THE EFFECTIVENESS OF COLOR AND COMMENT ERROR MARKER IN PROVIDING UNFOCUSED-IMPLICIT CORRECTIVE FEEDBACK FOR
UNDERGRADUATE STUDENTS’ WRITING GRAMMAR
(Submitted as partial requirement for SarjanaPendidikan degree)
RIZA PURNAMA 0906889
ENGLISH EDUCATION DEPARTMENT
THE EFFECTIVENESS OF COLOR AND
COMMENT ERROR MARKER IN
PROVIDING UNFOCUSED IMPLICIT
CORRECTIVE FEEDBACK FOR
UNDERGRADUATE STUDENTS’
WRITING GRAMMAR
Oleh RizaPurnama
Sebuah skripsi yang diajukan untuk memenuhi salah satu syarat memperoleh gelar Sarjana pada Fakultas Pendidikan BahasadanSeni
© RizaPurnama 2014 Universitas Pendidikan Indonesia
April 2014
Hak Cipta dilindungi undang-undang.
TABLE OF CONTENTS
Page of Approval ………. i
Statement of Authorization ………..……….. ii
Preface ……… iii
Acknowledgement ………. iv
Abstract ………. v
Table of Contents ……… vi
List of Figures ……… viii
List of Tables ………. ix
List of Appendix ……… x
Chapter I: Introduction I. Background of the Research ……….. 1
II. Statement of the Problems ……….. 2
III. Objectives of the Research ………. 3
IV. Significance of the Research ………... 3
V. Clarification of Key Terms ………. 3
VI. Organization of the Paper ……… 6
Chapter II: Literature Review I. Corrective Feedback & L2 (FL) Writing ……… 6
II. English Grammar in Indonesia’s L2 Writing ……….. 10
III. Corrective Feedback Amount: Unfocused Corrective Feedback ………. 14
IV. Corrective Feedback Strategy: Implicit Corrective Feedback ………. 18
V. Indonesia’s Adult Second Language Learner ……….. 20
Chapter III: Research Methodology I. Research Design ……….. 23
III. Operational Definition of Research Variables ………. 28
IV. Research Instruments ……….. 31
V. Data Collection ……… 32
VI. Data Analysis ……….. 36
Chapter IV: Findings and Discussion I. Research Findings ……… 41
II. Discussion ……… 62
Chapter V: Conclusion and Suggestion I. Conclusion ……… 67
II. Suggestion ……… 68
References
ABSTRACT
THE EFFECTIVENESS OF COLOR AND COMMENT ERROR MARKER IN PROVIDING UNFOCUSED IMPICIT CORRECTIVE FEEDBACK FOR
UNDERGRADUATE STUDENTS’ WRITING GRAMMAR
RizaPurnama (0906889)
Main Supervisor: PupungPurnawarman, M.S.Ed., Ph.D.
Co-supervisor: Muhammad Handi G. S.Pd., M.Pd.
Regardless being considered potentially ineffective by many (Lee, 2013), the provision of unfocused-implicit corrective feedback (CF) is considered an ideal way of addressing real-teaching problem (Ferris, 2010; Van Beuningen, 2010). This research was aimed at comparing the effectiveness of color and comment as error-marker in unfocused-implicit CF provision. The design used in this research was quasi-experimental design. Involving 64 participants, the data were collected through controlled observation and the observation of documents from10 randomly chosen participants. The findings had suggested significant difference between the two error-marking techniques where the use of comment was better than the use of color (sig 2-tailed = 0.011, p <.05). Such finding was further validated given that the use of comment was revealed to have facilitated better assistance towards revision of different type of errors (97.67%). In addition, it was also revealed that there were some circumstances where the provision of both strategies might have contributed to futile revisions. However, despite being outperformed, the use of color as error marker also resulted in noteworthy revising success which took up to 94.13%.
ABSTRAK
EFEKTIFITAS PENGGUNAAN WARNA DAN KOMENTAR DALAM MEMBERIKAN UMPAN BALIK MENYELURUH DAN TERSIRAT UNTUK TATA BAHASA TULISAN
MAHASISWA S1
RizaPurnama (0906889)
DosenPembimbing I: PupungPurnawarman, M.S.Ed., Ph.D.
DosenPembimbing II: Muhammad Handi G. S.Pd., M.Pd.
Meskipundianggaptidakefektifolehsebagianbesar orang (Lee, 2013), pemberianumpanbaliksecaramenyeluruhdantersirat (unfocused-implicit corrective feedback) dianggapsebagaijalan yang ideal untukmenyikapiberbagaimasalahpengajaran yang
nyataterjadi (Ferris, 2010; Van Beuningen,
2010).Penelitianiniditujukanuntukmembandingefektifitaspenggunaanwarnadankomentarseba gaipenandaletakkesalahandalampemberianumpanbalikmenyeluruhtersirat.Desain yang digunakandalampenelitianiniadalahquasi-experimental.Denganmelibatkan 64 partisipan, data didapatkanmelaluicontrolled ovservationsertapengamatanterhadapdokumendari 10 partisipan yang dipilihsecaraacak. Hasiltemuandaripenelitianinimenunjukkanperbedaan yang berartidiantarakeduacarapenandaankesalahandimanapenggunaankomentarmemberikanhasil yang lebihbaikdibandingkandenganpenggunaanwarna (sig 2-tailed = 0.011, p <.05). Temuaninilebihlanjuttervalidasiseiringtemuandimanapenggunaankomentarterbuktilebihefekti fdalammendampingipartisipanmelakukanrevisibergaijeniskesalahan (97.67%).Selainitu, jugaditemukanbahwaterdapatbeberapakeadaandimanapemberiankeduastrategijugabisaberakhi rdalamkegagalanrevisi.Namundemikian, terlepasdarinilai yang lebihkecil, penggunaanwarnasebagaipenandakesalahanjugamenghasilkankesuksesan proses revisi yang prosentasenyamencapai 94.13%.
