The Effects of Written Feedback in the Form of Recasts

THE JOURNAL OF ASIA TEFL Vol. 9, No. 4, pp. 27-50, Winter 2012
The Effects of Written Feedback in the Form of
Recasts
Rintaro Sato
Nara University of Education, Japan
This study examined the effects on 25 Japanese university students’ text
revisions following teacher feedback through written recasts. Written recasts
were categorized as recasts given on learners’: (1) grammatical errors; (2)
lexical errors, (3) unsolicited use of Japanese, (4) spelling errors and (5)
content. Recasts were also categorized according to (6) the degree of
differences between students’ original erroneous utterances and written recasts,
as well as (7) recast lengths. In total, 125 written recasts were provided on
students’ essays, and students were directed to revise their first draft referring
to these recasts. The students’ revised essays were quantitatively and
qualitatively analyzed. The results showed that written feedback in the form
of recasting is, in general, beneficial for learners to notice their errors or
mistakes leading them to repair regardless of the degree of differences and the
length of the recasts. However, their effects varied according to students’ error
types. The development of their writings from the first draft to a revised
version was examined from the points of accuracy, fluency and complexity. It
was found that written recasts contributed to the development of the three
aspects reducing trade-off effects between accuracy and fluency.
Key words: writing recasts, complexity, writing feedback, errors,
Japanese tertiary learners
INTRODUCTION
Writing is one of the crucial skills in students’ English learning—whether in
27
The Effects of Written Feedback in the Form of Recasts
English as a second language (ESL) or English as a foreign language
(EFL)—though even ESL learners struggle to produce linguistically correct writing
(e.g., Hartshorn, Evans, Merrill, Sudweeks, Strong-Krause, & Anderson, 2010).
Teachers may try to give the best feedback to help students improve their writing.
Written feedback can be focused on form or on content, and both have played
crucial roles in improving student writing quality (Coffin, Curry, Goodman,
Hewings, Lillis, & Swann, 2003). Previous studies found that not only teachers, but
students themselves prefer teacher written feedback (e.g., Nugrahenny, 2007).
However, since Truscott’s claim (1996) that written corrective feedback would
never improve learner writing ability—and may even be harmful—it has been
debated to what extent learners can benefit from written feedback. There seems to
be some agreement that learners can improve their writing in a second draft on the
same topic after being given corrective written feedback (e.g., Ellis, Sheen,
Murakami, & Takashima, 2008; Ferris, 2004; Truscott, 1996, 1999). However, it
has not been well researched whether accuracy, fluency or complexity in writing
improves and how many errors or mistakes in writing are corrected.
As for the effects of feedback, recasts have been gaining attention. Though the
beneficial effects of recasts as oral feedback on second language (L2) acquisition
have been reported (e.g., Doughty & Varela, 1998; Loewen & Philip, 2006;
Loewen & Nabei, 2007; Muranoi, 2000), several studies have also demonstrated
the ineffectiveness of recasts (e.g., Lyster, 1998a, 2004; Lyster & Ranta, 1997;
Varnosfadrani & Basturkmen, 2009). Studies on recasts both in classroom and
experimental settings have shown mixed results as to a facilitative role for learning.
Previous research showed that recasts in the form of written feedback would be
more effective than positive evidence (a model of what is grammatical and
acceptable) in the form of models and traditional grammar instruction (Ayoun,
2001). However, Ayoun (2004), which is a follow up study of Ayoun (2001),
concludes it is difficult to compare the effect of the three treatments (recasts, models,
instruction) on acquisition.
The present study with Japanese university students examines the effectiveness
of written feedback in the form of recasts by measuring: to what extent learners
would repair their errors or mistakes according to types of recasts; the development
28
The Journal of Asia TEFL
of accuracy, fluency and complexity in writing; and variation of the correlation
between accuracy and fluency.
BACKGROUND
Definition of Recast
Lyster and Ranta (1997) defined recasts as reformulation of all or part of the
students’ utterances. Long (2006) redefined corrective recasts as a reformulation of
learners’ preceding utterance in which non-target-like item(s) is/are corrected to
target language form(s) while the interlocutors’ focus is not on language but on
meaning. Recasts are, in general, considered as implicit corrective feedback
reformulating all or part of ill-formed utterances provided by learners without
changing the central meaning (Iwashita, 2003; Long, 1996; Lyster, 1998a, 1998b).
Long (2006) states that L2 research findings have shown that recasts in the L2
are effective, also adding that recasts are not clearly necessary for acquisition but
are facilitative and especially efficient for older, more proficient L2 learners in that
they do not interrupt the flow of conversation, and thus keep learners focused on
message content. Lyster (2007) also states that recasts help maintain the flow of
communication, keeping learners’ attention on content and enabling them to
participate in interaction in which their linguistic abilities can exceed their current
level. In fact, a number of previous experimental studies have provided positive
reports on the impact of recasts in L2 acquisition (e.g., Doughty & Varela, 1998;
Loewen & Nabei, 2007; Loewen & Philp, 2006; Muranoi, 2000).
However, some previous studies showed that recasts were less effective than
other types of feedback. Lyster (2004) compared the effects of recasts and prompts
(i.e., clarification requests, repetitions, metalinguistic clues and elicitation) in a
quasi-classroom study, and statistical analysis of the results of the written tasks
revealed that students receiving prompts performed better than students receiving
recasts. Varnosfadrani and Basturkmen (2009) argued that explicit correction
would induce learners’ awareness more than implicit correction such as recasts,
referring to the crucial role of attention in learning.
