CC BY
81
POSTGRADUATE STUDENTS ACADEMIC WRITING: CORRECTIVE FEEDBACK STRATEGIES
E. Belyaeva
(St Petersburg, Russia)
Abstract On the basis of a previously developed by R. Ellis typology of corrective feedback currently available to teachers and researchers this article presents the outcomes of a small-scale research of the most effective corrective feedback strategies in a context of a university course for postgraduate students. The reader can find a useful overview of the corrective feedback strategies backed up by the relevant research briefly examined. This study has investigated the students' response to the error correction strategy offered - direct corrective feedback, indirect, metalinguistic corrective feedback and reformulation. The surveyed group consisted of 18 postgraduate students of the St Petersburg State University. The research findings convincingly show that the target group learners 'preferences and briefly describes the implications for the language teaching: the explored ways of error correction and corrective feedback can meaningfully contribute to learners' language learning and their command of academic writing.
Keywords: error correction; corrective feedback; awareness raising; consciousness raising; discovery learning; data-driven learning; second language acquisition; input; output; cognitive strategies.
Introduction
The central research question of this article is to find out which error correction techniques and corrective feedback strategies work best for the university postgraduate students doing their research in physics. This research is clearly focused on English for Academic Purposes (EAP) and English for Specific Purposes (ESP) and is limited to English for Physics. The target population of this small-scale research is a group of 18 postgraduate students of the St Petersburg State University Faculty of Physics with the language proficiency ranging from B2 to CI language level.
The main aim behind this research is to explore postgraduate students' attitudes and insights into which error correction techniques and corrective feedback strategies they consider to be effectively facilitating their research report writing skills. The target group of 18 postgraduate students was surveyed with the help of a specially designed questionnaire. This convenient sampling - the surveyed group of postgraduate students may not be a representative sample of the target population as they all are the speakers of Russian, specialize in physics and at the same university. This is an important limitation of this research casting shadow upon its findings and whether it would be possible to generalize.
The Research Background and Rationale
Teaching English to postgraduate students seriously involved in both theoretical and experimental research in various fields of Physics and quite competent in a number of research methods, is quite a
challenge. Being a mixed-level language proficiency group of learners most of them are already quite mature and committed researchers with some background in data collection, data analysis and data interpretation. Language learners with such a predisposition to all kinds of investigation and exploration are more inclined to analyze, to contrast and compare and to generalize. English language classroom should offer postgraduate students an environment conducive to critical investigation of the linguistic features involving them in a study of the form and its use. Cognitive and metacognitive language learning strategies are typically over-used.
All the above-mentioned features of the observed post-graduate students together with their above-av-erage IQ make them the best audience for such inductive approaches to language learning as consciousness raising (CR) [9] (or awareness-raising in some publications) and data-driven learning (DDL) [8]. These are the approaches which require language learners to process input in a way that makes features of the language more salient. CR is aiming at encouraging learners to identify and think about the particular features of language form and then use that at their own time, sometimes much later. An information processing approach to language learning leads first to implicit intuitive knowledge of the language and later on it will take form of explicit knowledge of the language and both kinds of knowledge make a meaningful contribution to second language acquisition and learning. [6]
CR is quite frequently compared to discovery learning [6] where "learners develop process associated with discovery and inquiry by observing, inferring, formulating hypothesis, predicting and communicating" [6, p. 162]. However, unlike induction, guided discovery learning implies a certain degree in external intervention, typically directed by the teacher, in the form of exposure to carefully selected data and skillfully placed questions and cues. Guidance is typically offered through questioning techniques. One can always question the discovery learning evidence and the boundaries between teacher guidance and discovery learning.
CR is characterized by a learner-centered orientation where the learners rely on their intellectual capacities and use their cognitive modes of learning: the most important one being noticing leading to awareness about the use of a language structure. Different aspects of grammar are focused on without necessarily using explicit rules or metalanguage (or technical jargon) to help learners discover the rules by themselves. According to Ellis: "CR has proved to be really useful at an initial stage of language acquisition, the stage of controlled processing, to trigger the declarative knowledge, leading the way gradually to the procedural knowledge when learners attend to content rather than form." [2] This is very true and valid statement regarding the surveyed group of learners. The postgraduate students, whose language proficiency is above B2 level and whose motivation is reasonably high as they are genuinely interested in the subject matter, seek helpful feedback on their written output. What kind of feedback do these learners need and consider helpful?
The approach to developing postgraduate students' academic writing skill is based on the combination of the two approaches to developing writing: mod-el-based and process-oriented. Students are offered a wide range of samples of research reports and are exposed to a variety of samples to investigate and explore the typical organization of such a document. There are generalizations made and typical genre features are listed in order to develop deep understanding of all the requirements. Then students are guided to explore the language of the research report with the most frequent grammar and lexis. Students are encouraged to investigate the most frequent formulaic language, grammatical and lexical academic discourse features, written discourse markers and all possible ways of signposting. Students are also offered a couple of an-
ti-examples to critically examine those and to identify all the weaknesses in the structure and the organization as well as the critical analysis of the selected linguistic tools regarding their relevance and accuracy. All of those model-based techniques paved the way to students' work on their first draft of their research reports. Teacher provides corrective feedback using certain strategies. Students get teacher feedback, reflect on the effects and self-explore their reactions to a certain corrective feedback strategy. Then students revise their second draft, self-edit it and get feedback through a different correction strategy in order to further work on the questionnaire and the same cycle repeats around the third (final) research report.