Kata kunci:umpanbalikmenyeluruhdantersirat,
CHAPTER I INTRODUCTION
As an introductory part, elaboration of several points is presented in this chapter. The background of the research is firstly presented. Afterwards, the research question is listed out. Departing from the research background, too, the research objectives are listed. Furthermore, the research significance is elaborated as seen from theoretical and practical perspective. In providing limitation, the meaning of the key-terms is clarified.
I. Background of the Research
The application of corrective feedback (CF) as a mean of improving learner’s writing skill is not without intrigues. Currently, there has been contradicting study-findings causing vagueness on how the CF should be provided to achieve maximum result. As suggested by Lee (2013), who has summarized the results of CF studies around the globe, to selectively choose type of error to be treated is regarded as the best way in providing the feedback. It is fair to say that such assumption is justifiable. It has been suggested that process of revision is a cognitively demanding process requiring heavy demand on working memory (Kellog, 1994). It has also been suggested that each individual possesses different capacity of working memory (Dehn, 2011; Gathercole&Alloway, 2008). Especially in Indonesia, where learners attain inadequate amount of English exposure, to limit the scope of error to be treated in a CF provision is a wise choice. Nevertheless, such way of CF provision is considered unrealistic (Ferris, 2010; VanBeuningen, 2010). They argue that in reality, learners make broad range of errors thus to selectively choose errors to be treated only suits laboratory setting.
indirect ones. However, some studies had also proven that implicit/ indirect feedback, where an error is only marked without the provision of correct linguistic form, is more fruitful (e.g Ferris &Helt, 2000 in Bitchener&Knoch, 2008). Chandler (2003), in accordance, has revealed that implicit feedback is the strategy preferably chosen by students over explicit one implying positive attitude towards the strategy. However, the study done by Ferris &Helt was conducted by involving 92 ESL learners in American university where exposure on English attained by the participants was adequate. To claim such notion in Indonesian context needs further verification. Myhill& Jones (2010) have suggested that most
EFL learners faced difficulty in figuring out what to do with the error within their writing even when the location has already been identified. Given inadequate amount of English exposure to be utilized as reference, the provision of implicit CF may not be a wise choice to assist Indonesian learners in assisting learners’ revising process towards a grammatically-accurate piece of writing.
An Indonesian CF study done by Purnawarman (2011) has suggested that the best way in providing CF to Indonesia’s adult learners (university student) is through combination of both implicit and explicit strategies. In his study, Purnawarman utilized color and comment as error-marker during the provision of corrective feedback. However, there was a strict limitation of the errors that were being treated (focused CF) in this referred study. Thus, there is vagueness on whether or not these error-marking strategies will similarly result positively in the provision of an unfocused and implicit corrective feedback.
II. Statement of Problem
Referring to the elaboration above, to combine unfocused CF and implicit CF for Indonesian learners seems to be preposterous. However, a distinction
(comprehensive) as suggested by Ferris (2010) and VanBeuningen (2010), such assumption on unfocused-implicit CF potential ineffectiveness should be put into test. By scrutinizing the effectiveness of two error-marking strategies, color and comment, such assumption is attempted to be put into test. Applying sequential-explanatory mixed-method design, the research question is formulated as follow:
A. Is there any significant difference between the use of color and comment error marker in assisting participants’ grammar-revising progress?
B. Under what circumstance does the use of the two strategies end up in futile revisions?
III. Objectives of the Research
This research is aimed at comparing the effectiveness of two types of unfocused-implicit CF techniques both quantitatively and qualitatively. The comparison is focused on the effect towards participants’ revising progress.
IV. Significance of the Research
Two aspects, namely theoretical and practical aspects, are comprised as the significance of this current research. From theoretical aspect, the findings of this research are expected to provide further reference on CF discussion. More specifically, the findings of this research are expected to provide further validation towards the assumption suggesting the combination of unfocused and implicit CF to be potentially ineffective for adult learners by scrutinizing the effectiveness of color and comment error-marker in its provision. Seen from practical perspective, the findings of this research was intended to expand the guidance on how to and how not to provide corrective feedback for Indonesia adults learning to write in English.
V. Clarification of Key Terms A. Writing corrective feedback
provision of comment or response towards learners’ writing which is aimed at assisting learners’ revising process. Referring to Hayes’ current model of writing (2012) (see Figure 1, p.9), the evaluator and collaborator is the researcher himself, a final semester student of English Education Department, Indonesia University of Education; whereas the proposer, translator and transcriber is students enrolled in foundation of English grammar (FEG) class academic year 2012/2013. Thus, referring to O’brien’s classification of teacher and peer feedback (2004), it can be said that in this research, the feedback given is categorized under teacher-feedback category given that the feedback is given
under the supervision of the FEG-course lecturer.
B. Unfocused corrective feedback
The term unfocused corrective feedback used in this research is synonymous to the term comprehensive feedback as suggested by Lee (2013) where such CF is defined as the response to every single error identified within learners’ writing. Nevertheless, it has to be noted that the term unfocused CF here is limited only to grammatical aspects. In other word, other errors in terms of contents, mechanics and other ungrammatical errors are neglected. One feature distinguishing it from focused CF relies on the absence of limitation on what and how many grammatical errors should be marked in learners’ writing. In addition, there is also no categorization of treatable and untreatable errors.
