In the questionnaire, the scaled data collected for feedback scopes (comprehensive or selective feedback) were recoded (1 to 3 • The data collected from both sources were compared.

Questionnaire Document Analysis Form
• The scaled data collected for feedback scope were recoded. • The teacher beliefs on feedback types were identified. • Frequency counts were calculated.
• Each essay collected was checked for its scope of GWCF, i.e. comprehensive or selective feedback. • Each feedback given to learner's mistake was analysed for its type of GWCF, i.e. direct or indirect feedback.
• Frequency counts were were recoded as No, 4 to 5 were recoded as Yes). On the other hand, the teacher beliefs on the types of feedback (direct or indirect feedback) that should be given to the learners' grammatical mistakes according to the part of speech were first identified and then calculated for their frequency of responses.
As for examination of teachers' practice of GWCF, each essay collected was analysed whether GWCF was given comprehensively or selectively. Besides, in each essay collected, the researchers also calculated the total grammatical mistakes made by the learners, as well as whether GWCF was given directly or indirectly. Frequency counts of the scopes and types of feedback were calculated.
The data collected from the questionnaire responses and analysis of learner essays were later used to compare the teachers' feedback beliefs and practice. Section A of the questionnaire collected the teachers' demographic data (Table 1). Of the participants in this research, there were more female teachers (72.7%) than male teachers (27.3%). Besides, the majority of the teachers (36.4%) had teaching experience of more than 30 years. All the teachers (100%) held a Bachelor's degree, while most of them (90.9%) were English optionists. Apart from that, five teachers (45.5%) mentioned they had attended GWCF training, while six teachers (54.5%) mentioned they had never attended any GWCF training.  Table 2 summarises the teacher beliefs about the scopes of feedback. Overall, more teachers believed selective feedback should be given to the learners' writing (54.5%), as compared to comprehensive feedback (45.5%). In addition, the reasons for choosing selective feedback were to sustain "the learners' motivation to learn" (respondent 7), to motivate learners to "write better in the future" (respondent 7) and to not indicate repeated errors in learners' writing (respondent 6). On the other hand, one justification for giving comprehensive feedback reported by respondent 3 was "to improve the learners' writing" as a whole.

Findings Malaysian Primary ESL Teacher Beliefs of GWCF
Besides, more teachers (54.5%) believed that the learners' repeated errors should be given feedback. Four teachers (36.4%) mentioned that their feedback provision was prescribed by the schools, while the teachers' perception of their learners' needs (63.6%) was found to be the major factor that influenced the teachers' feedback provision.  Table 3, in terms of teachers' beliefs on the types of feedback (direct, indirect or both) based on the mistakes made on different parts of speech, the majority of the teachers (more than 70%) believed direct GWCF should be given.

Figure 2 Summary of Findings from DAF
As shown in Figure 2, of the 102 essays, all the essays (100.0%) were found to be marked comprehensively by the teachers, while none of the essays (0%) were given selective GWCF. Additionally, of the 102 essays, it was found that the learners made a total of 880 grammatical   Table 4, concerning young learners' grammatical mistakes, the learners received the most GWCF in verb (n=301), noun (n=140) and determiner (n=102) errors. Nevertheless, though the ratio of feedback types given to each part of speech varied, it can be seen that the teachers prominently marked the learners' grammatical mistakes directly.

The Comparison Between Malaysian Primary ESL Teacher Beliefs and the Practice of GWCF
The data collected from the questionnaire and learner essays were then compared. From the questionnaire, five teachers indicated that they preferred to give comprehensive GWCF, while six teachers preferred giving selective GWCF to the learners' grammatical mistakes. In the analysis of learner essays, it was found that all the 102 essays were given only comprehensive GWCF. The comparison suggested that there was a misalignment between teacher GWCF beliefs and practice in terms of scopes.
In terms of GWCF types, it was found that there was an alignment between the teachers' beliefs and practice of GWCF. The teachers viewed direct feedback should be given to learners' grammatical errors, and this was also reflected in the feedback given to learners' grammatical errors in their essays.

Malaysian Primary ESL Teacher Beliefs of GWCF
In GWCF scopes, more teachers preferred providing selective feedback to learners' writing, when compared to comprehensive feedback. However, the difference in the number of teachers choosing comprehensive or selective feedback was not significant, suggesting that both GWCF scopes were preferred by the teachers. A possible explanation for this belief could be the teachers' perceptions of the benefits learners could receive from each feedback scope, since seven teachers indicated that their choices of CF largely depended on their perceptions of the learners' needs. For example, five teachers believed they should correct all the learners' grammatical mistakes in writing (comprehensive feedback), and that they were likely to inform learners about their overall mistakes in writing which corresponds to Mulati et al.'s (2020) claim. In addition, six teachers preferred providing selective feedback to their learners, perhaps to avoid demotivating them and overloading their attentional capacity (Rahimi, 2019).
In terms of feedback types, the majority of the teachers believed they should indicate and provide answers to the learners' grammatical mistakes (direct feedback). This belief concurs with Abdullah and Aziz's (2020) argument that feedback could be given according to the learners' proficiency levels. For example, indirect feedback is believed to benefit the highproficient learners as the learners could be reminded of their mistakes and later make corrections independently when required (Chen, 2018). In contrast, Gultekin and Nystrom (2019) believe that direct GWCF could be more suitable for low-proficient learners as they might need more support in identifying their mistakes in writing. Similarly, in this study, perhaps most of the learners were of lower proficiencies since these were primary school learners (Aoyama, 2020), causing some teachers to provide answers for "the weaker students" (respondent 1) so they could "know the correct answers" to their mistakes (respondent 2).

