A comparative study of the engagement with written corrective feedback of Chinese private college students

Abstract

Previous research has explored how L2 students with one specific level of English proficiency engage with teacher written corrective feedback (WCF) underpinned by a tripartite dimensional construct of student engagement in the context of Chinese public universities. Yet, scant attention has been paid to how students of differing proficiency levels engage with teacher WCF in a Chinese private college context. Based on a quadripartite construct of student engagement, this case study has explored how L2 Chinese students with high proficiency (HP) and with low proficiency (LP) cognitively, affectively, behaviorally and agentically engage with teacher WCF. Data collected from multiple sources were examined, including drafts of student essays, teacher WCF, student immediate self-retrospective verbal reports and semi-structured interviews. The study has found imbalances among the four dimensions of engagement, with relatively high affective engagement and less extensive cognitive, behavioral and agentic engagement, which was mediated by the interplay of individual factors like language proficiency, writing self-efficacy and learner belief and contextual factors like student–teacher relationship. The findings contribute to an understanding of the multifaceted and dynamic nature of HP and LP students’ engagement with teacher WCF and provide some implications for both school administrators and teachers in Chinese private colleges.

Introduction

Engagement in any activity can have one or more of three forms, mainly including behavioral engagement, cognitive engagement, affective engagement and agentic engagement (Fredricks et al., 2004; Reeve, 2013; Reeve & Tseng, 2011). Written corrective feedback (WCF) mainly refers to error or grammar correction provided by instructors to improve students’ writing skills and accuracy (Bitchener & Ferris, 2012). Lee (2009) points out that since providing feedback to students’ errors is one of the pivotal elements in a writing class, teachers and students, as well as other stake-holders such as school administrators and curriculum explorers are all supportive and positive about WCF.

Although abundant research suggests that WCF is an effective strategy for improving L2 learners’ writing accuracy, Truscott (1996) posits that WCF may not be feasible and acceptable for all L2. Specifically, given that second language acquisition happens in a fixed sequence, WCF may only benefit highly proficient L2 learners and may be less effective for limitedly proficient L2 (Li & Vuono, 2019). Indeed, research demonstrates that while L2 students with average and advanced English proficiency effectively utilize WCF to improve their writing accuracy (Han & Hyland, 2015; Zhang, 2017), those with low English proficiency demonstrate difficulty using WCF to improve it (Zheng & Yu, 2018; Zheng et al., 2020). Although some research has examined how L2 students with intermediate language proficiency (e.g., Han & Hyland, 2015) and how L2 students with low proficiency (e.g., Zheng & Yu, 2018; Zheng et al., 2020) cognitively, affectively and behaviorally engage with teacher WCF in Chinese public universities, research has yet to explore in what ways L2 learners with high and low levels of English proficiency in a private college differentially engage with WCF under a quadripartite conceptualization construct. As suggested in Zheng and Yu (2018) and Han and Hyland (2015), future studies need to compare how participants of differing proficiency levels engage with teacher WCF. The exploration of how L2 learners with different levels of English proficiency engage with WCF in a private college can provide teachers as well as school administrators with some insights into their responses to the WCF. Based on this study, teachers can adjust the kinds or forms of WCF provided to the students and make it more effective and school administrators can reform the current curriculum arrangement to benefit the students. The findings of the study can provide pedagogical implications to enhance L2 learners’ engagement, improve their writing accuracy, and give some suggestions on instructions for private college English teachers as well as the school administrators.

In light of these considerations, this study has explored how low proficient (LP) L2 learners and high-proficient (HP) L2 learners engage with and utilize WCF provided by a foreign teacher of English in an English writing class. Specifically, the study aimed to answer this question: How do six LP and HP learners cognitively, affectively, behaviorally and agentically engage with teacher WCF?

Literature review

Before the discussion of the previous studies on how L2 students engage with written corrective feedback and the introduction to student engagement and its influencing factors, the controversies over the effectiveness of teacher written corrective feedback will be first included.

The controversies over the effectiveness of WCF

The controversies about whether or not the teacher WCF is helpful and effective to increase the writing accuracy of L2 learners were sparked by Truscott’s (1996) strong objection to the necessity and utility of written corrective feedback. Truscott (2007) pointed out that students were likely to use avoidance strategies such as shortening or simplifying their writings in an effort to avoid such situations as being corrected in his meta-analysis of the effect of error correction on L2 learners’ ability to write accurately. In other words, for some of the corrected students, the overall improvement of their writing scores is largely due to the fact that they have avoided using the structures or vocabularies of which they are less sure. However, other empirical studies have demonstrated the positive effects of WCF on the improvement of writing accuracy (e.g., Bitchner, 2008; Bitchner & Knoch, 2010; Bitchener & Storch, 2016; Lee, 2019; Sheen et al., 2009). By meta-analyzing 21 empirical studies, Kang and Han (2015) found that WCF could result in greater grammatical accuracy in second language writing, yet its effect was mediated by many variables, including learners’ proficiency, the types of WCF as well as the genre of the writing task. Similarly, in a meta-analysis, Abalkheel and Brandenburg (2020) explored ten quasi-experimental studies on the effects of WCF and found that WCF in general was a positive predictor of student improvement in writing and that direct and focused WCF was especially more effective and helpful than indirect and comprehensive one.