C. Implicit corrective feedback
The term implicit CF used in this research is synonymous to the term indirect CF used in other CF researches. Simply put, the term implicit CF is defined as a strategy of providing feedback by indicating an error without providing the correct form in assisting a learner correcting his error (Ferris &
D. Revising progress
It has to be noted that the term revising progress here doesn’t refer to the improvement of learners’ writing score as measured by tests. The term revising progress refers to the percentage-value of error of each learner during three drafting process (rough draft/ original draft, revised draft and final draft). A grammatical error is considered successfully revised when the reviser have made an accurate revision on the respective error through different means of revision resulting in a coherent clause. On the other hand, a grammatical error is considered unsuccessfully revised when the reviser either fails at performing
accurate revision towards the marked item or performs unnecessary omission of the marked item.
VI. Organization of Paper
CHAPTER III
RESEARCH METHODOLOGY
In this chapter, the research methodology applied in this research is comprehensively elaborated. Firstly, the elaboration starts with the application of sequential-explanatory mixed-method comprising informal experimental and descriptive design. Then, a comprehensive elaboration on sampling design which includes research sampling method as well as the determination of representative sample size is elaborated. Afterwards, operational definitions of the research variables are pointed out. Description on the research instrument, data collection
and data analysis follows afterwards.
I. Research Design
A. Quasi-experimental design
Quasi-experimental design was applied in accordance to one of the research objectives namely to test a hypothesis on degree of CF implicitness, from the two error-marking strategies, to affect revising-progress. As defined by Rasinger(2010, p. 59) and Kothari (2004, p. 39), experimental design (or also be
referred as hypothesis-testing research) is a type of study which involves the manipulation of variables in order to test hypotheses. Given that the manipulation of treatment (independent variable) was at the very core of the journey towards the scrutiny of effect-comparison between two unfocused-implicit CF techniques, to apply an experimental design was thus a necessity. Regarding respective research design, Kothari breaks down the term into more specific categorization. It can be said that such categorization was based on the complexity, existence of control group and the randomization of sample-drawing (2004, p. 41). Despite being considered the least complex and the most likely to be interfered by extraneous variable amongst other experimental-research types, before-and-after without control experimental research-design was chosen.
single group is measured prior to the provision of treatment and after the provision of research treatment (Kothari, 2004). Modification, however, was made in accordance to the research objective namely to statistically compare the two types of unfocused-implicit CF techniques in terms of treatment effect on grammar-revising progress. Thus, further definition on experimental unit was made (see figure 4). Instead of only involving one experimental group, two groups receiving different type of treatments were involved.
Figure 3. Before-and-after without control informal-experimental design scheme (taken from Kothari (2004. p.41)
Figure 4. Modification on Kothari‟s Before-and-after without control informal-experimental design.
To be more specific, the modification of Kothari‟s scheme on before-and-after without control experimental design comprised two major things. Firstly, instead of only involving one experimental group, two experimental groups were
involved. The treatment, unfocused-implicit CF technique, given to both group
Level of
were different each other. Secondly, instead of being provided only once, the treatment provision was extended into two provisions. The type of follow-up CF (the second CF) given was similar to the type of CF given in the first provision hence each experimental group only received one type of CF during the experiment.
1. The application of experimental design
Referring to Hayes‟ model of writing (figure 1, p.9), there are three levels comprised within a writing process. They are control level, process level and resource level. The followings are the clarification of how the experiment was conducted following Hayes‟ writing model.
In control level, the elements that were included were motivation, goal setting, and current plan. It has to be noted that this research was conducted in two Foundation of English Grammar (FEG) classes in English Department, Indonesia University of Education. Following the allowance from the instructor to conduct the experiment in her class, drafting tasks were designed where the final draft would be assessed and the score would be included in their final FEG-course score. Hence, it is fair to say that the motivation behind the participations of the participants was to attain an additional score to increase their final score. During the introduction of the research, the researcher socialized the steps the participants were going to undertake in attaining such score.
The second level comprised in Hayes‟ latest writing model is process level which is divided into two sub-levels, writing processes and task environment. Within the writing-process sub layer, the role of evaluator was taken by the researcher himself under the supervision of the FEG-course instructor. Hence, it is fair to say that the CF provided fell under the category of teacher feedback.
the supervision of FEG-course instructor. In terms of transcribing technology, word-processing software was utilized given that the submission of drafts and the provision of corrective feedback were done through internet interaction. In task material, written plans, the proposers (participants) were given a set of pictures in assisting them planning the writings they were about to produce. Lastly, regarding the text written so far part of Hayes‟ writing model, the term revised draft and final draft used in this research represent the product of transcription done by the participant (transcriber) after going through evaluation phase (provision of unfocused-implicit CF) and translation phase (making sense of the error-marker
implicitness).
The last level in Hayes‟ writing model is resource level which comprises attention, working memory, reading and long-term memory. It has to be noted that the elements of attention, working memory and long-term memory were not measured prior to the execution of experiment and hence being considered extraneous factor. The reading element, on the other hand, was being deliberately manipulated given that all participants were given similar material in the form of a set of pictures in order to assist them planning and generating ideas for their writing.
II. Sampling Design A. Research site
This research took place in English Education Department, Faculty of Language and Arts Education, Indonesia University of Education. More specifically, the researcher focused on the English-majoring undergraduate students who were taking Foundation of English Grammar (FEG) course in the academic year of 2012/ 2013.
B. Research population
similar amount of grammatical-knowledge input from FEG course when the
experiment took place thus the students possess a more relatively homogenous
grammatical knowledge.Cargan (2010) suggests that the purpose of the study is an essential determining factor in deciding the size of research-population (2007, p. 236). Considering the research being gearedtowardsthe scrutiny of the effect resulted from two types of error-markers in unfocused-implicit CF provision, the involvement of two FEG classes was considered adequate as a representation. In addition, given the circumstances, the lecturer external-factor was taken into
consideration given that the other two FEG classes were taught by different
lecturers.