29
The Effects of Written Feedback in the Form of Recasts
Types of Recast
Previous studies examined whether the degree of difference between the
learner’s initial utterance and the recast would affect their effectiveness by counting
the number of changes, and found that recasts with only a single change are more
likely to be noticed by learners more than recasts with multiple changes (Philip,
2003; Sato, 2009). Lengths of recasts (short or long according to the number of
morphemes) and their effects were also examined, and it was revealed that learners
can notice shorter recasts to reformulate their original erroneous utterances (Philip,
2003; Sato, 2009). As for error types of learners’ original erroneous utterances to
which recasts were given, previous studies showed grammatical recasts were less
likely to be effective (e.g., Kim & Han, 2007; Sato, 2009; Trofimovich, Ammar, &
Gatbonton, 2007; Williams, 1999). Trofimovich, Ammar, and Gatbonton, (2007)
suggest that in order for learners to notice their own errors through grammatical
recasts and to reformulate them after recasts, they should already have knowledge
of the form. Previous studies showed that recasts of lexical errors are noticed more
by the students, showing they have greater effectiveness than those directed at
grammatical errors (e.g., Kim & Han, 2007; Oliver, 1995). In Sato (2009), it was
reported that recasts addressed at students’ L1(Japanese) led to the highest rate of
immediate repair, and explained that this is attributed to the salience of the recast
and the cognitive process students experienced when they decided to speak
Japanese followed by the English equivalent (recast). He interpreted that when
feedback was given immediately after students found interlanguage deficiencies,
they could easily understand the corrective purpose of the recast, leading them to
produce what they had wanted to say.
Written Recast
Ayoun (2001) examined the effectiveness of written recasts by comparing it with
modeling (pre-emptive positive evidence) and grammatical instruction (explicit
positive evidence and negative feedback). In the study, in testing the effectiveness
of the three conditions, the acquisition of the aspectual distinction between the past
30
The Journal of Asia TEFL
tense in French was used, and recasts were given as implicit negative evidence.
Post-test results revealed written recasts were significantly more effective than
grammatical instruction, but not modeling. Although written recasts could have
served as implicit positive evidence in some cases, it was concluded that the result
partially supported recasting as the most effective form of feedback. In the follow
up study, analyzing the mixed results, Ayoun (2004) concluded that the acquisition
of subtle and complex structures requires both implicit negative feedback such as
recasts and explicit negative feedback as provided by traditional grammatical
instruction.
Trade- Offs in Writing
Skehan (1996) points out that there are three aspects of production: accuracy,
fluency and complexity. Accuracy is defined by Skehan (1996) as the extent to
which the target language is produced in relation to the rule system and how well
the learner can handle whatever level of interlanguage complexity he/she has
achieved. Ellis (1987, 2003) mentions that accuracy requires syntactic processing
with the availability of planning time. Fluency refers to learners’ ability to mobilize
their system to communicate meaning in real time, prioritizing meaning over form,
and is achieved when learners can exercise strategies to avoid or solve problems
quickly (Ellis, 2003). Complexity is defined as the extent to which elaborate
structured interlanguage is utilized (Skehan, 1996). In writing, referring to previous
studies (e, g., Ellis, 1987, 2003; Skehan, 1996), it can be argued that: accuracy
concerns how precisely the learner can write what he/ she wants to write; fluency is
likely to be indicated by a high rate of writing; complexity concerns to what extent
the language produced is elaborate (Hunt, 1970; Sato, 2008; Tong-Frederics,
1984;).
As for the relation between accuracy, fluency and complexity, Ellis (2003) argues
that there could be trade-offs in L2 learners’ production, meaning that when L2
learners attend to accuracy in their writing, it interferes with their ability to
conceptualize, formulate, and articulate messages, preventing them from showing
fluency. Skehan and Foster (1997) argue that in general fluency may be
accompanied by either accuracy or complexity but not both, referring to trade-offs
31
The Effects of Written Feedback in the Form of Recasts
in performance due to learners’ limited attentional resources. However, there also is
a contrasting view. Sato (2008) examined the relation between accuracy and
fluency in low-intermediate Japanese high school learners’ writing. One
experimental group was directed to be concerned about accuracy and the other
group was directed to be concerned about fluency. It was found that there was no
minus correlation between accuracy and fluency in the two experimental groups.
The control group, which was not directed to be concerned either about accuracy or
fluency, showed moderate correlation between them. He concluded that learners
can access multiple attentional resources. Some previous studies found that written
corrective feedback decreases fluency; letting learners focus more on accuracy at
the expense of fluency (e, g., Kepner, 1991; Sheppard, 1992). To my knowledge,
the effects of written corrective feedback on Japanese EFL learners—in relation to
the three aspects—have not yet been fully examined.