The research method and focus
The two basic questions of this small-scale research are aimed at investigating the surveyed group of postgraduate students' reflections on which corrective feedback (CF) strategies work for them best and how they respond to them.
This study consisted of the following seven stages:
1) the postgraduate students are instructed to produce the first draft of their research reports);
2) teacher provides CF using direct corrective strategy;
3) students self-correct their first draft and submit the second draft of their research reports;
4) teacher provides indirect CF;
5) students self-edit their second draft and submit their final draft;
6) teacher provides metacognitive corrective feedback with occasional reformulation;
7) students revise and self-edit their third draft on the basis of the received CF and submit the final version.
This research focuses on the two dimensions of CF: on the four strategies for providing CF - direct, indirect or metalinguistic strategy and reformulation - and on how students respond to the strategies used.
Direct CF is the most widely used by teachers all over the world and the easiest to describe: directly corrected piece of writing demonstrated teacher corrections, most often in red color, with an overall mark or score or a brief assessment comment. It can effectively bring an immediate insight and bring about the short-term accuracy improvement.
Indirect CF requires a teacher to develop or choose an error code to indicate clearly to students the place where something is wrong and to give a clue or a helping hint regarding what the problem is. Error code is
supposed to guide learners to a better language. One obvious disadvantage of indirect CF is that if learners don't know the correct form they cannot self-correct. Sometimes learners self-correct but they are not certain if they are correct. However, indirect CF leads to guided learning and problem solving and it definitely encourages learners to reflect on the selected linguistic tools. Besides, the indirect CF is considered to have stronger chances to lead to long-term learning.
The results of various studies that have investigated direct versus indirect CF are very mixed and no study to date has compared the effects on accuracy. There have been comparative studies of direct corrective feedback and error codes [4] to discover that indirect CF using error codes led learners to fewer errors that direct CF. Lalande research studied the L2 German learners. Ferris and Roberts [3] research discovered that error codes helped learners to self-edit their writing but no more than indirect feedback.
Metalinguistic CF involves providing learners with some form of explicit comment about the nature of the errors they have made using error codes or metalinguistic explanations of their errors comment. Ferris and Roberts [3] research demonstrated that error codes helped students to self-edit their writing but no more than indirect feedback. However, there is very limited evidence to suggest that error codes helped students to achieve greater accuracy over time. Besides, it would also seem error codes are no more effective than other types of corrective feedback in achieving self-editing.
Research on metalinguistic corrective feedback shows that it is less common than error codes due to its being quite time-consuming and requiring teachers to be able to write clear and accurate explanations for a variety of errors. This might be too challenging for the non-native speaker teachers who are often insecure about so many things. It has to be mentioned here that within this study metalinguistic CF was offered to surveyed group of students through a discussion (delayed public corrective feedback) during the lesson. Sheen [7] compared direct and metalinguistic CF and discovered that both were effective in increasing accuracy in the students' use of articles in subsequent writing completed shortly after the CF treatment but metalinguistic CF proved more effective that the direct CF in the long term.
Metalinguistic CF within this study was required every time when the students could not self-correct or self-edit their research reports on the basis of the indi-
rect CF given to them with the help of an error code. Therefore, metalinguistic CF often looked like remedial teaching with some elements of teacher explanations, investigations of the sources of errors (lack of knowledge, the complexity of certain language phenomena, first language interference or transfer, etc.), consciousness raising activities (researching authentic data - research reports samples) and comparing/ contrasting students' first language and the target language. Unlike focused CF which provides multiple corrections metalinguistic CF occasionally may seem lacking focus or completely unfocused as it is aiming at the deep understanding of the nature of the students' errors. Focused CF does seem more efficient in the short term as students are provided with the multiple corrections of the same error and these errors are attended to. However, unfocused metalinguistic CF may prove superior in the long term.
Reformulation requires "...a native speaker writing the student's text in such a way as to preserve as many of the writer's ideas as possible, while expressing them in their own words so as to make the piece sound native-like" [1:4]. The writer then revises by deciding which of the native speaker's reconstructions to accept. The described-above CF strategy requires a non-native speaker teacher of English with quite high level of language proficiency. Sometimes teachers avoid taking risks suggesting their reformulations in order to avoid loss of face over some very complicated cased, particularly when the student's messages are unclear in the first place. However, if teachers see ways of reformulating student's defective sentence and are capable of making more than one suggestion, then it makes reformulation work effectively. We would argue that reformulation is the best CF strategy in every case when student completely fails to self-correct or self-edit and when the problem is of a more subtle nature than an obvious grammatical error, but of an academic discourse nature.