Referring to Kothari (2004, p. 56), to know exactly whether the population (universe) is finite or infinite is essential given that such nature determines how the sample is chosen. Having finiteness of the population being determined by the measurability of the item-quantity (Kothari, 2004, p. 54), the population chosen in this research was labeled finite. The total student enrolled in the course was 92 students comprising 75 freshmen (81,5%), two sophomores (2,2%), eight third-year students (8,7%) and seven fourth-third-year students (7,6%).
C. Research sampling
1. Representative sampling size and sampling method
The number of participantsto be involved in the experiment was determined by an approach which was based on confidence level and precision rate. In this research, the confidence level was set at 95% along with precision-rate value which was set at .05. The decision on setting the precision-precision-rate value at .05 was based on the notion that such value could generate conservative sample size as suggested by Kothari (2004, p. 180).
obtained number of 10 was fair to be claimed as reliable given the formula had already taken precision rate and level of confidence into consideration.
Regarding the sampling method, non-probability sampling was used. Due to research limitation, immaturity in designing the research, the principle of randomization was absent in this research thus non-probability sampling-technique was applied instead. As defined by Kothari, such sampling sampling-technique allows the researcher to deliberately choose the individuals/ items to be taken as research sample, thus element of bias becomes inevitable (2004, p. 59).
It has to be noted that in the beginning of the experiment, all of the 92
students within the population were involved in the research and thus given research treatment. However, in the course of the experiment, 28 students did not finish all research tasks (see figure 8, p.39) leaving 64 students (69%) qualified to be involved in the analyses. Referring back to one of Kothari‟s criteria of reliable sampling design, the assurance that systematic bias is controlled (2004, p. 58), the quantity of the research sample was multiplied in pursuing Cargan‟s idea on greater quantity to strengthen research-finding quality (2010). Regardless 10 students being considered adequately representative to the chosen population, 54 more students were deliberately chosen resulting in the total of 64 participants to be further involved in this research. In the process of assigning the 64 participant into the two experimental groups (A and B), the research used the existing group based on the learning group (classes) determined by the institution.
III. Operational Definition of Research Variables A. Dependent Variable
1. Error reduction in the first revising attempt
As has already been mentioned in the clarification of key term section in Chapter I, the term revising-progress used in this research doesn‟t refer to the improvement of learners‟ writing score as measured by tests. The term revising progress refers to the percentage-value of error of each learner during three drafting process (rough draft/ original draft, revised draft and final draft). Given that the experimental counterpart of this research demands score under interval scale, the value of revising progress was obtained by calculating the percentage value of the number of error occurrence in each drafting process and the total
clause the participant had produced in the original draft into equation.
A grammatical error was considered successfully revised when the reviser had made an accurate revision on the respective error through different means of revision resulting in a coherent clause. On the other hand, a grammatical error was considered unsuccessfully revised when the reviser either failed at performing accurate revision towards the marked item or performed unnecessary omission of the marked item. For instance, in the clause “… as the national biggest annual
event called Eurovision Song Contest” taken from the essays written by
participant A1, the green error-marker implicitly indicated the need to reorder the words to form a proper noun-phrase modifier. In the revision progress, participant A1 omitted one of the words resulting in the clause as follow: “as the biggest annual event called Eurovision Song Contest”. Despite the clause being modified into a grammatically coherent one, in this research, the omission of the word
“national” was considered a failure in performing revision.
2. Error reduction in the final revising attempt
This second research variable takes the percentage of successfully revised
B. Independent Variable
1. Error-marking techniques
In distinguishing the degree of corrective feedback implicitness to be given
to the two experimental groups, a modification of error-marking strategies previously used in Purnawarman‟s study (2011) was utilized. Two types of error -marking strategies were used namely coloring and comment provision. Both of the error-marking strategies were implicit in nature given that no explicit
revised-revision was provided.
Regarding the coloring error-marking used as one of the unfocused-implicit CF techniques, several colors were used. This color error-marking was also used in Purnawarman‟s study on CF (2011). Taking the same research-setting, English Department – Indonesia University of Education, Purnawarman used three colors namely yellow for indicating article-error, bright green for indicating preposition-error and turquoise for indicating errors on past-tense verbs. Due to the scope of CF being unfocused, more colors were used during the CF provision. However, instead of providing different color for each category of error, the categorization was one-or-two-layer less-specified.
1.
Figure 5. Different colors used as error markers provided for experimental group A
It has to be noted that participants were not given adequate information on the meaning behind each color as error-marker. The instruction given to the
Syntactical error: absence of subject, verb, object; Jumbled order causing confusion.
Error on the use of possessive marker
participants was “some grammatical inaccuracies in your writing had been marked with different colors depending on the error-type”. Given that the participants were required to make sense of the meaning of each color given, it is fair to say that, the error-marking through provision of colors possess greater level of implicitness compared to the other unfocused-implicit CF strategy applied in other experimental group.
On the other hand, in this research, error-marking through provision of comment was considered attributed with milder level of implicitness. The utilization of MS-Word review/ comment feature was also applied in Purnawarman‟s study (2011). In contrast to Purnawarman‟s study in which review/ comment feature was used to provide explicit CF, the feature was used in providing explicit categorization of metalinguistic error-nature. It has to be noted even though the CF provider facilitated the participants with a more explicit guidance; the explicit expected-revision was not provided.
Figure 6. Example of comment as error markers provided for experimental group B
IV. Research instruments A. Participants‟ Essays
terms of error-occurrence frequency in participants‟ original draft, revised draft and final draft. In order to keep track of such frequency, a tally sheet was used. B. Ms-Excel 2010 worksheet as a tally sheet
Given that the expected raw data was in form of number of occurrence (frequency) instead of test score, no test was used as research instrument. Instead, a tally sheet was used to keep track of the occurrence of error found within each experimental phase. Ms-Excel 2010 worksheet was used to keep track of each participant‟s progress in order to put ease in the sorting and calculating process.