Rationale of the Investigation of Written Recasts
There is a general agreement that recasts are most commonly employed as oral
correction by teachers in EFL or ESL classrooms (e.g., Lyster, 1998b; Lyster &
Ranta,1997) and several studies showed the beneficial effects in learning (e.g.,
Ayoun, 2001; Doughty & Varela, 1998; Iwashita, 2003; Leeman, 2003). However,
Philip (2003) pointed out the limitations of working memory as one of the factors
which hinder the beneficial effect of recasts. Recasts in the form of oral corrective
feedback demand an immediate cognitive comparison also requiring learners to be
dependent on short-term memory. In the study which compared the effects of face
to face communication and computer-mediated communication on L2 development,
Payne and Whitney (2002) found greater improvements in oral proficiency in the
post-test for learners who were in the computer-communication group than those in
the face to face communication group. They interpreted that computer-mediated
communication supported students who were less able to maintain oral information
in memory: Interlocutors’ feedback was less fleeting as learners were able to trace it
by reading. Williams (1999) also suggested that if cognitive comparison does not
overtly tax learners’ attentional resources, learners with lower working memory
would benefit in feedback. As written feedback is delayed and imposes less
cognitive demand without requiring immediate on-line cognitive comparison, we
32
The Journal of Asia TEFL
could assume that written recasts would be better noticed by learners leading them
to L2 development than recasts provided as oral feedback.
Several studies, for example, conducted by Lyster and his colleagues (e.g., Lyster
& Ranta, 1997; Lyster, 1998b, Lyster, 2004; Panova, & Lyster, 2002), demonstrated
that recasts are the least effective means of oral corrective feedback. However,
some other studies conducted in the Japanese EFL context in which learners are
concerned with accuracy in their production, found recasts to be effective (e.g.,
Loewen & Nabei, 2007; Muranoi, 2000). Therefore, it is worthwhile to examine
whether written recasts are effective in the Japanese EFL context.
Description of the Research Context
English has long been taught as a knowledge-based subject in Japan. It is said
that Japanese learners have dual orientations for learning English: a practical,
realistic goal related to examinations and grades; a vague idealistic goal related to
using English for international or intercultural communication, and for many
learners, passing of knowledge-based exams is the primary objective (Yashima,
2000). The data for the present study were collected in an English class titled
“Foreign Language Communication” in a national university in an urban area in
western Japan. The purpose of the English class is to improve integrated skills of
reading, writing, listening and speaking in English.
RESEARCH QUESTIONS
Measures of Effects of Written Recasts on Writings
In this study, we investigated the effects of written recasts by type. It seems that
there is an agreement that learners demonstrate improvement in their writing in a
second draft on the same topic after being given corrective written feedback. (e.g.,
Ellis et al., 2008; Ferris, 2004; Truscott, 1996, 1999). In the present study, the extent
to which written recasts can help learners in improving accuracy, fluency and
complexity in text revision and whether or not there are trade-offs between
accuracy and fluency were examined. Research questions were operationalized as:
33
The Effects of Written Feedback in the Form of Recasts
1. Are written recasts noticed, leading Japanese EFL students to repair in the text
revision?
2. How effective are written recasts for Japanese EFL students to repair in the
text revision according to error type?
3. How effective are written recasts for Japanese EFL students to repair in the
text revision according to the degree of differences between errors or mistakes
and recasts?
4. How effective are written recasts for Japanese EFL students to repair in the
text revision according to the lengths measured by the number of morphemes?
5. Do written recasts contribute to the development of accuracy, fluency and
complexity in the text revision?
6. Do correlations of accuracy with fluency in the writing change between the
first draft and the revised draft after learners are provided written recasts?
METHOD
Participants
The class consisted of 27 students (12 males and 15 females) between 19 and
21years of age. All of them belonged to the department of teacher training and
school education. The Japanese EFL teacher taught the class. He is the researcher of
the study as well. Two students who were absent from either the first or second
week of the class were excluded from the study.
Procedures
In the first class after summer vacation in 2010, students were assigned to write
an essay on the topic of “My Summer Vacation. Students were given 30 minutes to
complete the essay and were not given any direction on whether they should focus
on accuracy or fluency. Although this was not a test but a classroom writing task
they were not allowed to use a dictionary as it was aimed at measuring their writing
34
The Journal of Asia TEFL
proficiency from the points of accuracy, fluency and complexity. After they
submitted the essay, the rest time of the class time was spent with students engaged
in listening, reading and speaking activities which were not a part of this study.
The teacher-researcher wrote written recasts in the blank space of each essay. On
average 5 recasts were given to each of the students with a minimum of 2 (two
students) and a maximum of 9 (one student).
There was no target focus in providing recasts in the study. Though Bitchener
(2008) argues that there should be only one or a few categories for providing
feedback to prevent information overload, it was assumed that this would be
impractical in actual classroom settings. As students, in general, want to improve
overall accuracy in writing (Hartshorn et al., 2010), focusing on one or a few error
categories may lead to students neglecting other areas (Xu, 2009). Written recasts
were provided randomly depending mainly on the teacher’s common sense
intuitions and experience so that students could revise the first draft well enough to
improve the overall quality of the writing as is usually done in EFL classroom
settings.
In the second session, one week after the first session, each of the essays was
given back to the students. Students were asked to read through the feedback for the
text revision and given 20 minutes to revise the essay. They were not allowed to use
a dictionary this time, either. After finishing the revision, they were given a B5- size
white paper and were asked to comment in Japanese, about how they felt about
written recasts and on using them in revising their first draft, because the
information they were asked to deliver was complex.
Analysis
Classification of written recasts
In order to examine the effectiveness of recasts according to types, recasts were
categorized as recasts given to learners’: (1) grammatical errors; (2) lexical errors,
(3) unsolicited use of Japanese, following Lyster and Ranta (1997). Grammatical
errors are errors in the use or lack of determiners, particles, verb forms, word order;
35
The Effects of Written Feedback in the Form of Recasts
lexical errors are inappropriate, imprecise or inaccurate choices of lexical items;
unsolicited use of Japanese is an instance where a student writes Japanese instead of
English. In addition, recasts given to (4) spelling errors and (5) contents were
examined. Content recasts were provided to an expression(s) whose meaning(s) is
(are) vague or awkward. The following are examples according to types. Written
recasts were written in italics. Changes were written in bold and will be explained
later.
