Introduction
Rationale of the Study
English proficiency is crucial for daily interactions and success in Vietnam, prompting many to enroll in English classes to enhance job prospects and promotions Writing, in particular, is regarded as the most challenging skill for English as a Foreign Language (EFL) learners, as noted by Richard (2002), who emphasizes the difficulty in constructing grammatically correct sentences This challenge is compounded by the need for a high level of language proficiency, as highlighted by Kern (2000).
Understanding how to write a summary or analysis in Mandarin or Spanish does not guarantee that students can perform the same tasks in English Therefore, effective instruction must consider the diverse educational, social, and cultural backgrounds that influence students' native language experiences Advanced learners often struggle with native-like collocation and idiomatic expressions As noted by Kim (2009), even when advanced students produce writing with few grammatical errors, their work may still sound unnatural due to differences in the frequency of certain linguistic structures or unfamiliarity with genre conventions Consequently, both students and teachers require access to comprehensive and authentic language materials that extend beyond traditional textbooks and reference sources.
The use of authentic materials in foreign language teaching is crucial, offering numerous advantages such as cultural richness, relevance to learners, and enhanced motivation (Mishan, 2004) In today's technology-driven era, the integration of computers and the Internet has transformed education, providing increased engagement for both instructors and students Educators now leverage easy access to a variety of authentic language samples across different media, including newspapers, magazines, TV, books, and online resources, to improve pedagogical quality and deepen understanding of foreign languages Additionally, there is a rising interest in utilizing corpora for educational purposes.
A corpus is a comprehensive database of authentic language use from diverse sources, as defined by Yoon and Hirvela (2004) Biber et al (1998) describe it as a large, organized collection of natural texts representative of a specific language or dialect Kim (2009) highlights that searchable data corpora serve as valuable resources for EFL students and teachers to verify the appropriateness of their language use With its effectiveness in enhancing language proficiency, the application of corpora in language learning has gained significant attention in recent research.
The study explores the effectiveness of EFL students in using corpus tools, specifically the Corpus of Contemporary English (COCA) and the British National Corpus (BNC), for self-editing their essays With the rise of computer availability in classrooms, students can now easily access these resources to enhance their writing skills COCA, launched in 2008 by Professor Mark Davies, features over a billion words from diverse spoken and written sources and is regularly updated, ensuring relevant language material Similarly, the BNC, created by Oxford University Press in the 1980s and 1990s, contains 100 million words across various genres, offering valuable examples for EFL learners Both corpora serve as effective self-correction tools aimed at improving students' academic writing and fostering their error self-correction skills and autonomy.
This research examines student performance, identifies the types of errors that students can self-correct, and explores their attitudes towards learning Additionally, it evaluates the advantages and disadvantages of using a corpus as a supplementary tool in language education.
Aims of the study
This study aimed to introduce corpus as a supplementary tool for enhancing EFL students' writing skills at HUFL, where many students are unfamiliar with its benefits It explored how EFL students utilize corpus during their essay self-correction process and their perceptions of its role in L2 writing Understanding students' responses to a classroom environment that emphasizes corpus use is crucial, as it can inform EFL educators in adjusting corpus-based techniques to foster more effective and meaningful learning experiences.
Based on the purpose of the study mentioned above, the study was carried out to answer the following research questions:
1/ What types of errors can students correct with the corpus?
2/ What are the students‘ experiences with the use of corpus as self-correction tool for their writings?
3/ What are the students‘ attitudes towards the use of corpus as self-correction tool for their writings?
Significance of the Study
This study aims to introduce corpus tools to HUFL students, believing it will yield significant benefits As corpus usage is a novel concept for many students, it is expected to spark their interest and encourage deeper exploration Utilizing corpus as a self-correction tool can enhance the quality of essays for senior students, improving grammar, collocation, and overall naturalness By providing authentic examples, corpus serves as a valuable resource for identifying grammatical and lexical errors, thereby refining writing skills Familiarity with native speakers' writing styles and analyzing real-life context extracts also enhances students' sociolinguistic and socio-pragmatic competence.