V. Data collection
It has to be noted that there were two types of data to be analyzed in this research. Firstly, the data was in the form of quantitative data depicting each reviser‟s success rate in revising grammatical error within their writing. Secondly, the data was in the form of records of reviser‟s essay comprising the original draft, revised draft and the final draft. The data, percentage value of grammar-revision success rate, were used in the experimental counterpart of the research whereas the essay records were used for the descriptive means. The qualitative data, participants‟ essay records, were collected sequentially after the course of experiment. For the quantitative counterpart, the raw data in forms of number of error occurrence as observed using tally sheet (MS-Excel 2010) was being converted into a percentage value by considering the number of error occurrence and total clause a participant had produced.
A. Data collection methods
to serve such mean. Controlled-observation method is defined as the way of observing a phenomenon through a pre-arranged plan in forms of the involvement of experimental procedure and instruments (Kothari, 2004, p. 96). MS-Excel 2010 worksheet was used to serve the function as a tally sheet in keeping track of the progress taking place within each reviser. Similar to the function of conventional use of tally sheet as a researcher instrument, the MS-Excel 2010 was used to measure the frequency of errors in the original draft, revised draft and final draft.
B. Data collection procedure
As has already been elaborated previously, the data for the experimental
half of the research was obtained through the application of controlled-observation method whereas the raw essay-records were used for the descriptive counterpart. The procedure in collecting the research data itself mostly follow the four stages of data processing as suggested by Kothari (2004, pp. 122-129).
1. Editing
The process of editing can be defined as the examination of raw data to detect errors and to correct them when possible (Kothari, 2004, p. 122). He further adds that “editing is done to assure that the data are accurate, consistent with other facts gathered, uniformly entered, as completed as possible and have been well arranged to facilitate coding and tabulation” (p. 123). Given the limitation on the provision of treatment, lack of inter-rater reliability, prior to the data analyses the accuracy of CF given was reevaluated. Either errors that were missed to be marked or errors that were unnecessarily marked were noted. Passing through the editing phase, out of 2.337 errors identified within the 64 chosen participants‟ essays, 1.701 error cases were rendered accurately marked and thus being taken as items to be analyzed.
2. Coding
classes (2004, p. 123). The coding phase, along with the editing process, was aimed at putting the data, error-cases, into two categories. Firstly, the error-cases which were accurately marked were put under the sample category whereas the error-cases which were either inaccurately marked or missed to be marked into non-sample category. From these evaluative phases, 1.701 error-cases were categorized under sample category out of 2.337 total error-cases.
This process was done by carefully reevaluating each participant‟s essay. Firstly, the sentences in the original draft were broken down into clauses and these clauses were being pasted into rows of MS-Excel 2010 worksheet. Afterwards, the renewed versions of clauses from participant‟s revised and final draft were also being pasted in line with the clauses from the original draft (see Appendix C). In cases where there were more than one grammatical errors within one clause, the respective clauses were pasted into another row resulting in each row only representing one error-case. Consequently, a clear view on how the revision took place was available in a three-columned table (see Appendix C). Having two programs running, MS-word file of marked participant‟s essay and MS-Excel worksheet, the error-markings were reevaluated in terms of accuracy. The accurately marked errors were labeled „1‟ whereas the missed or inaccurately marked errors were labeled „0‟. Using MS-Excel 2010 sort feature, the „0‟-labeled errors were being put together and thus could easily be removed from the worksheet. Afterwards, MS-Excel 2010 SUM formula was used in order to obtain the number of error-cases included as research sample.
3. Classification
Following the editing and coding process, the sample error-cases were further being classified into more specific categories. The research objective was
36 More specifically on the classification phase, 8 columns were prepared and labeled „1‟ to „8‟ representing each error-type. „1‟ for omission; „2‟ for double-markings; „3‟ for regularization; „4‟ for addition; „5‟ for miss formation; „6‟ for archi form; „7‟ for miss ordering and „8‟ for other types of error. Having each row representing one error case, the classification was done by typing the number „1‟ in the column corresponding to the error type.
Furthermore, concurrent to the classification of error category, each error-case was also labeled based on whether or not it was being successfully revised. Four more columns labeled „successfully revised in first attempt‟, „successfully
revised in final attempt‟, „unsuccessfully revised‟ and „omitted‟ were prepared (see Appendix D). Referring to the first three columns comprising the clauses‟ grammatical alterations, the classification was done by typing the number „1‟ on the column corresponding to the circumstance of revising success. In the end of the phase, the MS-Excel 2010 sort function could easily cluster the error-cases of the same category which put ease into the tabulation process.
4. Tabulation
Lastly, having the data already being classified, the tabulation to attain the value to be included in the analyses was executed. Referring to the individual record of data, a new MS-Excel worksheet was prepared to comprise all necessary values to be included in the analysis. As has already been explained in the sub chapter on operational definition of dependent variable, the value of revising progress was obtained by calculating the percentage value of the error occurrence in every research phase to the total clause produced by the participant.
Class A
Research task 1 Research task 2
Research task 3
Figure 7. Research scheme
V. Data analysis
As mentioned in chapter I, there was only one question formulated in this research. In order to answer this question, several statistic formulae were used to measure the difference on the value of dependent variables after the treatments were introduced. Sequentially, qualitative description was also used as a mean of answering the question. The steps that were taken in the process of analyses were as follow:
A. Steps taken in statistically comparing the two error-marking techniques
1. Setting null hypotheses
Assuming that undergraduate students had already possessed Saville-Troike‟s notion of adults‟ learner advantage in language learning which includes learning capacity and analytic ability (see table 5, p. 23), it could be assumed that the different degree of implicitness wouldn‟t matter as long as markings are given. Thus, there are two null hypotheses to be proven in this research: 1) There is no significant difference between the effect of color marker and comment error-marker on participants‟ first-revising attempt; 2) There is no significant difference between the effect of color error-marker and comment error-marker on participants‟ final-revising attempt. If the obtained value was more than the significance level set at .05, the difference would be rendered statistically insignificant.