Example (1) Grammatical recast.
S1: It is beautiful. ← (written recast) Oh, it was very beautiful.
Example (2) Lexical recast
S2: In the car I saw a dream. ← You had a dream in the car?
Example (3) Recast to unsolicited use of L1
S3: We ate a lot of Ika (squid in Japanese) in Hakodate. ← Oh, you ate a lot of
squids!
Example (4) Recast to spelling error
S4: I talked with a foreingner. ← You had a talk with a foreigner!
Example (5) Content recast
S5: We lost the games. ← You lost all of the games?
Recasts were also categorized according to the degree of differences and lengths
following the parameters of previous studies (e.g., Philip, 2003; Sato, 2009). To
examine the effects of the degree of difference between the learner’s initial writing
and the written recast, the number of changes was counted and coded following
Philp (2003), but for the study, recasts were divided into two categories according
to whether the recast had only a single change or more than one change. This
decision was made referring to Sato (2009) which revealed that recasts with more
than one change were less likely to be noticed by the Japanese learners. Conversion
36
The Journal of Asia TEFL
of the subject was not counted as a change and inversion counted as one change.
Exclamation and interjection were not counted. The following example was
counted as one change.
Example (6) Grammatical recast
S6: Kourijima is small island.← Oh, it is a small island.
Examples (1) and (5) were counted as two changes; examples (2) and (3), as one
change; example (4), as four changes.
As for the lengths, written recasts were categorized into long or short ones
according to the number of morphemes, based on Philp (2003) and Sato (2009):
recasts with more than five morphemes were coded as long. Example (7) was
counted as short, while (8) was counted as long.
Example (7) Short recast (lexical, one change)
S7: I pointed an umbrella. ← Oh, you opened an umbrella.
Example (8) Long recast (grammatical, four changes)
S8: I don’t know what should I teach to child then.
← OK, you didn’t know what you should teach to children.
Drawing on previous studies that insist on the importance of modified output in
L2 learning (e.g., Gass, 2003; Izumi, 2002; Shehadeh, 2002; Swain, 1985; Loewen,
2004), students’ correct written repair was measured in the study. When students’
errors, mistakes or inappropriate expressions to which recasts were given, were
corrected in the revision, it was counted as successful (repair) and rewarded one
point. If they failed to make the revision, it was counted as failed and a point was
not given. We computed success rates. In the situation when students did not use
original utterances to which recasts were given, it was excluded in calculating the
success rates. The following are examples of a successful revision, a failed revision
and one excluded in calculating the success rates.
37
The Effects of Written Feedback in the Form of Recasts
Example (9) Successful
S9: I thought it was neccessary. ← Yes. It was necessary.
(Student’s revision) →I thought it was necessary.
Example (10) Failed.
S10: I was belonged to the club. ← Oh, you belonged to the team.
(Student’s revision)→I was belonged to the team.
Example (11) Excluded
S11: I saw their swimming in the sea. ← You saw them swimming in the sea?
(Student’s revision)→My friends began swim in the sea.
In example 11, the student produced ill-formed output. However, it was not
counted as failed but excluded, because whether he noticed the recast or not is
unclear as he did not use the same structure in the revision.
One rater conducted categorization of written recasts. A week after the first
classification, the same rater conducted it again. This method of classification
follows Alderson et al. (1995), which explains that multiple rating sessions
increases the reliability of the rating. Where there were discrepancies between the
two ratings (4 cases), a second rater was invited to rate them. After discussion, the
disagreement was resolved.
Writing accuracy, fluency and complexity
To measure accuracy, the proportion of the number of T-units without lexical and
grammatical errors in the total number of T-units in the writing was calculated. This
means that the denominator was the number of T-units and the numerator was the
number of T-units which did not include lexical and grammatical errors. This
measure was taken in accordance with previous studies (e.g., Sato, 2008;
Wolfe-Quintero, Inagaki, & Kim, 1998). I decided not to count errors related to the
usage of articles as they present difficulties even for proficient learners. In
measuring fluency, the number of words written in the essay was counted for the
38
The Journal of Asia TEFL
current study, as is rationalized in Wolfe-Quintero et al (al.) (1998) and Ellis and
Barkhuizen (2005). As for measuring complexity, it was decided to measure the
mean length of T-unit by calculating the average number of words per T-unit
following previous studies such as Ortega (2003) and Wolfe-Quintero et al (1998).
In the relation of the three aspects, it was decided to examine whether or not there
are trade-offs between accuracy and fluency in writing referring to Ellis (2003) and
Sato (2008).
RESULTS
Success Rates of Recasts According to Types.
In total, 125 written recasts were recorded. Grammatical recasts were made 81
times: lexical recasts, 17 times; recasts to unsolicited use of L1, 11 times; recasts to
spelling error, 6 times; content recasts, 10 times. As for the lengths, long recasts
were recorded 81 times and short recasts were made 44 times. About the degree of
differences, single change recasts were recorded 48 times, and multiple change
recasts were recorded 77 times. Table 1 summarizes the number of recasts, repairs,
failed revisions, avoided revisions and success rates according to error types; Table
2 shows the results according to the differences; Table 3 shows the results according
to the lengths.