Corpus shows how language is used in real situations and reduces students‘ reliance on teachers to tell them directly what errors have been made and how to correct them
Teachers often identify students' mistakes without providing direct corrections, requiring students to self-correct By integrating corpus tools into the error-correction process, the burden on teachers can be significantly reduced.
Scope of the Study and Limitations
This study explored the experiences of EFL students regarding the benefits and challenges of using corpus tools, along with their reactions after exposure The research focused on 10 volunteers from third- and fourth-year students at the Faculty of English at HUFL Participant recruitment took place over a two-week period, concluding when the target number was reached or the time limit expired A one-week training session preceded a six-week experimental phase, highlighting the study's specific context within Vietnam, particularly among EFL courses at HUFL.
This study faced several limitations, primarily due to the researcher being a non-native speaker, which influenced the error evaluations in students' essays based on personal knowledge and experience Additionally, the research was confined to HUFL in Hue City, Vietnam, involving a deliberately selected, small participant group Consequently, the study's objectivity, representativeness, and reliability were challenging to establish.
Literature Review
Definition of Corpus
Advancements in electronic and computer sciences have transformed language teaching, significantly enhancing language instruction through corpus linguistics This field focuses on the construction, processing, and analysis of language corpora via computer systems According to Sinclair (2005), a corpus is defined as a collection of language texts in electronic form, selected based on external criteria to effectively represent a language or language variety for linguistic research purposes.
A corpus is a collection of linguistic data, encompassing both written texts and recorded speech transcripts, which serves as a foundation for linguistic description and hypothesis verification about a language (1991) These machine-readable texts facilitate various linguistic analyses, including investigations into lexis, syntax, discourse, regional variations, and differences between learners and native speakers (Koo, 2006) The corpus can comprise texts from diverse genres such as newspapers, journals, and books, making it a valuable resource for understanding historical language changes and usage.
A corpus is a substantial collection of written and spoken texts from various registers, often stored digitally in extensive databases (Dobrić, 2009) It serves as a sample of language that includes material from diverse genres or specific topics Ranging from thousands to hundreds of millions of words, a corpus enhances opportunities for empirical observation of linguistic phenomena across different levels of analysis, including word, clause, phrase, and text (Francis, 1993).
A general corpus is invaluable for understanding the global usage of vocabulary and phrases, serving as a foundation for dictionaries and textbooks In contrast, a specialized corpus is tailored for specific purposes, enhancing knowledge of language use within particular registers (Tribble & Jones, 1990) Each corpus possesses unique language characteristics that cater to diverse user needs For instance, a corpus of news articles related to significant events, such as wars or natural disasters, provides journalists with insights into the perspectives of specific groups during those times Likewise, a corpus of business letters reveals important information about grammar and phrasing in various commercial contexts Therefore, researchers must consider both the language variety and the corpus size to effectively conduct a meaningful study.
This study utilizes the Corpus of Contemporary American English (COCA) and the British National Corpus (BNC) as key research tools, focusing on the nuances of both American and British English.
British English) has been used in recent years This encompasses written and spoken language from a variety of registers, including academic, newspapers, blogs, and movies
The Corpus of Contemporary American English (COCA) is recognized as the largest freely accessible English corpus, featuring over one billion words across more than 485,000 texts It has seen an impressive addition of 20 million words annually from 1990 to 2019.
The platform offers a diverse range of content, including speech transcripts, fiction, popular magazines, newspapers, academic texts, and, as of the March 2020 update, TV and movie subtitles, blogs, and various web pages Its user-friendly interface enables searches for specific words, phrases, and parts of speech, while also allowing users to compare word frequency across different genres and time periods.