2. Processing the data obtained from the experiment
The obtained data from the process of experiment were firstly being processed hence interval data were attained. Such interval scores were in the form
Quantitative data findings
Qualitative data findings
of percentage-value of error-occurrence (unsuccessfully revised error) within each research phase to the total clause produce by respective participant. In the end of the tabulation process, each participant was being attributed with two percentage values representing both his progress in the first revising attempt and his progress in the second (final) revising attempt.
3. Determining the normality of data distribution
As suggested by Kranzler, normality of data distribution is considered essential in determining the choice of using parametric or non-parametric formulae (1999, p. 119). A parametric test, which in this case was t-test, is only
applicable when the data is normally distributed (Kranzler & Moursund, 1999, p. 119). To measure normality of a set of data, two means can be applied. Graphical interpretation can be used as well as numerical interpretation through the application of one-way ANOVA test (Kranzler & Moursund, 1999). In this research, graphical interpretation was used in interpreting the normality of distribution. In doing so, IBM SPSS Statistics 20 was used following these steps: a. Copy-pasting the tabulated interval data from MS-Excel database to SPSS
worksheet. The 36 percentage values attained from the 36 participants in experimental group A in the first revising attempt were pasted into a column in the SPSS worksheet whereas the interval values in the final revising attempt were pasted into different column. Afterwards, the interval values of experimental group B were pasted into similar column. Hence, there were 64 values in each column which normality of distribution was to be determined. The two columns were then labeled as first revision and final revision.
b. Choosing analyze menu of IBM SPSS Statistics to attain a graph depicting
c. A dialogue box was opened to enter the column (variable) to be analyzed. Having the two columns already being labeled, the first revision variable was chosen to be analyzed first. Given the goal was to attain a graph depicting data distribution, chart menu was clicked to choose the type of chart to be produced in the output. Histogram chart was selected and a tick was put on the option show normal curve on histogram.
d. Creating the histogram by clicking the OK button.
e. Determining if the data distribution, as shown in the histogram was normally distributed by identifying if the shape of the curve was
bell-shaped or skewed. Kranzler&Moursund (1999) state that normally distributed data will show a bell-shaped curve.
f. Determining the data distribution of final revision variable by following similar steps.
g. Applying parametric or non-parametric test.
The intention of the application of statistical test in the analyses was to determine the significance value of difference between the two experimental groups in terms of their grammar-revising progress. In doing so, the statistical tests needed were the ones that allow the comparison between two experimental groups. The options provided under parametric test were independent t-test whereas the options provided under non-parametric test were Mann-Whitney u-test.
Kothari states that seen from type of data being analyzed, data in interval and ratio scales are parametric (2004, p. 69). Regardless the data obtained fell under the interval scale category, to apply parametric test (independent t-test) might not be fair without the consideration on normality of data distribution. As
Statistics 20 was used in analyzing the significant difference between the effects of two strategies.
Having the necessary interval data had already been pasted into the SPSS worksheet, another column labeled “treatment” was provided. Number “1” was typed in representing error-marking through color whereas number “2” was typed in representing error-marking through comment. The provision of this additional column was necessary in the process of analysis either using independent t-test (parametric) or Mann-Whitney u-test (non-parametric). Both independent t-test and Mann-Whitney u-tests function could be found under IBM SPSS Statistics 20‟s analyze menu. Independent t-test function can be found under compare means sub-menu whereas Mann-Whitney u-test can be found under non-parametric test sub-menu on legacy dialogue – 2 independent samples options. The significance level was set under the value of .05 (α = .05) in a two-tailed test. Hence any number greater than .05 would be considered strong enough to reject the null hypotheses.
B. Complementary analyses
1. Exploring the effect of the two CF strategies towards the revision of negative-interference driven errors
a. The first step taken was done during the coding and codification process (see page 36-37) . In these phases, the error-cases that had already been verified in the editing phase were being categorized based on Dulay, Burt &Krashen‟s seven category of negative-interference driven errors. The categorization was done by putting the number “1” on one of the columns. b. Having the errors being labeled by “1” on different columns, the calculation of frequency of each error category was simply determined by
using MS-Excel 2010‟s SUM formula.
error-case before and after the treatment within an experimental group was being translated into graphs.
d. These steps were also being applied to the tally sheets of the other experimental group.
e. The last step taken was to interpret the comparison of effect of the two CF strategies towards the revision of negative-interference driven errors.
2. Determining the circumstance where the use of the two strategies ended up in futile revisions
As suggested before, out of the 92 students within the population, to
choose ten students as research sample was considered representative enough. Given that the essential part of a descriptive design is to randomly choose the samples; out of the 64 students finishing the research tasks, five students from each experimental group were randomly chosen. The randomization was done following these steps.
a. Listing the participants who did not perform successful revision in the end of the experiment.
b. Writing down the aliases of those participants on a piece of paper and putting them on a jar.
c. Randomly choosing five participants from the box whose records of revision were to be scrutinized.
d. Reapplying similar steps to choose another five samples from the other experimental group.
e. Having ten participants already been randomly chosen, the essay records of respective participants were scrutinized in order to investigate how the two CF strategies had contributed to the success and the failure of revising
CHAPTER V
CONCLUSION AND SUGGESTION
In this chapter, the concluding thoughts based on the findings already elaborated in the previous chapter are presented. Having the conclusion being drawn, suggestions are also pointed out. The suggestions comprise the suggestion for the teaching of writing in higher education as well as suggestion for further studies in respective subject.
