Individual Differences in Written Corrective Feedback : A Multi-case Study

Written corrective feedback (WCF) has been a long time practice in L2 writing instruction. However, in many cases, the effects are not satisfactory. There have been controversies about it both theoretically and empirically. This paper reports a multi-case study exploring individual differences that impact learners’ responses to WCF. Four students’ compositions in an English course with writing components were examined. Then, the students were interviewed. The results of analyses reveal, though all the four students were motivated to study by a compulsory provincial English test, students of significantly lower proficiency were motivated to significantly more independent learning activities after class. As a result, they achieved significantly greater improvement. It is also revealed that WCF did not lead to simplified writing in terms of lexical diversity and structural complexity, for the students knew the grading criteria for the written task in the compulsory test. Moreover, this study shows that Ferris (1999) classification of treatable and untreatable errors should be used with reference to the students’ present zone of proximal development (ZPD). Finally, pedagogical suggestions are provided.


Introduction
Written corrective feedback (WCF) is the input from the reader to the writer that provides information for revision (Keh, 1990).It has been applied in L2 writing instruction for a long time.However, the results often seem not to be satisfactory.Controversies about it can be seen in both SLA theories and empirical studies.The effect of WCF even varied significantly in participants receiving the same kind of WCF in the same experiment (Santos, Lopez-serrano, & Manchon, 2010).Ferris (2006) found the impact of individual differences (IDs) on the students' utilization of WCF to improve their L2 writing proficiency.Yet, IDs is an under-researched area in WCF research.For a better understanding of how IDs influence the function of WCF among L2 learners, this multi-case study explores the different responses of four Chinese learners of English to their teacher' s WCF and factors underpinning such responses.

Controversies in the Benefits of WCF
Controversies about WCF can be seen both in theories and empirical studies.

Theoretical Controversies
SLA research found error cannot be avoided in SLA.Truscott (1996) proposed, theoretically, WCF would disturb the natural development order of SLA.There is also a possibility of "pseudo-learning", which results in the learners' peripheral and superficial acquisition of language forms.Moreover, he pointed out, WCF could make the students nervous.To avoid error, they would try to use simple language, which results in simplified writing.He also considers that learners' concentration on language form while writing would negatively impact the content they write.Hence, he claims WCF is harmful to L2 writing and should be avoided.However, WCF finds supports in mainstream SLA theories-cognitive theories and Vygotskian theory.
SLA research reveals SLA is a process of input-intake-output (Brown, 1994).Krashen (1985) proposed only comprehensible input could be converted into intake, and comprehensible input is necessary and sufficient for SLA.Schmidt's (1983) case study of a Japanese working in the U.S. for three years showed comprehensible input alone could not guarantee intake.He (1990) proposed "Noticing Hypothesis": only the comprehensible input noted by the learner can be assimilated and become part of interlanguage.In 1985, Swain proposed "Comprehensible Output Hypothesis".On the one hand, she agrees with Krashen on the necessity of comprehensible input in SLA; on the other hand, she points out that comprehensible output can make the learner notice the gap between what he/she has already mastered and what he/she wants to express, thus notice the language form in input and evoke intake.Hence, in writing-a kind of comprehensible output, learners can notice the language form spontaneously. To compare, WCF-a source from outside-can draw learners' attention to the language form they have not mastered or mastered partially, which, in turn, can evoke intake.
According to Vygotskian theory, learning first takes place interpersonally, then intrapersonally.It emphasizes the scaffolding of interpersonal communications in learning (Mitchell & Myles, 2004).Hence, WCF-feedback from the reader to the writer-can serve as a scaffold of the learners.With the help of WCF, learners can fulfill tasks that they cannot when independently.Moreover, according to the concept of "zone of proximal development" (ZPD) in Vygotskian theory, WCF can work effectively as a scaffold only when it is within the learner's present ZPD.This implies, in order to offer appropriate WCF, the teacher must analyze and understand the students' present L2 proficiency in advance.Therefore, Vygotskian theory not only supports the benefits of WCF in L2 instruction, but also points out the matters that need attention in offering WCF, thus solve the problem raised by Truscott (1996) about WCF's disturbance of the natural development order of SLA.