Student engagement and influencing factors

In the field of English as a foreign language (EFL) and second language acquisition (SLA), students’ engagement has long been regarded as an overarching concept, bringing together students’ attention, curiosity, and willingness to exploit their language proficiency and learning strategies to make progress (Zhang & Hyland, 2018). Inspired by the tripartite conceptualization of student engagement, specifically, cognitive, affective and behavioral engagement, proposed by Fredricks et al. (2004), Ellis (2010) applied it into the study of corrective feedback and made some adjustments. He has defined cognitive engagement, affective engagement, and behavioral engagement as how students cognitively respond to feedback, how they attitudinally attend to feedback, and how and whether they revise their texts in response to feedback, respectively. Acknowledging the tripartite conceptualization construct, Reeve and Tseng (2011) proposed that agentic engagement could serve as the fourth dimension, which captures students’ intention and proactivity in trying to personalize or enrich what to be learned as well as the circumstances under which it is to be learned.

Students’ engagement with WCF can never be consistent, which has been found to be dynamic and vary across individuals instead, mediated by individual and contextual factors (Ellis, 2010; Han, 2019; Zhang & Hyland, 2018; Zheng et al., 2020). Individual factors affecting students’ engagement with WCF include students’ belief and personal experience (e.g., Han, 2017; Han & Hyland, 2015; Storch & Wigglesworth, 2010; Zheng et al., 2020), learning goals (e.g., Hyland, 2003), students’ motivation (e.g., Fredricks et al., 2004; Goldstein, 2006); language proficiency (Han & Hyland, 2015; Lee, 2008; Qi & Lapkin, 2001; Zheng & Yu, 2018), writing self-efficacy (Tsao, 2021), and foreign language enjoyment (Zhang et al., 2021). For example, learner’s beliefs about the language use may have a negative influence on L2 writer’s affective engagement with WCF (Storch & Wigglesworth, 2010). Specifically, students who focus more on the content of their writing, rather than on its linguistic accuracy, are more reluctant to accept the form-focused WCF (Hyland, 2003). Han (2017) found in the empirical study that student engagement with WCF can be directly and indirectly mediated by learner belief, with the direct impact including guiding students’ selection of learning strategies, external resources, and revision operations and the indirect impact of lowering negative self-concept students’ own expectation to their writing performance (Zheng et al., 2020). Contextual factors, such as characteristics of WCF (e.g., Bitchener & Ferris, 2012), teacher-student relationships (e.g., Dewaele & MacIntyre, 2019; Martin & Dowson, 2009), and instructional approaches (Wang & Holcombe, 2010) can also impact students’ engagement with WCF. For instance, Dewaele and MacIntyre (2019) found that a good relationship with and a positive attitude for the teacher can generate more foreign language enjoyment, making L2 students get more engaged with WCF (Zhang et al., 2021).

Empirical studies on student engagement with WCF

Although researchers have apportioned greater weight to explore student engagement with WCF from a multi-dimensional perspective, related studies are still scant and “in its infancy” (Zheng & Yu, 2018). Specifically, Han and Hyland (2015), focusing on four non-English major students with intermediate English proficiency, found that students’ engagement with WCF varied, partly caused by their beliefs, experiences, learning goals, and interactional context. Zheng and Yu (2018) also conducted a qualitative research on the engagement of twelve L2 students with lower English proficiency on teacher WCF and found that students’ cognitive and behavioral engagement could be negatively influenced by their lower English proficiency, causing imbalance among the three dimensions of engagement. Zheng et al. (2020) explored two low-proficiency L2 students’ engagement with teacher WCF and why they (dis)engage in the ways they do. It was found that both of the students’ engagement was distinctively different because of their different beliefs, goals, and teacher-student relationship.

It is worth noting that each of the aforementioned studies focused on students with one specific language proficiency, specifically, intermediate language proficiency (e.g., Han & Hyland, 2015), and lower English proficiency (e.g., Zheng & Yu, 2018; Zheng et al., 2020). Considering “the richness and complexity of classroom life” (Guerrettaz & Johnston, 2013, p. 782), it is never possible to have only one group of students with one specific English proficiency in a real teaching context. It is more common to have students with various levels of English proficiency instead, meaning more voices from students with different English proficiency should be heard. Moreover, the extant studies used the three-dimensional construct proposed by Ellis (2010) and paid no attention to how students agentically engaged with WCF, the fourth dimension proposed by Reeve and Tseng (2011), which asserts the significance of students’ agency in taking initiatives to enrich and modify learning activities “rather than merely reacting to them as a given” (p. 258).

To narrow the research gaps, this study following a case-study approach to explore how students with high English proficiency and low proficiency engaged with teacher WCF cognitively, affectively, behaviorally and agentically in a Chinese private college. The case study approach was selected, considering the complexity of the four dimensions in students’ engagement as well as the meditating individual and contextual factors, which cannot be captured in quantitative methodology that can “single out one factors as main source” (Lee, 2008, p. 157).

Theoretical framework

Based on the previous research (e.g., Ellis, 2010; Finn & Zimmer, 2012; Han & Hyland, 2015; Martin & Rose, 2002; Reeve & Tseng, 2011; Zhang, 2017; Zheng & Yu, 2018), we fine-tuned a theoretical framework for studying student engagement with teacher WCF, which views student engagement with teacher WCF as a multi-faceted system, involving the student’s cognitive, behavioral, affective and agentic response to feedback.

In specific, cognitive engagement is manifested in the depth of processing of WCF, meta-cognitive and cognitive operation used to process WCF and make revisions (Han & Hyland, 2015; Zheng & Yu, 2018). The depth of processing refers to how deep the L2 writers can process the WCF, for instance, they can notice errors they made and can give the reasons, or notice errors and fail to give the reasons, representing different quality of noticing (Qi & Lapkin, 2001). Cognitive operation refers to L2 writers use cognitive strategies, such as making mental notes (Ferris et al., 2013), memorization and visualization (Storch & Wigglesworth, 2010) while dealing with WCF. Meta-cognitive operation means how L2 writers monitor and regulate their mental effort to process WCF, for instance, evaluating the effectiveness of WCF and deciding whether to accept that WCF in their future writing (Ferris et al., 2013; Storch & Wigglesworth, 2010).