British National Corpus (BNC) (english-corpora.org/bnc/) is a collection of 100 million words of contemporary British English text held in computer-readable form
The British National Corpus (BNC) comprises a diverse range of texts, with 90% consisting of written materials such as regional and national newspapers, specialized journals, academic books, and various forms of fiction, alongside unpublished letters and essays The spoken component, making up 10% of the corpus, includes transcriptions of informal conversations from a variety of demographic backgrounds, as well as recordings from formal settings like business meetings and media interactions The BNC's interface and features closely resemble those of the Corpus of Contemporary American English (COCA).
In this experiment, COCA and BNC were chosen for their extensive, freely accessible databases and multifunctional capabilities COCA has demonstrated its value in Kim (2009)'s study for several compelling reasons.
Users can search for words or phrases, grammatical constructions, parts of speech, and collocations; they can see how frequently each of these types has been used over the past 15-
The COCA database, which has been developed over 20 years, allows users to explore different genres of English usage It enables comparisons between related words and their collocates across various contexts and registers, providing insights into their frequency With part of speech (POS) tags included, users can efficiently search for specific word classes to refine their results Additionally, the COCA website offers a feature for users to create and review their query history, enhancing their research experience.
BNC, a comprehensive corpus with an interface similar to COCA, is favored for its accessibility, user-friendly design, and extensive data It offers a wealth of words and functions, making it an invaluable resource for EFL students By utilizing both BNC and COCA, learners can enhance their language skills with insights often missing from traditional English learning materials Together, these corpora serve as effective tools for exploring modern English usage across various genres and time periods, regardless of the English variety.
The influence of global English usage significantly shapes language learning, making the combination of the COCA and BNC corpora valuable resources for EFL learners Utilizing these corpora enhances learners' understanding of English and supports their efforts to improve writing accuracy.
Data-driven learning (DDL)
The Data-driven Learning (DDL) approach in teaching English as a Foreign Language (EFL) enables learners to access linguistic data to achieve their educational objectives, as introduced by Tim Johns in 1990 DDL encourages students to explore language data like detectives, examining authentic language use through real-life contexts This method involves learners analyzing concordance lines—text samples focusing on specific words—to identify patterns and rules inductively, rather than relying on pre-provided linguistic rules By utilizing a corpus, which is a collection of digital texts from natural language contexts, DDL allows foreign language learners to understand typical language usage, thereby enhancing their speaking and writing skills.
Praise (2014) identified two approaches to Data-Driven Learning (DDL) The first approach involves direct consultation of corpora, either through desktop applications or online platforms, allowing learners to independently search for language forms and word choices during their writing process For example, students can utilize extensive online corpora like COCA and BNC, which encompass various genres, or specialized corpora tailored by teachers to meet specific professional needs, such as business writing or law (Flowerdew, 2010, p 448) The second approach is mediated use, where learners engage with materials created by teachers based on corpus data, a concept referred to as pedagogic processing by Flowerdew (2009) and linked to corpus data transferability by Widdowson (1991) This mediation may include printed resources or annotations on written texts to guide students in corpus consultation (Boulton, 2010; Gilmore, 2009).
Chujo et al (2012) categorized data-driven learning (DDL) into three distinct approaches: computer-based, paper-based, and combined The computer-based approach fosters direct interaction between learners and technology, promoting a student-centered and inductive learning experience where students act as researchers due to the limitless data available (Gavioli, 2001) In contrast, the paper-based approach involves students analyzing printed concordance lines provided by instructors, making tasks more straightforward and time-efficient, as they do not require computer access (Boulton, 2010) The combined approach integrates elements of both methods; however, research on this approach remains limited.
2.2.2 The Effectiveness of DDL in Foreign Language Acquisition
The emergence of corpus-driven Data-Driven Learning (DDL) has significantly enhanced the quality and sustainability of sentence-level error correction in second language (L2) writing, demonstrating clear advantages for L2 writers (Johns, 1991) A variety of empirical studies have validated these benefits, highlighting the effectiveness of DDL in improving writing skills.