TABLE 1
Number of Recasts, Repairs, Failed Revisions, Avoided Revisions and
Success Rates by Error Types
Types
Frequency
Repair
Failed
Avoided
Success rate
Grammatical
Lexical
81
17
43
10
31
5
7
2
58%
60%
L1 use
Spelling error
11
6
9
4
1
2
1
0
90%
67%
Contents
Total
10
125
4
70
4
43
2
12
50%
62%
39
The Effects of Written Feedback in the Form of Recasts
TABLE 2
Number of Single Change Recasts, Multiple Change Recasts, Repair, Failed
Revisions, Avoided Revisions and Success Rates
Types
Frequency
Repair
Failed
Avoided
Success rate
Single change
48
28
16
4
64%
Multiple
Total
77
125
42
70
27
43
8
12
61%
62%
TABLE 3
Number of Long and Short Recasts, Repair, Failed Revisions, Avoided
Revisions and Success Rates
Types
Frequency
Repair
Failed
Avoided
Success rate
Short
44
23
17
4
58%
Long
Total
81
125
47
70
26
43
8
12
64%
62%
To answer the first research question (RQ1. Are written recasts noticed, leading
Japanese EFL students to repair in the text revision?), the success rate was
calculated. 70 written recasts successfully led students to repair and 43 failed. In 12
cases students did not use the same structures, words or expressions to which the
recasts were given, and were excluded in calculating the success rates. The success
rate calculated was 62%. To examine whether there was a statistical difference
between the number of repairs (70) and failed revisions (43), a chi-square statistic
with Yates’ continuity correction was calculated, finding a significant difference
between them. (χ2= 6.45, df = 1, p=.011<.05).
The second research question asked about the success rates of written recasts
according to their types (RQ2. How effective are written recasts for Japanese EFL
students according to error type?). Among 125 written recasts in total, grammatical
recasts occurred 81 times with a 58% success rate, lexical recasts occurred 17 times
with a 60% success rate, L1 use recasts occurred 11 times with a 90% success rate,
spelling error recasts occurred 6 times with a 67% success rate, and contents recasts
occurred 10 times with a 50% success rate. A chi-square statistic with Yates’
40
The Journal of Asia TEFL
continuity correction revealed there was a significant difference between the
number of repair and failed revisions only in L1 use (χ2= 6.40, df =1, p=.011
<.05).
The third research question concerned the success rates according to the degree
of difference between the learner’s initial writing and written recasts (RQ3. How
effective are written recasts for Japanese EFL students according to the degree of
differences?). It was reported that single change recasts occurred 48 times with a
64% success rate and that multiple change recasts occurred 77 times with a 61%
success rate. A chi-square statistic test with Yates’ continuity correction revealed
that there was not a statistically significant difference in the success rates between
single change recasts and multiple change recasts.
The fourth research question asked whether there is a difference in the success
rates according to the lengths of written recasts (RQ4. How effective are written
recasts for Japanese EFL students according to length?). It was recorded that short
recasts with five morphemes or less occurred 44 times with a 58% success rate and
that long recasts with more than five morphemes occurred 81 times with a 64%
success rate. A chi-square statistic test with Yates’ continuity correction revealed
that there was not a statistically significant difference in the success rates between
short recasts and long recasts.
Accuracy, Fluency and Complexity
The fifth research question concerned the development of accuracy, fluency and
complexity from the first draft to the second draft (RQ5. Do written recasts
contribute to the development of accuracy, fluency and complexity in the text
revision?). Mean scores of accuracy, fluency and complexity improved. The results
of t-tests indicated that there was a statistically significant difference with a large
effect size between the first draft and the second draft in fluency ( t (24)=-6.55,
p=.00<.05, r =.80) and complexity ( t (24)=-4.53, p=.00<.05, r = .68). However,
that was not the case in accuracy ( t (24)= -1.1, n.s. r = .22). Table 4 summarizes
the results.
41
The Effects of Written Feedback in the Form of Recasts
TABLE 4
Mean Scores and Sds of Accuracy , Fluency and Complexity in the First and
Second Drafts
Accuracy
First draft
n
25
Second draft 25
Fluency
Complexity
Mean
59.8
SD
13.7
Mean
120.3
SD
41.2
Mean
8.2
SD
1.2
62.8
15.3
139.1
40.1
8.7
1.4
Note. Accuracy score is the proportion (%) of the number of T-units without lexical and
grammatical errors in the total number of T-units in the writing. Fluency score is the
number of words written in the essay. Complexity score is the average number of words
per T-unit
Research question 6 concerned trade-offs of accuracy and fluency in writing
(RQ6. Do corrections of accuracy with fluency in the writing change between the
first draft and the revised draft after being provided written recasts?). To
investigate the correlation of accuracy with fluency in writing, we examined
Pearson’s correlation coefficients. As table 5 shows, in the first writing, little
correlation was observed, but in the second writing, a significant weak correlation
was observed.
TABLE 5.