In terms of the affective engagement, it emphasizes more on the student’s attitudinal responses toward WCF (Ellis, 2010), consisting of affect, judgement and appreciation (Martin & Rose, 2002). Zheng et al. (2020) further explain that affect can be judged as students’ feelings and emotions expressed when they receive WCF and the changes in these feelings and emotions in the process of revising text; judgment as student’s personal judgements of admiration or criticism and moral judgments of praise or condemnation towards WCF; appreciation as valuing the worth of WCF.

Behavioral engagement mainly revolves around whether L2 writers revise their writing after receiving WCF and what strategies they take to avoid future errors in an attempt to improve their writing (Han & Hyland, 2015). Common strategies that facilitate the processing of WCF to improve future writing include keeping an error book (Hyland, 2003), and checking a dictionary or seeking teacher’s explanation (Han & Hyland, 2015).

Agentic engagement was defined as students’ intention or direct attempts to “enrich and improve both what is to be learned and the conditions under which it is to be learned” (Reeve & Tseng, 2011, p. 258), conceptually and statistically different from the other three counterparts. Agentic engagement referred to students’ proactive contribution to the flow of instruction both to improve their own learning together with learning condition and to ask for interpersonal support to get motivated in task-related learning activities (Michou et al., 2021; Reeve, 2013; Reeve & Shin, 2020). It can be represented by students’ direct communication with the teacher about their preferences and suggestions for teacher WCF upon receiving the it during the semester. Based on the previous relevant theories and research literatures, the theoretical framework for understanding student engagement with teacher WCF can be illustrated in Fig. 1.

Fig. 1

The theoretical framework for student engagement with teacher WCF

Methodology

The study took a multiple-case approach to explore how six individual L2 writers with either high proficiency or low proficiency engage with teacher WCF, cognitively, behaviorally, affectively, and agentically within an authentic context of a Chinese private college.

Context and participants

The study took place in a private college, located in eastern China. Generally, the operation of private colleges in China mainly relies on tuition fees and funding from enterprises, social groups or individuals, barely getting financial support from the government (Su, 2012). In other words, the investment in private colleges is expected to gain reasonable economic returns (Li & Morgan, 2008), which is the same to the college in this study. Upon the students entering the college, the freshmen choosing the same major are grouped into administrative classes based on their overall score of the College Entrance Examination. To save cost, generally, up to 35 students are assigned into an administrative class. Since the passing rate of the College English Test (CET) Band 4 is an important index to demonstrate the schools’ quality of teaching, which in turn influences the school reputation, the newly recruited students are required to take a compulsory English proficiency test organized by the school to get the overall knowledge of their English level. To ensure a high passing rate of CET and improve students’ overall English proficiency, a rich English curriculum system has been set up, including courses like Integrated English I-IV, English writing, public speaking, and the like.

The study was conducted in a sophomore English class which was taught by a native speaker of English, Linda (a pseudonym), who has the experience of teaching writing in other Chinese universities for three years and in this college for eight semesters. Considering the fact that Linda is a native English speaker having a Master’s Degree in TESOL and her previous experience of teaching English courses in China, the researchers did not provide any training to her. In her previous writing classes, students followed a feedback-revision cycle, beginning from composing a draft, to receiving teacher WCF, and ending with completing a revised draft. In this study, students majoring in English education with a wide range of English proficiency from the 2021 cohort were targeted.

Purposeful sampling (Creswell & Creswell, 2017) strategies were used to select student participants. According to the score of English proficiency test which the students had taken after entering the school, as well as their English scores of the college entrance examination, three voluntary students with high proficiency and another three with low proficiency from the same class were recruited as typical cases.

Data collection

Data triangulation was utilized in the research design. Data sources included students’ first and final drafts, teacher WCF, student immediate oral reports, and student interviews (See Appendix 1). With the help of the first and final drafts, students’ revisions were coded in order to address their behavioral engagement with WCF. With regard to the immediate oral reports and the final interviews, they could provide a more deep and comprehensive understanding of students’ cognitive, affective, behavioral, and agentic engagement with teacher WCF. The students generated the self-retrospective oral report after receiving teacher WCF in their first draft. Specifically, after reading the feedback alone for the first time, the participants were required to self-record their oral report in Chinese and then send the recordings to the first researcher through email. Considering that if the participants were requested to provide an oral report no less than some minutes, their attention would be diverted to how long they had recorded, therefore, they were encouraged to express their feelings and thoughts in detail as much as possible with no time limitation. Semi-structured interviews were conducted at the end of the semester, each lasting 40–60 min. The photocopied first and final drafts were collected at the end of the semester to explore students’ behavioral and cognitive engagement. The data collection procedure is illustrated in Fig. 2.

Fig. 2

Data collection procedures

Data analysis

The date analysis consisted of three parts, namely, the text analysis of the participants’ two drafts, the types of WCF Linda provided, and the qualitative analysis of transcriptions of interviews and oral reports. First, linguistic errors in the first drafts were identified and categorized according to Ferris’s (2006) taxonomy with minor adaptations (See Appendix B). A measure of errors per 100 words in the first and revised drafts was calculated to indicate the success of student revisions (Chandler, 2003) which could indicate their behavioral and cognitive engagement.