One of the key benefits of Data-Driven Learning (DDL) is its use of authentic language, which allows both teachers and learners to analyze naturally occurring language for grammatical patterns, word usage, and semantic features (Romer, 2008; Flowerdew, 2009) This approach exposes students to language themes that are often absent from traditional textbooks, enhancing their vocabulary and grammatical understanding (Shaw, 2011) Additionally, DDL equips students with the necessary skills to navigate the spontaneous language they encounter in English-speaking environments, fostering greater awareness and adaptability (Romer, 2008).
In Data-Driven Learning (DDL), teachers act as coordinators, guiding students to explore corpus data and discover language patterns independently, which is often unattainable without corpora This approach shifts away from traditional inductive learning, allowing students to learn rules and analyze authentic language examples through concordancers—tools specifically designed for linguistic analysis—rather than relying solely on definitions from dictionaries or grammar textbooks Consequently, learners can identify patterns and derive further examples with the support of teacher insights (Chujo, Anthony, & Oghigian, 2009).
Data-driven learning (DDL) significantly enhances long-term language acquisition by fostering cognitive and metacognitive skills such as noticing, hypothesizing, and verifying This approach increases learners' sensitivity to authentic language use and boosts motivation through personalized learning experiences Consequently, students become more active participants in their education, fostering greater autonomy for lifelong learning Traditionally focused on concordance lines, recent studies and corpora now include texts with expanded contexts, thereby broadening learning opportunities.
Research has shown the effectiveness of Data-Driven Learning (DDL) and corpus tools in language acquisition and writing error correction A study by Yunus and Awab (2014) demonstrated that law undergraduates using a DDL approach significantly outperformed their peers in traditional learning settings in tasks related to colligational competence Similarly, Gilmore (2009) found that L2 Japanese learners improved their writing strategies after utilizing online corpora like the BNC and COBUILD, enhancing the naturalness of their writing Furthermore, Quinn (2015) developed a training sequence for L2 writing teachers, emphasizing the use of concordance instead of traditional dictionaries, which included targeted feedback and a structured revision process Participants reported that corpus referencing positively impacted their written expression, highlighting the benefits of integrating corpus tools in language education.
Despite the increasing support for Data-Driven Learning (DDL) and corpus-based language instruction in linguistics, its implementation in language classrooms remains limited Factors contributing to this include teachers' and students' unfamiliarity with DDL, a lack of understanding of its processes, hesitance to adopt new teaching methods, insufficient integration knowledge, and inadequate resources Additionally, students may resist DDL due to unfamiliar methods, varying personal learning styles, technical skill gaps, and low face-validity This study aimed to introduce basic DDL by incorporating corpora as a supplementary tool for essay self-revision among students at the Faculty of English at HUFL, promoting a deeper understanding and awareness of language use.
DDL has proven to be an effective method for enhancing foreign language and second language (L2) learning, offering significant benefits While earlier research focused mainly on theoretical aspects, recent empirical studies have increasingly validated the positive impacts of DDL on language acquisition These studies highlight its role in fostering learner autonomy, increasing language awareness, enhancing noticing skills, and expanding cognitive abilities in learners.
In recent years, the popularity of computer-based DDL (Data-Driven Learning) approaches has surged, driven by advancements in the digital age Over the past few decades, numerous computer-based tools for language acquisition have emerged, with online corpora becoming essential resources in language teaching and learning This study focuses on the use of the COCA (Corpus of Contemporary American English) as a key instrument in this educational process.
BNC are chosen as a supplementary tool for students‘ writing revision for its versatile functions and large free online database.