Pearson’s Correlation Coefficients of Accuracy with Fluency
First draft
Second draft
n
Pearson’s correlation
25
25
.148
.380*
Correlation is significant at the .05 level (2-tailed)
DISCUSSION
Success Rates of Recasts According to the Type
42
The Journal of Asia TEFL
It was found that students were more likely to show repair to written recasts than
not. A 62% success rate is higher than those of oral recasts examined previously
(e.g., Lyster & Ranta, 1997). It is assumed that written recasts can be more effective
than oral recasts in leading learners to repair their erroneous production. Eleven
students mentioned in the free comments that they found they had made mistakes
or errors when they read recasts and revised their work. Six students wrote in the
comments that they found they had not yet learned the correct grammatical
knowledge about the structures and that by reading the recasts they learned the
grammatical rules correctly. This can imply that written recasts activated students’
previously existing learned systems and partially acquired knowledge. In addition,
in some cases, it can be interpreted that written recasts provided students with
opportunities to learn structures they had not yet learned. Four students wrote that
by reading the recasts they learned how to write what they had wanted to write but
couldn’t. In the situation, the students were able to fill the gap between what they
wanted to write and what they actually wrote after referring to the written recasts.
Variation of success rates by the error types suggests the effectiveness of recasts
varies according to the type. Recasts to L1 use recorded the highest success rate,
then recasts to spelling errors, followed by lexical recasts, grammatical recasts and
content recasts ( L1, 90%> S, 67%> Lex, 60%> G, 58% > C, 50%). This is
similar to Sato (2009) which compared the effectiveness of oral recasts according to
the types. As the reason that L1 recasts led to the highest success rate, we can point
out its salience as is the case in oral recasts (Sato, 2009). In the first draft when
students used the Japanese language, they found interlanguage deficiency or a lack
of knowledge. In reading recasts it was easy for them to find the English
equivalents of what had been written in Japanese, so that they could use them just
by changing the Japanese to English. The lowest success rate of content recasts can
be attributed to their vagueness. Although content recasts had corrective intent,
students often must have perceived them as confirmation, paraphrasing or just brief
comments.
Example (12) Content recasts
S12: I played tennis every day, six days a week.
43
The Effects of Written Feedback in the Form of Recasts
← Oh, You played tennis very hard, almost every day.
(There was no revision from the student.)
In this case, we can assume that the student perceived the recasts not as
corrective feedback but as a brief comment from the teacher, as is evidenced by his
comment that he could not understand the purpose of some written recasts.
Grammatical recasts also recorded a lower success rate (58%) than the total success
rate (62%). This result is compatible with previous studies on oral recasts (e.g., Kim
& Han, 2007; Sato, 2009; Trofimovich, Ammar, & Gatbonton, 2007; Williams,
1999). Trofimovich et al. (2007) suggested that students should already have
knowledge of the form so that they can notice their own errors through recasts and
to reformulate them after recasts. This argument could be partially applied to
written recasts in this study. It can be assumed that some students could not show
repair because of a lack of knowledge on the form to benefit from the enhancing
effect of recasts, even though they were not oral but written.
The results showed that effects of written recasts were not affected by the length
or the degree of difference between the learner’s initial writing and the written
recast. One unexpected finding is that long recasts recorded a higher success rate
than short recasts. These findings are incompatible with previous studies (e.g.,
Philip, 2003; Sato, 2009) which showed shorter and single change oral recasts are
better noticed, leading to learners’ repair. As to the possible reasons for this, it can
be interpreted that even written long recasts with multiple changes can let learners
repair their previous erroneous production regardless of their limitation of working
memory. Unlike oral feedback, written feedback was not fleeting and thus enabled
students to trace it back by reading (Payne & Whitney, 2002). This nature of written
feedback was beneficial enough for students with shorter working memory or
difficulty in maintaining information in memory to utilize not only short-single
change recasts but long multiple change recasts. A student’s comment that she
repeatedly read the written recasts to repair her errors confirms this interpretation.
44
The Journal of Asia TEFL
Accuracy, Fluency and Complexity
From the first draft to the second draft, accuracy, fluency and complexity
developed in the writing. This is consistent with the argument that written
corrective feedback is beneficial in editing in revision (e.g., Ferris & Roberts, 2001;
Truscott, 1996, 1999). In the second writing, some of them wrote additional
information in the revision, contributing to the statistically significant development
in fluency. This smaller cognitive load (in that they did not have to think about what
to write) also encouraged them to write more complex sentences, resulting in
statistically significant development in complexity. However, in writing additional
information and complex sentences, errors and mistakes naturally occurred. This
can be the reason that the development of accuracy was not statistically significant.
Another possible reason why increases in accuracy were not significant is the small
number of recasts given to each student.
As for the trade-off effect between accuracy and fluency, the correlation between
the two was little in the first draft, but a significant weak correlation was observed
in the second draft. This means that the extent of trade-offs decreased with the
correlation of accuracy and fluency improving in the second draft. This is
compatible with Ruegg (2010). A higher correlation of the two in the second draft is
thought to be due to students’ making revisions while referring to the written recasts
contributing to the development of both accuracy and fluency. That is to say,
better-balanced writing. One student mentioned in the comment that she used
written recasts as models and also added other expressions in the revision of
original erroneous production. It can be assumed that written recasts provided the
opportunity where students could repair their errors and mistakes, sometimes with
additional words and expressions as the following example shows.
Example (13) Grammatical recast
S13: I enjoyed sing song. ← You enjoyed singing songs.
→(Student’s revision) I enjoyed singing songs with children this summer.
It is assumed that this type of revision resulted in a higher accuracy and fluency
score and a higher correlation between the two.