Malaysian Primary ESL Teacher Practice of GWCF
Though all the teachers indicated that they have marked the essays comprehensively, a small proportion of the learners' mistakes (9.7%) were not given any feedback. This might be due to the repeated grammatical mistakes committed by the learners since nearly half of the teachers (n=5) indicated in the questionnaire that they would not mark repeated mistakes found in the learners' writing. However, the finding contradicts the study conducted by Gultekin and Nystrom (2019) whereby the primary school teachers would choose to correct their learners' repeated mistakes to prevent the learners from making similar mistakes in the future.
Besides, the teachers' knowledge of the language might also be a factor, hence leaving some mistakes in the learners' essays unmarked. In this study, a teacher indicated that she was not an English optionist, which could have affected the teacher's content knowledge when marking the learners' essays. In fact, this in line with Nemati et al.'s (2017) study when they found that some of the teachers may not have the required writing proficiency to assess their learners' writing. However, further investigation is needed to know the reasons behind the unmarked grammatical mistakes in the learners' writing. In terms of GWCF types, it was observed that the teachers mostly corrected the learners' grammatical mistakes directly. The finding contradicts studies conducted at tertiary (Aquino & Cuello, 2020) and secondary (Mahmud, 2016) education levels, where the learners' mistakes were marked indirectly. An explanation for the teachers to provide direct GWCF could be due to the teacher beliefs on the support needed by young learners, who possess limited work memory capacity (Aoyama, 2020). Therefore, the teachers' perception of the young learners' ability to analyse and to self-correct their mistakes may influence the way the teachers provide CF to the learners' grammatical mistakes (Mulati et al., 2020).

The Comparison Between Malaysian Primary ESL Teacher Beliefs and the Practice of GWCF
When the data collected from the questionnaires and analysis of essays were compared, misalignment was found between the teacher beliefs and practice of GWCF in terms of scopes. In this study, more teachers believed in the use of selective GWCF than comprehensive GWCF. However, teacher GWCF practice showed that all the essays collected were only given comprehensive GWCF. Furthermore, one could also relate the discrepancy between teacher GWCF beliefs and practice from the Activity Theory perspective, where the teachers' provision of feedback could result from a network of connections between six main factors, such as tool, subject, object, rule, community and division of labour (Storch, 2018). It is possible that the teachers had to follow the feedback scope pre-determined by the schools although it might not correspond to their beliefs (Lee, 2014), since four teachers indicated in the questionnaire that their provision of feedback to the learners' grammatical mistakes was prescribed by the schools. Nonetheless, in-depth interviews with relevant key stakeholders (school administrators and parents) could be useful in knowing the policies or expectations about the appropriate forms of feedback (Storch, 2018).
When looking into the feedback types given to the mistakes according to parts of speech, an alignment was found between the teacher GWCF beliefs and practice. It was indicated that the majority of the teachers believed they should provide direct GWCF to the learners' mistakes in writing, regardless of the parts of speech. Similarly, it was also revealed in the analysis of learner essays that the teachers prominently marked each part of speech directly. This could possibly indicate the teachers' strong belief that feedback accompanied by the teacher's correction is needed for young learners, and consequently, this is manifested behaviourally through feedback provision. Direct feedback is not only a quick and clear way to help resolve learners' complex errors (Budianto et al., 2020), but also especially necessary for learners of lower proficiency levels (Gultekin & Nystrom, 2019).

Conclusion
This study investigated the Malaysian primary ESL teacher beliefs, the practice of GWCF and the alignment between teacher GWCF beliefs and practice. The study revealed that more teachers believed they should provide selective feedback to the learners' grammatical mistakes, while the majority of the teachers believed that learners' grammatical mistakes should be marked directly. Besides, teacher feedback practice revealed that all the essays were marked comprehensively, while the majority of the teachers provided direct feedback on the learners' grammatical mistakes. When compared between the teacher feedback beliefs and practice, it was found that teacher beliefs misaligned with their practice in terms of scopes but aligned in feedback types.
Although the study has its limitations with regards to the small sample size and lack of qualitative data, it has contributed to the scant literature on feedback in the primary school setting, especially with regards to the teachers' beliefs and practice about the types and scopes of feedback that should be given to learners' errors. The study has also disclosed how teachers may have particular views about feedback and yet have different feedback practice, implying that perhaps there are other factors dictating teachers' feedback practice which warrants further investigation.
This study contributes to the limited studies on feedback in the primary school setting and aims to encourage ESL teachers to reflect on their current teaching practice, especially in terms of correcting learners' grammatical mistakes in writing. Moreover, this study further confirms that teachers' feedback provisions can be influenced by interconnected factors, such as school rules and expectations, as explained by Activity Theory (Lee, 2014). Therefore, it is suggested that key stakeholders grant teachers with more autonomy in their teaching, including how they provide GWCF to their learners. Moreover, this study contributes to the understanding of the most prominent types and scopes of GWCF provided in the Malaysian primary school settings. Thus, primary ESL teachers could experiment direct and comprehensive GWCF in their classrooms to maximise learner engagement with feedback. Discussions could also be conducted with learners to reach a consensus on the kind of feedback that should be provided in their writing. For future studies on feedback, having a bigger sample size and incorporating interviews in the study will increase the generalisability of the findings and depth of the study.