Writing skill and Self-correction in Writing
Writing skills are essential for effective communication, allowing individuals to express their thoughts and feelings According to Nunan (2003), writing involves both physical and mental processes, where the act of writing includes committing ideas to a medium and organizing them for clarity Harmer (2001) emphasizes that writing is a process comprising planning, drafting, editing, and finalizing a piece Richard and Schmidt (2010) further describe writing as a complex interplay of planning, drafting, reviewing, and revising Additionally, Shokrpour & Fallahzadeh (2007) highlight the social dimension of writing, viewing it as a reflection of the writer's expertise and communication skills Mastering writing skills, particularly in English as a second language, can be challenging yet vital for effective expression.
Before the 1960s, writing skills in foreign languages received little attention, but their significance has grown in recent decades as a crucial communication skill in language learning Harmer (1998) emphasized that teaching writing to English as a foreign language students serves multiple purposes, including reinforcement, language development, and accommodating diverse learning styles, highlighting writing as an essential skill Despite often being the least favored skill among students, writing is vital for both daily life and academic success Mastery of writing enables students to effectively use appropriate language and style in their future studies and careers.
In today's technologically advanced era, effective written communication is crucial as information exchange predominantly occurs through digital means.
Learning a foreign language is a gradual journey, and errors are a natural part of this process Implementing error correction is essential, as it significantly enhances the proficiency of English as a Foreign Language (EFL) learners.
Effective language acquisition requires a supportive learning environment, as learners often struggle with the fear of making errors, which can hinder their receptiveness and responsiveness in the learning process To address this issue, language classrooms should foster a friendly and relaxed atmosphere, allowing learners to feel comfortable and confident in their language development By incorporating up-to-date techniques tailored to individual learners' needs, educators can create a more engaging and inclusive learning experience, ultimately promoting successful language adoption.
Traditionally, teachers address student errors by providing feedback, taking an active role in error detection and correction, while students often adopt a passive stance, relying on teachers to fix their mistakes (Ferris, 2001) However, this method proves to be inefficient over time, particularly for fossilized errors—mistakes that students subconsciously incorporate into their language use (Ancker, 2000) To foster effective learning, teachers must encourage students to view errors as valuable learning opportunities rather than setbacks.
Recent trends in language teaching have shifted towards minimizing the teacher's role while maximizing student engagement, emphasizing the importance of learner autonomy (Chen, 2007; Wu, 2003) Ferris (2001) highlights the necessity of training students to self-correct, as teachers cannot always be present to provide feedback This approach aligns with Allwright and Bailey's (1991) assertion that a key objective of second language acquisition is enabling learners to independently identify and correct their own errors.
As a result, self-correction can be viewed as a crucial part of language learning
Research indicates that self-correction significantly reduces the number of errors made by students over time It has been found to be more effective than teacher-led correction and revision, fostering a positive attitude towards error correction among learners Additionally, self-correction prompts meta-cognitive discussions in the classroom, creating valuable learning opportunities (Hanrahan & Isaacs, 2001; Kubota).
A study by Rana and Perveen (2013) demonstrated that students could enhance their writing skills by identifying and correcting errors in their own work This self-correction process allowed students to recognize specific issues in their writing, leading to significant improvements.
In conclusion, self-correction is essential for students to achieve long-term learning goals, as it enhances their motivation for self-development and creates valuable opportunities for learning from their mistakes.
Corpus as a Self-correction Tool in Writing
Self-correction is an effective method for helping students recognize and address their writing errors, yet both students and teachers often struggle with identifying and correcting mistakes in target language writing Lou and Liao highlight the ongoing challenges in this area.
In 2015, it was observed that EFL students often struggle to identify their own errors, necessitating teacher assistance for effective self-correction Some learners rely on intuition for corrections, which can lead to inaccuracies, while others consult dictionaries or reference materials However, it's important to note that not all errors can be resolved solely through these resources.
Collocation poses a significant challenge for students, particularly EFL writers, as it involves specific lexical items that co-occur with mutual expectancy (Nattinger & DeCarrico, 1992) While collocations allow for some flexibility, they are less appropriate when components are replaced, leading to difficulties in usage Smirnova (2017) highlighted that EFL students often struggle with limited vocabulary, resulting in frequent collocation errors even after resolving grammatical issues Furthermore, research indicates that collocational competence does not develop in tandem with overall vocabulary knowledge (Bahn & Eldaw, 1993), causing learners to misuse or underutilize idiomatic expressions in favor of less common word combinations.