To answer the research question, the qualitative analysis of transcriptions of interviews and oral reports was carried out, with the theoretical framework in Fig. 1 acting as a scheme for coding. The recordings of the oral reports and the interviews were manually transcribed and checked by the participants for accuracy. Deductive coding was used under the guidance of the theoretical framework (See Fig. 1), specifically four first-level codes and ten second-level codes were created to examine the transcripts (See Appendix 3).

Results

Cognitive engagement

How L2 students cognitively engage with WCF incorporates the awareness of the WCF (noticing and understanding), metacognitive operations used to regulate their mental processes and practices to the WCF, as well as the cognitive operations to process and respond to the WCF. In terms of the primary level of awareness—noticing, whether or not the students can recognize the teacher’s corrective intention depends on the types (direct or indirect) of WCF provided by the teacher, irrelevant to their language proficiency. Judging from the students’ first draft, Linda provided the indirect WCF to two of the participants, Xu (HP) and Luo (LP), while for the rest of the participants, direct WCF was offered. Both Xu and Luo reported their confusion about what Linda wanted to convey in the interview. As mentioned by Xu, “To some of the WCF, I can’t understand the teacher’s intention. For example, she put a question marker beside a sentence. I don’t know whether this sentence is grammatically wrong or unclear in meaning”.

Regarding the deeper level of awareness (specifically, understanding), manifesting the extent to which the students can diagnose the error and is able to provide accurate metalinguistic explanations, all of the students in the HP group had no difficulty understanding the feedback on sentence patterns and grammatical errors. For instance, Cao (HP) mentioned that he used the article “the” wrongly in his draft and Linda corrected all the wrong use of it. In the interview, he provided accurate metalinguistic explanations for why “the” should be used.

When using the structure “the more…the more…”, I omitted the second “the”. In the expression like “people beside you”, since “people” is modified by the preposition phrase “beside you”, there is supposed to be a “the” before people, but I didn’t use it.

On the contrary, most of the LP students couldn’t fully understand the feedback on grammatical errors even though they could notice the WCF. When asked to give a metalinguistic explanation for why the sentence should be corrected in that way (as Linda corrected), Luo (LP) first kept silence for a while and then said she would ask her teacher for explanation later. The same situation happened to both Jin and Lu. When asked the same question, instead of giving a metalinguistic explanation, they answered in a more general way, attributing their errors to their low proficiency in English grammar.

Interestingly, all of the students had difficulty understanding the WCF on word choice errors upon receiving the feedback even though Linda had provided the correct word or phrase aside. However, the two groups of students reacted differently to the feedback. For the HP students, they were all curious about why the word or phrase they used in their first draft was incorrect or unacceptable. Taking Li’s response as an example, she insisted in her self-retrospective verbal report that the expression like “a qualified Internet user” was correct if it was translated into Chinese and she was wondering whether she could keep her original version instead of correcting it into “a responsible Internet user” as suggested by Linda. After getting the explanation from Linda, she figured it out and had it revised. On the contrary, Lu (LP) said she could totally understand it but when asked to elaborate on why her original version was incorrect, she said, “Since Linda has corrected it, it means my version was wrong…”. Similarly, both Lu and Jin failed to clarify why their original word choice was incorrect when asked to give further explanation even if they said they could totally understand the feedback on word choice at first.

The conduction of meta-cognitive operations seemed to be limited in both groups. One common operation used in the two groups was that all the participants self-reflected on why they made these errors in their first draft. They mainly attributed the errors to their low proficiency in English grammar and their unsophisticated command of vocabulary. Xu (HP) mentioned that he knew it very well that he was poor at advanced English grammar, and that it was not uncommon for him to make grammatical errors in English compositions. Similarly, Luo (LP) put it simple that her English level was low and that she was weak in English grammar. She mentioned that even though compound sentences had been taught in high school, she still could not use them correctly in her writing and thus tried to avoid using them. Unlike students in the LP group only using self-reflection strategy, two HP students (Li and Xu) tried to link their errors to the meta-linguistic rules and categorize the errors. For instance, Li (HP) mentioned the follow words in the interview: “From the teacher WCF, I find that I am relatively poor at using transitional words to connect two sentences”. It could be seen that self-reflection was a common meta-cognitive operation to the participants from both groups and that most of the HP students would think deeper, cognizant of categorizing their mistakes.

In terms of cognitive operations, the general trend was that they were used limitedly, with the HP students using relatively more complicated ones than the LP students. For the latter, they just automatically followed the feedback and revised their first draft literally according to it. As mentioned by Luo (LP), “Linda is a native speaker. All I need to do is to accept all her corrections. And it is quite easy for me to correct the first draft since she has corrected most of the errors”. Likewise, Jin (LP) voiced the same opinion and said that it was beyond her capacity to figure it out why her original sentence was grammatically incorrect. To save time and efforts, totally correcting the errors literately according to the feedback is a favorable strategy for her. Contrary to the students in the LP group, those from the HP group reasoned why the grammatically errors should be corrected in that way. The following sentence was reported by Xu (HP) in the interview:

In the sentence of “…they want to relax after overloading work or study”, the subject is “they”, so the subject of the part behind “after” is also “they”. In this sense, “overload” is supposed to be used in the passive voice instead of the active voice, which should be corrected into ‘after being overloaded by…”.