Controversies in Empirical Studies
Empirical studies on WCF focus on the effectiveness of selective feedback, among which many explored the effect of different types of selective feedback simultaneously.Though these studies disagree with one another on the latter issue, most show, compared with the control group which received no WCF, the experimental group which received selective feedback ( though in different ways ) improved their L2 writing accuracy significantly (Bitchener, Young, & Cameron, 2003;2005;Bitchener & Knoch, 2010;Ellis, Sheen, Murakami, & Takashima, 2008;Ferris, 2006;Zhang & Deng, 2009).However, Truscott (2007) adopted qualitative analyses on some WCF experiments, together with meta-analyses of the findings of these experiments.According to him, the results showed, WCF slightly harms the accuracy of L2 writing.If it has any benefit, the benefit would be too small to count.This is a conclusion contrasting to those in the reports of the experiments being analyzed.
Similarly, empirical studies on comprehensive feedback have already generated contrasting results though they are rare.Truscott and Hsu (2008) reported experimental group with comprehensive WCF did not improve their writing accuracy, while Van Beuningen, De Jong, & Kuiken (2012) reported the opposite.
The different effects of WCF have also been observed among individuals receiving the same kind of WCF in the same experiment (Santos, et al., 2010).Hence, IDs may have played a role in the controversies in WCF empirical studies.

IDs in WCF
L2 learners' utilization of WCF for improvement is impacted by IDs (Ferris, 2006).Yet, IDs in WCF is under-researched.Sheen (2011) explored some IDs in WCF in detail, but only focused on anxiety, aptitude and attitude.
In comparison, IDs in SLA have been research foci for a long time.Language learning motivation (LLM) research has found integrative motivation significantly and positively correlated with learning outcome, while language learning strategy (LLS) research has shown learners of higher proficiency tend to adopt more and deeper LLS, and they have a wider LLS repertoire than their peers (Dornyei, 2005;Ellis, 2008).
Since such IDs as motivation, strategy and L2 proficiency have been found important in SLA, how they impact learners' responses to WCF deserves exploration.To fill the gaps in WCF research, this case study tries to answer the following questions: 1) Does WCF result in simplified writing?Why/why not?
2) Do the students achieving significantly different improvement vary significantly in their strategies when responding to WCF?If so, how?
3) Are the students aware of their strategies used in responding WCF? 4) What underpins their strategies used in responding WCF?

Participants
A quasi-experiment was conducted in an intact class of 45 students in a vocational college from June 2011 to January 2012.The provincial CET2 in June 2011 served as the pre-test, while the CET2 in January 2012 served as the post-test.During this period, the students wrote 5 compositions in responding to the written tasks in the "Simulation exercises with detailed explanations of College English Test, Band 3" (Yi, 2010).Draft 1 of each composition was finished within 30 minutes in class.Then, the compositions were collected and graded by the teacher who underlined every error and provided error type nearby.After that, the compositions were handed out to the students for revision.Every student was entitled to consult WCF to him/her with the teacher in the subsequent classes.
The subjects in the case study were chosen from those in the quasi-experiment.The only criterion for the choice was their progress in their English proficiency measured by pre-and post-test.Altogether, four students were chosen: two made the greatest progress and two the least progress.They are: S1: scored 58.5 in the pretest, but 43 in the post-test; S2: scored 57 in the pretest, but 57.5 in the post-test; S3: scored 46 in the pretest, but 63 in the post-test; S4: scored 51 in the pretest, but 66 in the post-test.

Instruments
1.The four students' original drafts and revisions of five compositions during the quasi-experiment.

Interviews:
To explore the students' awareness of their learning strategies in writing after the teacher's WCF as well as factors underpinning their strategies, the four students were interviewed by an outside researcher who didn't and would never teach them.The Interview guides are as follows: 1) Why did/didn't you use more simple language to reduce errors in the compositions?
2) Did you read the teacher's WCF each time after your composition was handed out?
3) After reading the WCF, what did you do next?

Data Collection and Analyses
The study consists of two phases.
Phase 1: Since Truscott (1996) claimed WCF would do harm to the complexity in L2 written language, the complexity of every student's five original drafts was compared to testify such a claim (in terms of lexical diversity and structural complexity).
Then, the students' original drafts and revisions were examined, their responses to WCF in each composition were recorded.Times of response to WCF were counted, and frequency of response was calculated.Also, ways of response and error types were matched.Thus, categories of their learning strategies emerged.Next, differences and similarities among the students were compared.
Phase 2: After the students' responses to WCF in the compositions were analyzed, the students were interviewed in the teachers' office after class.The interviewer was a teacher outside the college where the students were studying.The interviews were recorded and transcribed by the interviewer.
Content analysis was adopted in the analysis of each student's verbal report.Categories emerged in Phase 1 were applied here to check the students' awareness of their learning strategies.Factors underpinning their strategies formed new categories.Then, differences and similarities were compared among the students.