Many students, despite having a substantial vocabulary in English, struggle to construct essays that reflect native fluency, primarily due to a lack of practice in immersive language environments As a result, they often turn to their teachers for corrections, believing that educators can provide accurate guidance However, teachers may rely on their own limited knowledge and intuition, which can lead to incorrect feedback (Kim, 2009) This reliance on teachers as a trusted source for correcting local errors in EFL writing classes is echoed by Liou and Peng (2009), highlighting the challenges faced by student writers.
Recent research has explored the use of corpora as tools for addressing linguistic challenges in writing, particularly in correcting grammatical errors A notable study by Gaskell & Cobb (2004) focused on the impact of concordance feedback on L2 learners' sentence-level writing errors In this study, 20 Chinese EFL participants submitted ten essays over a fifteen-week semester and revised them using corpus examples The findings indicated a significant improvement in error correction accuracy when learners had access to online corpus links with instructional support Additionally, Mizumoto et al (2012) examined the influence of learner corpus size on grammatical error correction in ESL writing, revealing that corpus size significantly affects error rectification, with varying degrees of improvement depending on the type of error, such as article, preposition, lexical choice, and adjective errors.
Research on collocation error correction highlights the significance of corpus usage in language learning Yoon's (2008) study involving six advanced ESL graduate students in an academic writing course demonstrated that utilizing the Collins COBUILD Corpus, one of the largest general corpora, effectively enhanced learners' awareness of English collocation This awareness is crucial for improving writing skills in English.
Similarly, the findings of Satake‘s study (2020), with the help of 55 undergraduate students from a university in Tokyo, exposed that corpus use promoted the accurate correction of colligation and collocation errors
The use of corpus applications significantly enhances students' ability to self-correct errors in their writing, particularly in identifying and correcting grammatical and collocation mistakes.
The application of Corpus in the Use of Error-correction Tool
2.5.1 Benefits of Corpus in the Use of Error-correction Tool
There are several studies have been conducted to examined the effectiveness of corpus research as the use of a self-correction tool for EFL student Chambers and O‘Sullivan
In 2004, a study involving 8 advanced French learners examined the impact of corpus use on error correction and learners' evaluation of the corpus consultation process After a three-week training in concordance, these graduate students used a semi-specialized corpus to revise teacher-marked errors, achieving a 75% accuracy rate in their corrections Similarly, O'Sullivan and Chambers (2006) found that 14 undergraduate French majors had a 73% success rate in correcting errors through corpus consultation Both studies demonstrated that corpus use effectively reduces native language interference and is superior to traditional resources like dictionaries and grammar books for correcting prepositions and idiomatic expressions However, undergraduates reported a slightly less favorable evaluation of corpus consultation compared to post-graduate students, aligning with Granath's (2009) assertion that advanced learners gain more from using corpora.
A study by Lou and Liao (2005) examined the effectiveness of corpora versus traditional online dictionaries in assisting 30 undergraduate EFL students in China with essay revisions Utilizing the BFSU CQP web corpus, the experimental group demonstrated greater success in correcting lexico-grammatical errors compared to those using online dictionaries The results indicated that corpora serve as more effective reference resources, helping learners make accurate corrections and minimize errors in their writing Furthermore, participants expressed positive attitudes towards the use of corpora, recognizing their significant value in enhancing writing competence and raising language awareness through engagement with authentic texts.
Quinn (2015) found that training L2 writers to use corpora as a self-correction tool significantly enhanced their revision process Participants reported that utilizing the corpus alongside dictionaries, thesauruses, and bilingual resources enriched their lexical choices and informed their corrections Many students acknowledged their limited vocabulary but realized, through corpus research, how to express themselves more diversely.