Affective engagement

Students’ affective engagement can be manifested through their affect, judgement and appreciation. Affect is connected with feelings and emotions expressed upon receiving WCF and changes in these feelings and emotions in the process of revision. Li (HP) felt very shocked upon receiving her first draft with the WCF on it. She mentioned that she thought this composition was well written and that it was unbelievable for her to receive so many “red pens”. In terms of Cao (HP), he felt it was normal to have so much WCF on his first draft because he held that his English writing skills were always not that good, attributing to his low proficiency in advanced English grammar. He also reported worry about whether his bad performance in this composition would influence his final grade of this course. Xu (HP) expressed complicated feelings: disappointed, worried, and happy in the interview:

Upon seeing so many corrections in red, I felt very disappointed. One of the reasons is that it's really beyond my anticipation that I made so many grammatical errors in this composition. Besides disappointment, I also felt a bit worried about the possibility that so much WCF might influence my final grade. However, when I revised the draft, I also felt happy to some extent because the more she corrected for me, the more likely it is for me to avoid the similar errors in the future, which is really a good opportunity to learn and improve the accuracy of my writing.

Contrary to the students from the HP group expressing their negative emotions and feelings, two LP students expressed some positive feelings. Jin voiced that she was deeply moved by the teacher for she had provided so much WCF to her in great detail. She reasoned that since the teacher was teaching three classes, meaning she had to grade approximately 100 copies of this composition and provide detailed feedback to each student, it must be very laborious. She was so deeply moved by the time and efforts spent by the teacher that she made up her mind to work harder in this course. Similarly, Lu appreciated the teacher very much for giving so much feedback in detail. However, Luo (LP) voiced her frustration upon receiving the feedback and remarked in the interview:

I clearly know that my English is not as good as others. It’s very depressing to see so many “red pens” in the passage, reminding me of the number of errors I have made. It seems that I get nothing except so much feedback, a kind of humiliation to me, after several hours of labor. I would appreciate her more if she could also give me an overall comment, especially on the good points in my work.

Another dimension underlying students’ affective engagement is their judgment of admiration or criticism towards WCF. It was obvious that all of the students appreciated and accepted the feedback on grammatical errors, four of whom (Cao and Xu from the HP group and Lu and Luo from the LP group) emphasized the teacher’s identity as a native speaker. Li (HP) mentioned that the corrected sentence by the teacher was much clearer and more advanced than her original version. However, their judgement of the feedback on word choice errors differed. Both Li and Xu from the HP group expressed their skeptical attitude. As Li remarked in the self-retrospective verbal report: “I think ‘qualified’ is the word that I want to use, which is appropriately used here. There is no need for her to replace it with ‘responsible’”. In the final interview, she mentioned that she finally figured out the inappropriateness in using this word after consulting the teacher.

For Lu and Jin from the LP group, they just totally accepted the feedback on the word choice errors since it was offered by the teacher, who was a native speaker and professional in their eyes. On the contrary, Luo (LP) voiced a similar opinion to the students from the HP group that the word choice was more of a personal preference and that feedback on word choice was useless and unnecessary. As she remarked in the interview:

It is useless for the teacher to give the WCF on word choices, because they are something fixed in my prior knowledge and something of my personal preference. It is hard to correct their wrong use with once for all. Perhaps, for the next time, I will make the same mistakes.

Appreciation is the third sub-dimension, pertaining to students’ valuing the worth of WCF. All of the students straightforwardly expressed their recognition of and admiration for the teacher’s feedback. Li (HP) and all the students from the LP group expressed their inclination for direct feedback from the teacher. Li’s remark in the interview seemed to best explain their inclination for the feedback of this kind: “The explicit and direct feedback can reduce the learning difficulty for me. If she gives me indirect feedback and I fail to figure out why the marked part is incorrect, the feedback will be useless”.

Unlike these four students, two HP students (Cao and Xu) expressed their preference for mixed forms of WCF and indirect form of WCF, respectively. Cao mentioned that for the errors containing simple grammar points (like tense and aspect), he preferred indirect feedback (e.g., underlining or circling), which could give him some room for deep thinking, while for the errors containing advanced grammar points, direct form of feedback (e.g., revision) was more welcomed, which could save him time and efforts. In terms of Xu, indirect feedback was more acceptable for her since it could inspire her to think thoroughly.

Behavioral engagement

Both revision operations and observable learning strategies to improve the overall writing skills constitute the students’ behavioral engagement with WCF. The changes in error rates (as shown in Table 1) between the first draft and the revised version could serve as a good indication of their revision operations, while self-retrospection report and interview could be effective to show their learning strategies. Based on the error types adapted from Ferris (2006) in Appendix 1, the first and second drafts were marked and analyzed for errors. Even if the students were requested to write a composition of no less than 250 words, three students wrote far more words than required. Cao, Jin and Luo wrote 456, 484, and 310 words respectively. Not yielding texts of exactly as required is a common phenomenon (Zheng & Yu, 2018). To address the unevenness in the number of words, the error rate of the students’ first draft and the revised version were calculated based on errors per hundred words by following the previous research (e.g., Han & Hyland, 2015; Zheng et al., 2020).

Table 1 Summary of error rates, WCF, and revision

It could be concluded from Table 1 that the revision rates of all the students from the HP and of one student (Lu) from the LP group were higher than 80%, while two students (Jin and Luo) from the LP group had relatively lower revision rates than their peers (64.29% and 74.42%, respectively). The textual examination indicated that Linda solely provided direct feedback to Jin and Lu, and that they didn’t attend to it with heart and soul. For instance, in the case of Jin, the teacher corrected her original wrong sentence “The usage of Internet” into “The usage of the Internet”, but Jin kept her original sentence in the revised draft. Similar to Jin, Lu kept her original wrong sentence “Whether the blue light…” instead of revising it into “Whether it’s the blue light…” as corrected by the teacher. Unlike Jin and Lu, Luo was provided with some indirect feedback to some unknown reason. She carefully revised the errors literally according to the direct feedback and attempted to revise the errors with the indirect feedback.