45
The Effects of Written Feedback in the Form of Recasts
CONCLUSION
This study implied that written feedback in the form of recasting is beneficial for
learners to notice their errors or mistakes leading them to repair. The quality of the
writing from the points of accuracy, fluency and complexity would improve in the
second draft written with the help of feedback. The findings reported are suggestive,
in that as recasts are not intrusive as explicit correction (e.g., Doughty, 2001; Lyster,
2007), we can utilize the beneficial effects of written recasts for Japanese learners
who often have difficulty writing essays in English.
However, a number of limitations should be noted. The sample size was small
with just 25 national university students. Ideally we should have more participants
with different English proficiency levels so that we can make general productive
claims. In providing written feedback, there was no systematic rule: Written recasts
were provided randomly, mainly depending on the teacher’s intuition so that
students could use them to improve the overall quality of the writing in the second
draft. Some grammatical features or structures are more teachable or treatable than
others (Xu, 2009). Depending on which structures are targeted, the success rate of
repair can be different. In addition, the number of recasts given to students was
different. It can be argued that students tend to correct only the parts to which
recasts are given. Finally, the lack of a control group, in which students are not
given any feedback or given different types of feedback, is an important limitation.
Further investigation with experiment design would make the findings more
convincing. This study implied the effectiveness of written feedback in the form of
recasts in text revision. However, to confirm the findings of the study, a further
study with more samples in a controlled setting and with a control group is needed.
THE AUTHOR
Rintaro Sato is associate professor in the Department of English Education at
Nara University of Education. His research interests include intake and output
processing, feedback and negotiation of meaning. His recent publications include
46
The Journal of Asia TEFL
Reconsidering the effectiveness and suitability of PPP and TBLT in the Japanese
classroom (2010) and Self-initiated self-repair attempts by Japanese high school
learners while speaking English (2012).
Email: [email protected]
REFERENCES
Alderson, J. C., Clapham, B. J., & Wall, D. (1995). Language test construction and
evaluation. Cambridge: Cambridge University Press.
Ayoun, D. (2001). The role of negative and positive feedback in the second language
acquisition of the passé compose and the imparfait. The Modern Language Journal,
85(2), 226–243.
Ayoun, D. (2004). The effectiveness of written recasts in the second language acquisition of
aspectual distinctions in French: a follow-up study. The Modern Language Journal,
88(1), 31–55.
Bitchener, J. (2008). Evidence in support of written corrective feedback. Journal of Second
Language Writing, 17(1), 102-118.
Coffin, C., Curry, M. J., Goodman, S., Hewings, A., Lillis, T. M., & Swann, J. (2003).
Teaching academic writing: A toolkit for higher education. New York: Routledge.
Doughty, C. (2001). The cognitive underpinnings of focus on form. In P. Robinson (Ed.),
Cognition and second language instruction (pp. 206-257). Cambridge: Cambridge
University Press
Doughty, C., & Varela, E. (1998). Communicative focus on form In C. Doughty & J.
Williams (Eds.), Focus on form in classroom second language acquisition
(pp.114-138). New York: Cambridge University Press.
Ellis, R. (1987). Interlanguage variability in narrative discourse: Style shifting in the use of the
past tense. Studies in Second Language Acquisition, 9(1), 1-12.
Ellis, R. (2003). Task-based language learning and teaching. Oxford: Oxford University
press.
Ellis, R., & Barkhuizen, G. P. (2005). Analyzing learner language. Oxford: Oxford University
Press.
Ellis, R., Sheen, Y., Murakami, M., & Takshima, H. (2008). The effects of focused and
unfocused written corrective feedback in an English as a foreign language context.
System, 36(3), 337-352.
Ferris, D. (2004). The “grammar correction” debate in L2 writing: where are we, and where
do we go from here? (and what do we do in the meantime…?). Journal of Second
47
The Effects of Written Feedback in the Form of Recasts
Language Writing, 13(1), 49-62.
Ferris, D., & Roberts, B. (2001). Error feedback in L2 writing classes: How explicit does it
need to be? Journal of Second Language Writing, 10(3), 161–184.
Gass, S. M. (2003). Input and interaction. In C. Doughty & M. H. Long (Eds.), Handbook of
second language acquisition (pp. 224–255). Malden, MA: Blackwell.
Hartshorn, K. J., Evans, N. W., Merrill, P. F., Sudweeks, R. R., Strong-Krause, D., &
Anderson, N. J. (2010). Effects of dynamic corrective feedback on ESL writing
accuracy. TESOL Quarterly , 44(1), 84-109
Hunt, K. (1970). Do sentences in the second language grow like those in the first? TESOL
Quarterly, 4(3), 195-202.
Iwashita, N. (2003). Negative feedback and positive evidence in task-based interaction.
Studies in Second Language Acquisition, 25(1), 1-36.
Izumi, S. (2002). Output, input enhancement, and the noticing hypothesis: An experimental
study on ESL relativization. Studies in Second Language Acquisition, 24(4), 541–577.
Kepner, C. (1991). An experiment in the relationship of types of written feedback to the
development of second-language writing skills. The Modern Language Journal, 75(3),
305-313.
Kim, J-H., & Han, Z-H. (2007). Recasts in communicative EFL classes: Do teacher intent and
learner interpretation overlap? In A. Mackey (Ed.), Conversational interaction in
second language acquisition (pp. 269-297). Oxford: Oxford University Press.
Leeman, J. (2003). Recasts and second language development. Studies in Second
Language Acquisition, 25(1), 37-63.
Loewen, S. (2004). Uptake in incidental focus on form in meaning-focused ESL lessons.
Language Learning, 54(1), 153-188.