Results of Phase 1
Examination of the students' original drafts showed no change in the complexity of their written language when lexical diversity and structural complexity are concerned.
Revisions of S1 were mainly copies of her original drafts.The types of error corrected were misspelling and subject-verb agreement.S2 corrected most of his errors in Composition 1 and 4 according to the WCF before copied them into his revisions, while for Composition 2, 3 and 5, he only corrected few errors.The types of error corrected were misspelling, subject-verb agreement, tenses and word choices.S3 not only corrected most of the errors in the original drafts before each revision, but also copied each misspelled word in the original drafts three times and some grammar rules with illustrations after the revisions.For example, she copied the table of possessive pronouns (I, me, my, mine) after the revisions of Composition 1 and 2 after corrected her errors in this field in draft 1 of both compositions.The errors she corrected were: misspelling, subject-verb agreement, tenses, word choices, word orders, clauses and infinitives.However, some of the errors in sentence structure in all the five compositions were left uncorrected such as a complete sentence serving as the subject/object of a sentence.S4 also corrected most errors before revision.After revision, she copied words wrongly used and the appropriate words in pairs, together with rules and illustrations for their usage.The errors she corrected were: misspelling, subject-verb agreement, tenses, word choices, clauses and infinitives.Yet, like the case of S3, S4 also left come errors in sentence structure uncorrected.For instance, dangling participle remained in the revisions of all five compositions.In Composition 2, such an error occured twice in both draft 1 and revision.

Awareness of Their Learning Strategies
According to the students' reports, they all read the teacher's WCF carefully each time.However, their learning actions after that varied greatly except that no one asked for direct answers, which they all considered useless for improvement.
Both S3 and S4 tried to correct all the errors and mistakes.They first tried themselves, consulting dictionaries and grammar books.If this didn't work, they would turn to their friends for help.If their friends couldn't help them, they would turn to the teacher in class.Such self-reports are in line with the high rate of response to WCF in their revisions.
In contrast, S1 reported that she usually didn't respond to the WCF but just corrected some mistakes subconsciously in revisions.This concords to the low rate of response to WCF in her revisions.Likewise, S2's self-report is also supported with his responses to WCF found in the examination of his compositions.He said when he was in the mood for study, he would consult the dictionary and grammar books, then correct the errors himself; otherwise, he would just put the composition book aside after read the WCF.Moreover, he felt he was often not in the mood for study.
Hence, these students are aware of their learning strategies after receiving the teacher's WCF, and can describe them objectively.Therefore, the factors underpinning these behaviors they provided are reliable to a large extent.

Reasons for not Reducing the Complexity of Their Written Language
All the four students explained the scoring criteria for written task in CET3 were listed in the CET3 simulation exercise book they were using.They found language complexity (grammatical structure and vocabulary) was one of the criteria.Besides, the teacher reminded them of these criteria in class.Hence, they preferred to take the risk of making errors than losing marks on language complexity.

Factors Underpinning Their Learning Strategies
All these students expressed their eagerness to pass CET2, which is related to their diplomas.According to the college regulations, diplomas will not be granted to those who cannot pass CET2 in three years (Note 1).Though passing CET2 is the first thing to them, they demonstrated different degrees of willingness to take measures in study.
S1 -unwillingness to study due to satisfaction with her present English level S1 elaborated that, with a high score in the pre-test in the class, she believed she could pass CET2 next time with little work."58.5 means I'm a top student in the class (60 means pass).I don't have to work hard.My friends can't teach me…I didn't ask the teacher for help.When giving me some hints for further study, which I don't like, the teacher would know my weaknesses in English.And that hurts my image as a top student."S2 -struggle between willingness to work hard or not due to satisfaction with his present English level S2 expressed his struggle between willingness and unwillingness to work hard."I scored 57 in the pre-test.That was high.Moreover, I worked as a part-time waiter in a four-star hotel downtown.My English enabled me to communicate with the foreign guests there with the help with some body language.I'm quite contented with my English.I know, as a student, I should always study hard to make progress.But occasionally, I was unwilling to make efforts.I tried to force myself to study, but failed occasionally."S3 and S4 -strong will to work hard due to discontent with their present English level Both S3 and S4 expressed their determination to work hard."I was so poor in English.I had to work harder than others.The teacher's WCF helped me locate the deficiencies in my English.I worked on them, and felt I was making up for the deficiencies.So, I was happy with the hard work all the time", said S3.
Similarly, S4 expressed her strong will to work hard because she saw a large gap between her English level and her goal of English learningpassing CET2.In her opinion, WCF showed her the exact points in her English that demanded improving.Thus, "it pays to work hard with the guidance of WCF."However, in the examination of the students' compositions, it was noted that both S3 and S4 left some errors not responded (All are problems with sentence structures).This seems contradicting to their strong will to work hard.Regarding this phenomenon, S3 replied all the uncorrected ones are sentence structure errors."I've been reading a grammar book, and know I did badly on sentence structure in middle school.It's tough to make up deficiencies in this field.I don't even know clearly the sentence structure in Chinese language.And we don't have Chinese classes in this college.My friends can't help me, for they are confused about sentence structure themselves…I turned to the teacher often.But my problems in sentence structure can't be solved within 10 minutes in most cases.I dare not take the teacher more time, fearing my classmates would get unhappy with me.They also want the teacher's help.So, sometimes, I had to pretend to know the answer to conclude the one-to-one conferences with the teacher.Therefore, such errors were left uncorrected."S4 also expressed her difficulties in learning sentence structure by herself.Besides, she was also unwilling to take the teacher too much time for a single problem.