Similar to the cases of Jin and Lu from the LP group, Xu (HP) failed to attend to the feedback wholeheartedly even if she successfully self-revised the errors with the indirect feedback. Linda explicitly added a “their” between “enrich” and “spiritual life” and deleted “so much” in the sentence “People want to travel so much”, but Xu kept her original versions. On the contrary, the other two HP students, Li and Cao treated the feedback more carefully and seriously. Li revised all the errors according to the teacher’s feedback and Cao still had an error in his revised version. When examining Cao’s first and revised drafts, it could be indicated that he mistook “separating” into “sepirating” due to the teacher’s bad handwriting and kept “sepirating” in his revised version (Fig. 3). Generally, even if direct feedback was widely and commonly offered to the students, lowering the difficulty in the students’ deep processing of WCF and in their revision, half of the students failed to carefully and seriously attend to the feedback compared with the rest.

Fig. 3

An original example of teacher WCF to Cao

Learning strategies used to improve their own writing skills and English language competence were various among the students. Students from the HP group and Lu from the LP group firstly looked up the dictionaries and consulted the Internet to figure out why the words or phrases had received WCF. They wouldn’t consult the teacher unless they could not find out the reasons for their errors on their own. As Lu (LP) remarked in the interview, the process of looking for the reasons was beneficial to deepen her understanding as well as to help her memorize the language points. Li contended that she asked for help from the teacher about the “qualified” used in her first draft after she failed to find an appropriate explanation in the dictionary and on the Internet. She finally accepted the teacher’s explanation and then corrected “qualified” into “responsible”. On the contrary, instead of depending on themselves, Jin and Luo, the other two students from the HP directly asked for explanation from their peers and the teacher. To avoid making the same errors in the future, HP students and Lu (LP) mentioned that during the process of looking up the dictionaries and consulting the Internet, they would make a mental note about the errors and then memorize them after figuring out the errors. Jin mentioned that she kept an error book for future review.

Agentic engagement

Agentic engagement is the fourth dimension of student’ engagement, revolving around their attempts to enrich the conditions and circumstances under which the WCF was made to be more helpful and beneficial. It can be represented by students’ direct communication about WCF or suggestions for providing WCF with the teacher during the course. Among all the students, two HP students (Li and Xu) mentioned their experience of directly expressing how to make the WCF clearer and more helpful to the teacher. As recalled by Li in the interview, an active learner, she firstly went to the teacher and asked for an explanation for the use of “qualified”, and expressed the similar situations encountered by other students in class: even if direct feedback on the word choice errors had been given, most of the students were still unclear about why their original words were not acceptable. She then suggested the teacher sparing some time to have a face-to-face question-and-answer session in class. Luckily, the teacher took her suggestion and spared one class time for the students to have an in-person communication in the following class after they received the feedback. As Lu (LP) mentioned in the interview:

I am very thankful for such a change in this course. Talking directly with the teacher about the feedback really benefited me a lot. She not only told me why the word I chose was not correct but also elaborated on the grammatical points that I had not mastered. Compared with just offering me the WCF, this form of communication really helps!

Similar to Li, Xu also went to the teacher and suggested that it would be better to have a peer group discussion on explaining the feedback they had received, because this form of group study was found effective in their Integrated English I & II courses in the previous academic year. She reasoned in the interview that since they were all English Education major students with a higher possibility to be an elementary school English teacher in the future, it would be a good chance for them to practice teaching English writing. However, even if the teacher praised Xu for her good idea, she, in the words of Xu, rejected this suggestion for it would be much more time-consuming if some of the group members failed to give a correct explanation for the feedback in such a tight teaching schedule. Unfortunately, regarding to the rest of the students, they didn’t express their preferences for and needs in teacher WCF in the whole semester, albeit they mentioned that they provided some suggestions in the course evaluation at the end of the semester. When asked why, Cao (HP) voiced his concern that communicating his preferences in front of the teacher had the possibility of influencing his final grades, since this kind of behavior, as far as he was concerned, was a way to challenge the authority of the teacher. With respect to two of the LP students (Jin and Luo), even though they didn’t expect a good final grade, they also expressed the concern for incurring some troubles if they provided some suggestions to the teacher. Luo’ s comment in the interview seemed to sum up their opinions: “All I want from the course is that I can pass it and then learn something if possible. I don’t want to make some troubles to the teacher”.

Discussion

This multiple-case study has uncovered the complexity of how students with high English proficiency and with low English proficiency engaged cognitively, affectively, behaviorally and agentically with teacher WCF.

In terms of the cognitive engagement, all of the students provided with direct WCF could notice the feedback and figure out the teacher’s intention and purpose, while Xu (HP) and Luo (LP) could not get the intention of the indirect feedback and felt confused, since indirect WCF could sometimes cause confusion and direct feedback was more direct and effective (Bitchener & Ferris, 2012). As to the understanding of teacher WCF, HP students could understand the feedback on grammatical errors, each of whom could provide correct meta-linguistic explanations for the errors they made while all the LP students could have some difficulty understanding the feedback, manifested by their failure in providing meta-linguistic explanations to the errors involving intermediate and advanced grammar points. This finding corroborated Lee’s (2008) argument that language proficiency played a crucial role in students’ understanding of the WCF and that LP students, lacking adequate linguistic competence, could not thoroughly understand the feedback. With regard to the understanding of feedback on word choice errors, all students, regardless of their language proficiency, could not adequately understand it, even if the correct word was provided. As Zheng and Yu (2018) has suggested, students tended to use literal Chinese—English translation in writing and paid less attention to its appropriateness and correctness, indicating their immature command of vocabulary and the context in which it should be used. Meanwhile, students’ meta-cognitive and cognitive operations were quite limited. All the students used the meta-cognitive strategy of self-reflection to find out why they made so many errors in the composition. Compared with LP students, HP students used another strategy of categorizing their errors. Similarly, HP students used reasoning, a more complicated cognitive strategy, to process the WCF compared with LP students who took superficial strategies like automatically accepting the WCF instead of probing into grammatical rules behind their errors. Apparently, LP students had a relatively low writing self-efficacy, who attributed all their errors to their low English proficiency, which in turn made them deploy limited cognitive strategies to process the WCF. It corresponds to Tsao’s (2021) argument that without (a high level of) writing self-efficacy, students might fail to respond to and process teacher WCF.