I. Conclusion
After receiving two feedbacks, the difference of effect between the two
compared error-marking strategies was proven statistically significant implying the use of comment to facilitate better progress. This finding was also validated by the finding suggesting that the use of comment facilitated better revision towards
most of Dulay, Burt &Krashen’s categories of errors. However, under the
circumstance where the feedback was only provided once, there was no significant difference between the two strategies. Despite such statistical finding, the use of color was also facilitative in assisting participants to revise their grammatical error. It was shown from the noteworthy percentage of error reduction in experimental group A and the insignificant difference between the two error-marking strategies in the first CF provision.
According to the essay records, it was also revealed that there were several circumstances where the two strategies caused futile revisions. From the scrutiny
of 10 randomly chosen participants’ essays, it was revealed that majority of the futile revisions were caused by participants’ grammaticality. There were minor
cases where the futile revisions where identified under the circumstances as
follow: 1) inexistence of similar marking in one’s essay; 2) two different markers
II. Suggestions
1. Suggestions for the teaching of writing in higher education
As suggested by VanBeuningen (2010) and Ferris (2010), students tend to make broad range of error thus the provision of focused CF is not suitable to be applied. This research had provided evidence regarding the effectiveness of unfocused feedback through comparing two techniques of such CF method, thus there should no longer be doubt on the use of unfocused CF to treat undergraduate
students’ grammatical errors. As for the implicit counterpart, the two techniques
being compared in this research may not be the most practical techniques to be
applied. Regardless, through the comparison of these two techniques, it was vividly seen that undergraduate students were able to comprehend such implicit instructions.
2. Suggestion for future research on unfocused-implicit corrective feedback
Firstly, it is to be noted that to justify the need of further research on respective topic, reference to urgency of English-writing mastery which implies necessity to provide a facilitative learning needs to be made. Then, reference on
the state of CF debate inconclusiveness should also be fused into the brainstorming. Thus, given the urgency to keep on peeling off the layers to get to the core revealing how corrective feedback is best provided, further studies on respective topic should keep on being actively conducted. Preferably, further research focusing on unfocused-implicit corrective feedback is highly suggested regardless possible negative findings that may be resulted. This research is not without flaw and limitation, henceforward, the followings are suggestions to be taken into consideration in designing and conducting researches focusing on similar topic, unfocused-implicit corrective feedback.
inter-rater reliability aspects is taken more into consideration in designing the research.
2. This research has only attempted to scrutinize the effectiveness of unfocused-implicit CF through the comparison of effects of two different techniques under such CF provision method. Extraneous factors such as level of grammatical knowledge, working memory, long-term memory and attention
(as suggested in Hayes’ writing model (2012)) were not included in the
scrutiny, thus these extraneous factors should be further analyzed.
3. This research couldn’t put strong claim towards the cases of complex marking
to be the main cause of futile revision, thus another research on respective subject should be conducted.
4. This research has attempted to study the subject only from language learning perspective, thus no effect of the treatment on grammatical acquisition was analyzed. Henceforth, it is highly suggested for the future research, investigating ineffectiveness of unfocused-implicit corrective feedback, to go an extra mile and see the longer-term effect of this seemingly less fruitful CF strategy in terms of whether or not the strategy help learners towards language acquisition.
5. To know exactly the detail of progress dynamics taking place as result of unfocused-implicit corrective feedback provision is considered essential in addition to investigating statistical significance value in attempting to understand the effect of respective subject. Thus, a research designed in mixed-approach focusing more on the qualitative aspects is preferably conducted.
6. The two techniques compared in this research may not be the most practical
Bibliography
Aisyiah, A. (2009). Students' Grammatical Errors Analysis In Writing English Composition (A Case Study in Class XI of SMK BPP Bandung). Indonesia University of Education: Unpublished.
Bitchener, J., & Ferris, D. (2012). Written Corrective feedback in second language acquisition and writing. New York: Routledge.
Bitchener, J., & Knoch, U. (2008). The value of written corrective feedback for migrant and international students. Language Teaching Research 12 , 409-431.
Björk, L. (2003). Teaching Academic Writing in European Higher Education. Dordrecht: Kluwer Academic Publisher.
Brookhart, S. (2008). How to Give Effective Feedback to Your Students. United States of America: ASCD.
Brown, H. D. (2001). Teaching by Principles: An Interactive Approach to Language Pedagogy (Second Edition). Longman.
Cargan, L. (2007). Doing Social Research. Maryland: Rowman & Littlefield. Chandler, J. (2003). The efficacy of various kinds of error feedback for
improvement in accuracy and fluency of L2 writing. Journal of Second Language Writing, 12 (3) , 267-296.
Chanquoy, L. (2001). How to make it easier for students to revise their writing: A study of text revision from 3rd to 5th grades. British Journal of Educational Psychology, 71 , 15-41.
Cohen, B., & Lea, R. (2004). Essentials of Statistics for the Social and Behavioral Sciences. New Jersey: John Wiley & Sons.
Dehn, M. (2011). Working Memory and Academic Learning: Assessment and
Intervention. Hoboken: New Jersey: John Wiley & Sons.
DeVaus, D. (2001). Research Design in Social Research. London: SAGE.
Ellis, R., Loewen, S., & Erlam, R. (2006). Implicit and Explicit Corrective Feedback and The Acquisition of L2 Grammar. Studies on Second Language Acquisition, 28. , 339-368.
Evans, N. W., Hartshorn, J., McCollum, R. M., & Wolfersberger, M. (2010). Contextualizing Corrective Feedback in Second Language Language Writing Pedagogy. Language Teaching Research , 445-463.
Evans, N., & Hartshorn, K. (2010). Contextualizing Corrective Feedback in Second Language Writing Pedagogy. Language Teaching Research , 445-463.