As the corpus is a collection of authentic texts, the authenticity of materials helps students expand their understanding of the word functions in particular genre contexts (Yoon
Hilvera (2004) and Flowerdew (2009) emphasize the importance of studying naturally occurring language to identify grammatical patterns, word usage, and semantic and pragmatic features This approach enables students to produce more natural expressions in their essays and explore language topics that may not be covered in traditional textbooks or by their teachers.
In 2008, Romer emphasized that teachers are unable to alter the language students experience in English-speaking settings However, utilizing a corpus to provide authentic examples can significantly aid students at all proficiency levels This approach fosters "noticing" or "awareness-raising," which is essential for developing the skills needed to navigate the language they encounter.
Learner autonomy is essential for academic success, as highlighted by Little (1995), who asserts that making learner autonomy a key objective enhances learner outcomes Promoting this autonomy through corpus work varies according to learners' language proficiency and independence levels The use of concordances, especially for error correction, has shown significant potential O'Sullivan and Chambers (2006) emphasize the value of indirect teacher feedback in learner error rectification Actively consulting corpora engages learners in enhancing their essay quality and finding answers to their inquiries Yoon (2008) supports the notion that students are increasingly familiar with corpus usage, which fosters a habit of self-examination and grammar inquiry Consequently, this process gradually nurtures students into becoming independent language learners.
Utilizing a corpus can significantly enhance students' confidence in their writing skills A study by Yoon in 2008 found that students who engaged with a corpus demonstrated increased attention to detail during the editing process and reported greater self-assurance in their writing abilities.
A study conducted in 2019 investigated the effectiveness of BYU-BNC as a reference tool for learners in correcting writing errors The results indicated that 75% of participants felt that the training course significantly improved their writing techniques and boosted their confidence in utilizing the corpus for revision.
2.5.2 Limitations of Corpus in the Use of Error-correction Tool
In spite of the above-mentioned positive effects of corpus consultation, Tono et al
Research by 2014 indicated that not all error types are effectively corrected through corpus consultation, focusing on three specific errors made by learners revising English compositions The study found significant differences in correction accuracy, with omission and addition errors being easily identified and corrected, while misformation errors showed lower accuracy rates Similarly, Baghestani (2011) identified that students could effectively correct word choice, preposition, and word form errors using corpora, but struggled with spelling errors, which were the least susceptible to correction through this method.
Another possible drawback of corpus is the overload of information Yoon & Hilvera
Learners often feel overwhelmed by the extensive linguistic input available through corpus resources, leading to frustration and hesitation in error-checking (2004) Thurstun and Candlin (1998) noted that students expressed concerns about the challenges of consulting authentic materials, citing the daunting amount of data in example sentences Similarly, Vasiljevic (2014) observed that both teachers and students struggled to navigate the vast information presented in concordance lines when utilizing corpora in writing classrooms.
Many students struggle with interpreting results, making the process of correcting their errors time-consuming According to Sun's study (2000), the lengthy nature of data analysis poses significant challenges for students.
The two main reasons for this problem was the large number of example sentences given by the corpus and the difficult and unfamiliar language for students
Students frequently struggle to create effective search queries, often failing to ask the right questions to address their errors (Baghestani, 2009) This issue is compounded by their tendency to focus solely on underlined words or phrases, leading to misinterpretation of search outcomes (Kennedy & Miceli, 2010) The ability to conduct appropriate searches is closely linked to students' understanding of the problems they face, as those with only a vague grasp of the issue may find it challenging to articulate their queries effectively (O'Sullivan & Chambers, 2006).