From the affective perspective, the inconsistency between the errors made by the HP students (Li and Xu) and the teacher WCF triggered their epistemic affects such as shock and disappointment (Han & Hyland, 2019), since both of them originally had a very high level of writing self-efficacy (Tsao, 2021) and positive self-belief for writing (Han, 2017), which was lowered by the reality of so many errors. Outcome affect like worry was evoked in all three HP students because they perceived an unexpected failure to achieve their writing goal (Han & Hyland, 2019) – to get high points in this writing task. With regard to all the LP students, who had a relatively low writing self-efficacy and negative self-belief for writing, outcome affect like tranquility was elicited because the outcome (too much WCF) was exactly as what they had anticipated. The nature of HP and LP students’ affect was also dynamic. Apart from disappointment and worry, Xu (HP) also expressed her happiness because the WCF was a good chance to improve her writing skills. In addition, two LP students (Jin and Lu) conveyed their thankfulness and trust for the teacher because of her labor and workload in providing the feedback, which was contrary to the argument of Zhang et al. (2021) that HP students were more concerned about teachers’ workload. With regard to the judgement on the feedback on grammatical errors, all the students accepted it, as reported by the four students that the teacher was a native speaker who had a better command of the fixed grammatical rules, indicating their high belief in the teacher (Han, 2017), especially in her identity. However, in terms of the feedback on word choice errors, HP students (Li and Xu) expressed their skeptical attitude, showing their high belief in themselves. Despite the various affects and judgements of the six students, all of them expressed their appreciation for the feedback because it was useful and effective to improve their writing skills, contrary to the Truscott’s (1996) strong objection to the necessity and utility of WCF.

With respect to the behavioral engagement, students did spare some effort into correcting their errors with the help of teacher WCF, but the amount of effort varied between and within the groups. Compared with LP students, HP students (Li and Cao) attended to the WCF more wholeheartedly, correcting almost all the errors. It could be attributed to the fact that HP students, because of the individual factors like higher writing self-efficacy (Tsao, 2021) and language proficiency (Lee, 2008; Zheng & Yu, 2018), were more cognitively engaged with the WCF than the LP students were, even though both of the groups self-reported their positive affective engagement with the WCF. Nevertheless, the positive affective engagement with teacher WCF triggered by the teacher’s responsibility, heavy workload and identity as a native speaker had the potential to get students become more behaviorally engaged with the feedback, since a positive teacher-student relationship, one of the contextual factors, could conduce to more extensive student engagement with WCF (Zheng et al., 2020). The deployment of learning strategies for writing improvement could also be reflected by students’ learner belief (Han, 2017). Like HP students, Lu (LP) resorted to herself upon receiving teacher WCF by looking up the dictionaries and consulting the Internet to figure out the errors instead of directly asking for help from peers and the teacher like the other LP students (Jin and Luo) did. Meanwhile, taking a mental note (Ferris, et al., 2013) and memorization (Storch & Wigglesworth, 2010) were other learning strategies used by HP students and Lu.

From the agentic perspective, two HP students offered suggestions and communicated their preferences to the teacher after receiving the WCF, while the rest dared not to do that during the semester, since low writing self-efficacy could result in students’ inadequate confidence in expressing their preference for teacher WCF (Tsao, 2021). As in the cases of LP students, Luo mentioned that what she really needed was some trainings on English grammar, and that it was impossible to do such an adjustment to the teaching plan at the cost of other HP students. In addition, if she dared to give such a suggestion to the teacher, she would attract the teacher’s attention in class and then would be asked to answer questions. The failure in answering the questions would put her in an embarrassing situation in class. In the case of Cao (HP), his reluctance to express preference reflected students’ tenacious belief in teachers’ authoritative role, deeply rooted in the Chinese culture of learning (Han, 2019). Michou et al. (2021) contended that cognitive, behavioral, emotional, and agentic components constituting the multidimensional construct of engagement were interrelated and inter-connected. It is worth mentioning that Li’s and Xu’s behavioral engagement were impacted by their agentic engagement. Specifically, Li’s suggestion was taken by the teacher, which encouraged her behavioral engagement with the WCF by correcting all the errors as shown in Table 1. On the contrary, the teacher’s refusal of Xu’s suggestion incurred her careless revision behavior and inadequate engagement (Reeve & Shin, 2020), with two errors left uncorrected.