The specially designed for this survey questionnaire had 2 questions in Russian (to avoid any kind of misunderstanding). One closed yes/no type of question where students were asked to tick all the CF strategies they found useful for themselves. The second question was an open one and required to explain the reasons behind their responses.
The study findings
The survey outcomes demonstrate that the surveyed postgraduate students preferred reformulation (94%), indirect CF (83%) and metalinguistic CF
(66%). Direct CF was mentioned only by 3 students translated into English. Students were also encour-
(16%) and those with the lowest language proficiency aged to express their reactions to carious CF used and
in the group. The Data collected through the group their preferences on how CF should be offered,
survey is presented in the table 1 below and shows Table 1.
the numbers of students' preferences among the dis- Summary of the 18 postgraduate students' respons-
cussed CF strategies together with their comments es
CF strategy N of Ss who found this CF effective Students9 reflective comments on how they responded to a particular CF strategy
Direct CF 3 students (with the lowest language level) • briefly looked at the teacher corrections • looked for the final mark • compared the total amount of corrections with the grade given • reflected on the grade objectivity • copied the corrected text (almost mechanically) • never thought much about the source of errors • could easily self-correct many grammatical errors if given a chance • delayed and individual
Indirect CF 15 students • self-correction is time-consuming • didn't know how to correct/improve • didn't know which reference materials to use • decided to replace the wrong sentence with a totally different one • didn't take risks and played safe with the structure I definitely knew • grammatical errors are easier to correct than the lexical ones • the errors marked "st" (stylistic errors) are the most challenging • it was sometimes easier to change the original text completely than to improve it • delayed and individual
Metalinguistic CF 12 students • teacher explanations were useful • in teacher feedback there were many new and useful things about academic writing conventions in general and about report writing in particular • it is useful to understand the reason behind the error • comparing and contrasting the Russian and the English languages helps a lo • deep insights into the academic discourse make me think about how language works • thinking and speaking about the language is very stimulating • delayed and public
Reformulation 17 students • teacher suggestions provide a difficult choice - which option to choose • teacher reformulated sentenced make me think about the message and then search for the best wording • it is about comparing messages I need to send and how to formulate those messages best in English • teacher makes a valuable contribution to my writing • teacher does a good job and is very helpful • delaved and individual
Evidence-based conclusions
This study has explored the surveyed group of postgraduate students' preferences among the suggested CF strategies and identified reformulation, indirect CF and metalinguistic CF as most effective and useful for them through a process of self-reflec-tion on the basis of their hands-on experience. The study findings have helped the teacher to make more informed choices of the CF strategies that work best with the surveyed group. The hypothesis about the postgraduate students preferring more intellectually stimulating CF strategies involving a lot of awareness raising and consciousness-raising thus contributing to their deeper understanding of the language in general and academic discourse in particular has been proved.
The study findings contribute to our deeper understanding of the CF strategies effectiveness when teaching postgraduate students with a certain predisposition to all kinds of investigation and research. Therefore it can be recommended to extensively employ the above discussed CF strategies as a meaningful part of data-driven and consciousness raising approach to language teaching postgraduate students.
Future directions of the research
Electronic CF has not been extensively used in Russian universities so far for many good reasons. However, in our view, electronic extensive corpora of written English can assist students in their learning English in a very powerful way. Electronic resources provide learners with the models and examples of the usage of more experienced writers and make comparisons of their own writing to the given examples and discover meaningful differences. An example of an electronic CF is "Mark My Words" developed by Milton in 2006 consists of 100 recurrent lexico-gram-matical and stylistic errors frequently occurring in
the writing of Chinese students. It is a big question regarding as to what extent such a resource can be of use and help to Russian students but to develop a similar electronic CF tool would certainly make an impact upon corrective feedback strategies further development.
Bibliography:
1. Cohen, A. 'Reformulation: A technique for providing advanced feedback in writing'. Guidelines 11/2: 1-9, 1989.
2. Ellis, R. Task-based Language Learning and Teaching. Oxford: OUP, 2003, 387 pp.
3. Ferris, D. R. and B. Roberts. 'Error feedback in L2 writing classes: "How explicit does it need to be?", Journal of Second Language Writing 10: 161— 84, 2001.
4. Lalande, J. F. 'Reducing composition errors: An experiment'. Modern Language Teaching Journal 66: 140-9, 1982.
5. Richards, J. and Schmidt, R. Longman Dictionary of Applied Linguistics and Language Teaching (3d edition), Harlow: Longman, 2002, 389 pp.
6. Rutherford, W., & Sharwood Smith, M. (1985). Consciousness-raising and universal grammar. Applied Linguistics, 6(3), 274 - 282. In: W. Rutherford and M. Smith (eds.), Grammar and Second Language Teaching.
7. Sheen, Y. 'The effect of focused written corrective feedback and language aptitude on ESL learners' acquisition of articles'. TESOL Quarterly 41: 255-83, 2007.
8. Sinclair, J. Corpus, Concordance, Collocation. Oxford: OUP, 1991, 116 pp.
9. Skehan, R A Cognitive Approach to Language Learning. Oxford: OUP, 1998