Evans, N., & Levinson, S. (2009). The myth of language universals: Language diversity and its importance for cognitive science. Behavioral and brain sciences, 32 , 429-492.
Ferris, D. (2003). Responding to Writing. In B. Kroll, Exploring the Dynamics of Second Language Writing (pp. 119-140). Cambridge: Cambridge University Press.
Ferris, D. (2010). Second Language Writing Research and Written Corrective Feedback in SLA. Studies in Second Langauge Acquisition, 32. , 181-201. Ferris, D., & Roberts, B. (2001). Error Feedback in L2 Writing Classes: How
Explicit Does It Need To Be? Journal of Second Language Writing , 161-184.
Garcia, L. (2009). Hormones and brain plasticity. New York: Oxford University Press.
Gathercole, S., & Alloway, T. (2008). Working Memory and Learning: A Practical Guide for Teachers. Thousands Oaks: California: SAGE.
Grauberg, W. (1997). The Elements of Foreign Language Teaching. Bristol:
Multilingual Matters.
Harmer, J. (2002). The Practice of English language Teaching Third Edition. Longman.
Hawkins, R. (2004). The contribution of the theory of Universal Grammar to our understanding of the acquisition of French as a second language. French Language Studies , 233-255.
Hayes, J. R. (2012). Modeling and Remodeling Writing. Written Communication , 369-388.
Hyland, K. (2013). Writing in the university: education, knoledge and reputation. Language Teaching Volume 46/ Issue 01 , 53 - 70.
Hyland, K., & Hyland, F. (2006). Feedback in Second Language Writing: Context and Issues. New York: Cambridge University Press.
Kellog, R. (1994). The psychology of writing. New York: Oxford University Press.
Kenkel, J., & Yates, R. (2009). The Interlanguage Grammar of Information Management in L1 and L2 Developing Writing. Written Communication Volume 26 Number 4 , 392-416.
Kirkpatrick, A., & Sussex, R. (2012). English as an International Language in Asia: Implications for Language Education. New York: Springer.
Kothari, C. (2004). Methodology: Methods and Techniques (Second Revised Edition) (2nd ed.). New Delhi: New Age International (P) Limited Publisher.
Kranzler, G., & Moursund, J. (1999). Statistics For The Terrified: Second Edition. New Jersey: Prentice-Hall, Inc.
Kusdini, R. D. (2005). Students' Grammatical COmpetence in Classroom Presentation: An Ethnographic Research of the Freshmen of Public Relations Department of STIKOM Bandung. Indonesia University of Education: Unpublished.
Lamb, M. (2009). Situating the L2 Self: Two Indonesian School Learneres of English. In Z. Dörnyei, & E. Ushioda, Motivation, Language Identity and the L2 Self (pp. 229-247). Ontario: Datapage International Ltd. .
Leijten, M., Van-Wae, L., & Ransdell, S. (2010). Correcting Text Production Errors: Isolating the Effects of Writing mode From Error Span, Input Mode, and Lexicality. Written Communication , 189-227.
Llach, M. (2011). Lexical Errors and Accuracy in Foreign Language Writing. Maria Pilar Agustin Llach.
Manser, M. H. (2006). The Facts on File Guide to Good Writing. New York: Facts on File, Inc. .
Maree, K., & Westhuizen, C. (2008). Head Start in Designing Research Proposals in the Social Research. Cape Town: Juta and Company Ltd.
Myhill, D., & Jones, S. (2007). More Than Just Error Correction: Students' Perspectives on Their Revision Processes During Writing. Written Communication, 24 , 323-343.
Nesi, H., & Gardner, S. (2012). Genres Across the Disciplines: Student Writing in Higher Education. Cambridge: Cambridge University Press.
O'brien, T. (2004). Writing in a Foreign Language: Teaching and Learning. Languaeg Teaching , 1-28.
Purnawarman, P. (2011). Impacts of Different Types of Teacher Corrective Feedback in Reducing Grammatical Errors on ESL/ EFL Students' Writing. Virginia Polytechnic Institute and State University: Unpublished. Rasinger, M. (2010). Quantitative Methods: Frameworks and Issues. In L.
Litosseliti (Ed.), Research Methods in Linguistics (pp. 50-113). New York: Continuum International Publishing Group.
Ruhlen, M. (1991). A Guide to the World's Languages: Classification. Stanford: Stanford University Press.
Saville-Troike, M. (2006). Introducing Second Language Acquisition. Cambridge:
Cambridge University Press.
Setiyadi, A. B. (2006). Teaching English As A Foreign Language. Yogyakarta: Graha Ilmu.
Sharpless, M. (1999). How we write: Writing as creative design. London: Routledge.
Siemund, P. (2011). Linguisic universals and language variation. New York: Walter de Gruyter.
Sneddon, J. N. (1996). Indonesian: A Comprehensive Grammar. New South Wales: Allen & Unwin Pty Ltd.
Stringer, D. e. (2011). Straight on through to Universal Grammar: Spatial
modifiers in second language acquisition. Second Language Research , 289-311.
Truscott, J. (1996). The case against grammar correction in L2 writing classes. Language Learning, 46 , 327-369.
Van-Beuningen, C. (2010). Corrective feedback in L2 writing:: Theoretical perspectives, empirical insights, and future directions. International Journal of English Studies, 10.2. , 1-27.
Vasquez, C., & Harvey, J. (2010). Raising teachers' awareness about corrective feedback through research replication. Language Teaching Research, 14 , 421-443.
White, L. (2003). Second Language Acquisition and Universal Grammar. Cambridge: Cambridge University Press.
Wolfe-Quintero, K., Inagaki, S., & Kim, H. (1998). Second Language Development in Writing: Measures of Fluency, Accuracy, and Complexity. National Foreign Language Resource Centre.
Yuliana, Y. (2008). Students' Error Analysis In Writing Narrative Text: A Case