Implications and limitations

Research findings can provide some implications to both school administrators and English writing teachers for enhancing students’ engagement with teacher WCF. With regard to school administrators, firstly, English writing classes should be set up according to students’ language proficiency. Specifically, students with similar English proficiency should be grouped together and then take the English writing course that matches their proficiency. Considering the current situation of private colleges in China that up to 35 or even more students with various levels English proficiency crowding in one classroom, it is really difficult for teachers to provide appropriate and detailed feedback to students and for students to extensively engage with teacher WCF. From the perspective of school administrators, it would be impossible to reduce class size, which can increase cost, but grouping students according to students’ language proficiency and providing proficiency-matching classes would be helpful and beneficial for both teachers and students’ overall engagement. Secondly, in view of LP students, school administrators can take advantage of extra-curricular activities or student learning clubs, in which HP students can be hired to help LP students with grammar and even scaffold them.

In terms of teachers, firstly, to LP students, they can mainly provide direct WCF, which is believed to be more explicit and effective than indirect feedback (Bitchener & Ferris, 2012). To HP students, direct WCF and indirect WCF can be alternately offered. Indirect feedback can be challenging and demand more cognitive and meta-cognitive operations, which can inspire HP students to improve. Secondly, to raise students’ writing self-efficacy and learner belief, apart from WCF, teachers can also provide a positive comment to recognize the good points in the composition. Thereafter, the encouraged students can voluntarily engage more with the feedback. Thirdly, to reduce confusion and enhance students’ cognitive, affective and behavioral engagement, teachers should spare some time in class or their office hours in which students can ask questions and clarifications about the feedback. The interaction with students can build a strong teacher-student relationship, making teachers more trustworthy and raising students’ belief in the teachers. Last but not least, optional anonymous survey can be periodically distributed to the students to collect students’ opinions about and suggestions for WCF, which is a good way to foster an autonomy-supportive learning environment (Reeve & Shin, 2020).

Despite the possible contribution of the current study, there were some inevitable limitations to this study. Firstly, since the participants were recruited according to their own willingness, there was a gender imbalance, with five female students and only one male student. Future research might recruit more male students to investigate their engagement with teacher WCF. Secondly, previous teaching experience in English writing classes and a native speaker holding a Master’s Degree in TESOL do not necessarily mean that the teacher has a good command of implementing WCF. Therefore, future studies can first assess the teachers’ knowledge of how to use WCF effectively, specifically, how exactly to and when to implement WCF. If the teacher is tested with less knowledge of WCF, follow-up training programs can be offered to him or her. Thirdly, considering the interrelated and interconnectedness of the four dimensions of engagement, future research is expected to make an experiment in which an optional anonymous survey can be periodically distributed to the students to check how agentic engagement influence the other three dimensions and get students more engaged with teacher WCF. Fourthly, future research can cover more than one feedback-revision cycle in data collection to capture the changes and dynamic nature in student engagement. Lastly, future studies can pay more attention to some other learner variables that impact students’ engagement with WCF, for instance, their future L2 writing selves and achievement emotions.

Appendices

Appendix 1 Interview protocol

• 1. Tell me about your learning experiences of English writing.

• 2. Teachers may give feedback on linguistic errors in your writing. In general, what do you think of teacher feedback on these errors?

• 3. What do you do upon receiving the WCF?

• 4. How do you feel when you receive the WCF on your first draft?

• 5. To what extent do you understand the teacher’s feedback on your errors?

• 6. There are many types of feedback on linguistic errors, such as underlining, correction, giving clues, and the like. What type of feedback do you prefer? Why?

• 7. Tell me about your experiences of writing two drafts of this English essay.

• 8. What did you do with the linguistic errors in your first draft?

• 9. What were you thinking about when reading your teacher WCF?

• 10. What were you thinking about when revising your first draft?

• 11. During the semester, did you provide some suggestions on the teacher WCF to the teacher? If you did, what are the suggestions?

• 12. Do you have some comments on or suggestions for the teacher WCF now?

Appendix 2 Error Categories Adapted from Ferris (2006)

Error type	Description
Word choice	Excluded spelling errors, preposition errors, pronouns, informal and unidiomatic usage
Verb tense	Tense and aspect errors
Verb form	Excluded verb tense errors
Word form	Excluded verb form errors and verb tense errors
Articles	The misuse of zero, definite, and indefinite articles
Singular-plural	Noun ending errors
Pronouns	The misuse of pronouns
Run-on	Included comma splices
Fragment	Incomplete clauses
Punctuation	Inappropriate choice of punctuation marks. Excluded run-ons and fragments
Spelling	Misspelled words
Sentence structure	Included missing and unnecessary words and phrases and word order problems. Excluded run-ons and fragments
Subject-verb agreement	Excluded other singular-plural or verb form errors
Preposition	Inappropriate choice of prepositions
Miscellaneous	Errors that could not be otherwise classified

Appendix 3 Categories and sub-categories of student engagement with teacher WCF

Categories/Sub-categories	Definition
Cognitive engagement
Awareness: noticing	The extent to which a learner detects WCF, recognizes the teacher’s corrective intention, and attends to linguistic accuracy
Awareness: understanding	The extent to which the learner successfully diagnoses the error, and is able to provide accurate metalinguistic explanations
Meta-cognitive operations	Strategies and skills the learner employs to regulate his or her mental processes, practices, and emotional reactions
Cognitive operations	Cognitive strategies and skills that the learner uses to process and respond to WCF,
Affective engagement
Affect	Feelings and emotions expressed upon receiving WCF and changes in these feelings and emotions over the revision process
Judgement	Personal judgements of admiration or criticism to WCF
Appreciation	Valuing the worth of WCF
Behavioral engagement
Revision operations	Correct revision, incorrect revision, deletion, substitution, and no revision
Revision and learning strategies	Observable strategies taken to enhance the accuracy of the draft, and/or to improve the future writing pieces and even L2 competence
Agentic engagement	Direct communication with the teacher about the teacher WCF