Loewen, S., & Philip, J. (2006). Recasts in the adult English L2 classroom: Characteristics,
explicitness, and effectiveness. The Modern Language Journal, 90(4), 536-556.
Loewen, S., & Nabei, T. (2007). Measuring the effects of oral corrective feedback on L2
knowledge. In Mackey, A. (Ed.), Conversational interaction in second language
acquisition (pp. 361-377). Oxford: Oxford University Press.
Long, M. (1996). The role of the linguistic environment in second language acquisition. In W.
Ritchie & T. Bhatia (Eds.), Handbook of Second Language Acquisition (pp. 413-468).
San Diego: Academic Press.
Long, M. (2006). Problems in SLA. Mahwah, NJ: Erlbaum.
Lyster, R. (1998a). Recasts, repetition and ambiguity in L2 classroom discourse. Studies in
Second Language Acquisition, 20(1), 50-81.
Lyster, R. (1998b). Negotiation of form, recasts, and explicit correction in relation to error
types and learner repair in immersion classrooms. Language Learning, 48(2),
183-218.
48
The Journal of Asia TEFL
Lyster, R. (2004). Differential effects of prompts and recasts in form-focused instruction.
Studies in Second Language acquisition, 26(3), 399-432.
Lyster, R. (2007). Learning and teaching languages through content. Amsterdam: John
Benjamins.
Lyster, R., & Ranta, L. (1997). Corrective feedback and learner uptake: Negotiation of form
in communicative classroom. Studies in Second Language Acquisition, 19(1), 37-66.
Muranoi, H. (2000). Focus on form through interaction enhancement: Integrating formal
instruction into a communicative task in EFL classrooms. Language Learning, 50(4),
617-673.
Nugrahenny, T. Z. (2007). Teacher and students attitudes toward teacher feedback RELC
Journal, 38(1),38-52.
Oliver, R. (1995). Negative feedback in child NS-NNS conversation. Studies in Second
Language Acquisition, 17(4), 459-481.
Ortega, L. (2003). Syntactic complexity measures and their relationship to L2 proficiency: A
research synthesis of college–level L2 writing. Applied Linguistics, 24(4), 492–518.
Panova, I., & Lyster, R. (2002). Patterns of corrective feedback and uptake in an adult ESL
classroom. TESOL Quarterly, 36(4), 573–595.
Payne, J. S., & Whitney, P. J. (2002). Developing L2 oral proficiency through sychronous
CMC: Output, working memory, and interlanguage development. CALICO Journal,
20(1), 7-32.
Philip, J. (2003). Constraints on “Noticing the Gap”: Nonnative Speaker’s Noticing of
Recasts in NS-NNS Interaction. Studies in Second Language Acquisition, 25(1),
99-123.
Ruegg, R. (2010). Who wants feedback and does it make any difference? In A. M. Stoke
(Ed.), JALT 2009 Conference Proceedings. Tokyo: JALT.
Sato, R. (2008). The relationship between fluency and accuracy in writing. ARELE, 19, 71-80.
Sato, R. (2009). Considering the effectiveness of recasts on Japanese high school learners’
learning. The Journal of Asia TEFL 6(4), 193-216.
Shehadeh, A. (2002). Comprehensible output, from occurrence to acquisition: An agenda for
acquisitional research. Language Learning, 52(3), 597–647.
Sheppard, K. (1992). Two feedback types: Do they make a difference? RELC Journal, 23(1),
103-110.
Skehan, P. (1996). A framework for the implementation of task-based instruction. Applied
Linguistics, 17(1), 38-62.
Skehan, P., & Foster, P. (1999). The influence of planning and post-task activities on accuracy
and complexity in task based learning. Language teaching Research, 3, 215-247.
Swain, M. (1985). Communicative competence: Some roles of comprehensible input and
comprehensible output in its development. In S. Gass & C. Madden (Eds.), Input in
49
The Effects of Written Feedback in the Form of Recasts
second language acquisition (pp. 235–253). Rowley, MA: Newbury House.
Tong-Frederics, C. (1984). Types of oral communication activities and the language they
generate: a comparison. System, 12(2), 133-145.
Trofimovich, P., Ammar, A., & Gatbonton, E. (2007). How effective are recasts?: The role of
attention, memory, and analytic ability. In A. Mackey (Ed.), Conversational
interaction in second language acquisition (pp. 171-195). Oxford: Oxford University
Press.
Truscott, J. (1996). The case against grammar correction in L2 writing classes. Language
Learning, 46(2), 327–369.
Truscott, J. (1999). The case for ‘‘the case for grammar correction in L2 writing classes’’: A
response to Ferris. Journal of Second Language Writing, 8(2), 111–122.
Varnosfadrani, A. D., & Basturkmen, H. (2009). The effectiveness of implicit and explicit
error correction on learners‘ performance. System, 37(1), 82-98.
Williams, J. (1999). Learner–generated attention to form. Language Learning, 49(4),
583-625.
Wolfe-Quintero, K., Inagaki, S., & Kim, H. (1998). Second language development in writing:
Measures of fluency, accuracy, & complexity. Honolulu: University of Hawaii Press.
Xu, C. (2009). Overgeneralization from a narrow focus: A response to Ellis et al. (2008) and
Bitchener (2008). Journal of Second Language Writing, 18(4), 270-275.
Yashima, T. (2000). Orientations and motivation in foreign language-learning: A study of
Japanese college students. JACET Bulletin, 31, 121-133.
50