Interrelationship between WCF, Independent Learning and Progress
The learning strategies of S3 and S4 were alike.Both involved themselves in considerably more independent learning activities than S1 and S2 after reading the teacher's WCF.Also, they made considerably more progress than the latter two.This coordinates with Ferris and Helt's (2000) finding that students who maintained error logs after WCF -a kind of independent learning activityimproved their L2 writing accuracy more than those who didn't.Hence, we can infer that independent learning has positive effects on SLA, and WCF can evoke independent learning activities.

Benefits of One-to-one Conferences with the Teacher
Both S3 and S4 made great progress during the quasi-experiment.Both recalled they had talked to the teacher often for further hints to correct errors themselves after reading WCF.This supports Bitchener, Young and Cameron's (2003) suggestion that WCF plus one-to-one conference benefit L2 writing accuracy.It also supports Bitchener's (2005) view that teacher is a source of knowledge in error correction when the students have exhausted their own efforts.

On the Classification of Treatable and Untreatable Errors
Though both S3 and S4 worked hard to improve their English according to the teacher's WCF, both left some errors in sentence structure uncorrected.According to their explanations, it's difficult for them to solve these problems by themselves although they tried hard both after reading the WCF and the subsequent one-to-one conferences with the teacher.This may suggest that Ferris's (1999) classification of treatable and untreatable error should be used with reference to the students' present ZPD: going beyond it, any treatable error in Ferris' classification can become untreatable ones.Thus, detailed grammar instruction and direct feedbacks are expected.

Effects of Instrumental Motivation
It is noted that all the four students received the same kind of WCF, and held the same motivation and goal in English learning: to pass CET2.However, their amount of independent learning after receiving WCF differed greatly: the two students with significantly lower proficiency made significantly more efforts in learning.As learning to pass tests is a kind of instrumental motivation, the great difference in its efficacy with these two pairs of learners may suggest that instrumental motivations can hardly motivate learners who perceive a little gap between their goals and their present L2 level.To contrast, for those who perceive a large gap between the two, instrumental motivations can be very effective in L2 learning.Hence, students shouldn't only be driven to study English by tests as they are now.Cultivation of their integrative motivations should also be emphasized.

Conclusion
This multi-case study examined the students' original drafts and revisions of five compositions in a quasi-experiment, compared the differences and similarities in the students' responses to the teacher's WCF as well as the progress they made measured by the pre-and post-test.It reveals WCF did not lead to simplified writing.It also reveals, though the students received the same kind of WCF, and have the same motivation and goal in English learning, the two students whose English level was near the goal involved themselves in significantly less independent learning after reading the teacher's WCF than the two whose English level was significantly lower.Moreover, the latter two improved greatly.They passed CET2 at the end of the quasi-experiment, while the former two failed again.
Such findings support that independent learning benefits L2 accuracy (Ferris & Helt, 2000) and WCF, which can evoke independent learning, does not harm SLA.They also suggest, when comparing with their goals in L2 learning, learners who consider themselves good may be motivated little by instrumental motivations though for those who consider themselves poor, such motivations may work efficiently.Due to the features of foreign language learning context such as lacking in the need/chance for real communication in the target language and contact with a different culture, it is hard for the students to develop integrative motivations all by themselves.Therefore, in such a context, in order to improve the students' willingness to involve themselves in independent learning activities, special attention should be paid to the cultivation of the students' integrative motivation.
In addition, both S3 and S4 left some errors in sentence structure uncorrected despite of their eagerness to pass CET2 and their efforts to solve such problems.This may suggest flexible application of Ferris' (1999) classification of treatable and untreatable errors in L2 writing instruction.When the treatable errors go beyond the students' present ZPD, they become untreatable ones.Thus, detailed grammar explanation and correct forms should be provided.
This study shed some new lights on learners' differences in responding WCF.However, due to the limitations of case studies, more research into learners' individual differences in responding WCF is needed for our understanding of how WCF works in SLA, which, in turn, will inform L2 writing instruction.