Learners‘ Attitudes towards the Use of Corpus on Writing Self-correction
Research has shown that EFL learners respond positively to the use of corpora for self-correction in writing Vannestal and Lindquist (2007) found that students who engaged in corpus-based grammar training exhibited improved attitudes towards learning grammar compared to those who relied solely on traditional grammar books Feedback from questionnaires and interviews indicated that students preferred corpora for academic writing tasks Similarly, Sun (2000) evaluated the responses of 37 Taiwanese EFL college students to corpus-based activities, revealing a strong preference for authentic language texts following a 3-week online training course Students reported that using corpora enhanced their language awareness and sensitivity to specific words and phrases.
Tung et al (2015) investigated the effectiveness of the Corpus of Contemporary American English (COCA) in helping EFL students self-correct language errors in their writing The study involved 12 English-major students from a private university in central Taiwan, who were trained to utilize COCA for revising lexical and grammatical mistakes identified by their instructor over a two-week period Findings indicated that most participants had a positive attitude towards COCA, viewing it as a reliable resource that provided authentic examples, thereby enhancing their understanding of real language usage.
(2004) examined ESL learners‘ behaviors towards the use corpus use in L2 writing
In a study involving 23 ESL students, participants were divided into two groups: 8 intermediate-level and 15 advanced-level students Both groups received identical corpus materials and training from the same instructors, though the intermediate group underwent more extensive training Findings indicated that learners generally viewed corpus use positively, recognizing its effectiveness in improving L2 writing skills Notably, the intermediate group expressed more favorable responses than the advanced group, suggesting that increased exposure to corpus training correlates with more positive attitudes towards its use.
A study by Xue-hua and Ming-ying (2013) investigated the impact of corpus use on L2 writing instruction The findings revealed that incorporating corpus tools significantly enhances vocabulary acquisition in second language classrooms.
The study highlighted that corpus consultation is not equally effective for all learners, as participants displayed mixed attitudes towards its use in writing classes Those with a positive outlook found corpus resources beneficial for understanding word usage in context and making word comparisons Conversely, learners with negative attitudes perceived corpus consultation as time-consuming, feeling overwhelmed by the extensive analyses required.
In summary, the research literature highlights numerous studies on the effectiveness of corpus in second language (L2) writing, demonstrating that corpus-based activities significantly enhance students' language proficiency, grammatical knowledge, and vocabulary acquisition While extensive research has focused on high-proficiency learners with advanced computer skills, there is a notable lack of studies examining undergraduate college students in non-immersion environments Previous findings indicate that students generally hold positive attitudes toward using corpus tools for writing and error correction However, there is a scarcity of research on English as a Foreign Language (EFL) students' experiences and attitudes toward corpus use in Vietnam This study aims to fill this gap by exploring the effectiveness of corpus as a self-correction tool for EFL students in writing, along with the associated benefits, challenges, and student attitudes toward its application.
Methodology
Research Design
This study utilized a mixed-methods approach to examine how a corpus can aid students in self-correcting errors in English as a Foreign Language (EFL) writing The research focused on three key questions regarding the effectiveness of this method.
1/ What types of errors can students correct with the corpus?
2/ What are the students‘ experiences with the use of corpus as self-correction tool for their writings?
3/ What are the students‘ attitudes towards the use of corpus as self-correction tool for their writings?
Johnson et al (2007) gave a definition of mix-methods research:
Mixed methods research integrates both qualitative and quantitative approaches to enhance understanding and provide corroboration, allowing researchers to achieve a comprehensive analysis of their subject matter.
The mixed-methods approach was selected for this study due to its ability to integrate qualitative and quantitative research, providing a comprehensive understanding of research issues while reducing the limitations inherent in each method.
Creswell and Clark (2018) highlight that a mixed methods approach enhances research by combining quantitative and qualitative data, providing deeper insights into a research problem than either method alone This approach grounds findings in participants' experiences and mirrors the natural ways individuals gather information By integrating both data types, researchers achieve a more thorough and nuanced understanding of the research issue.
This study investigates the effectiveness of a corpus as a tool for students' error revision in essays by collecting both quantitative and qualitative data The quantitative analysis focuses on error correction, evaluating how effectively students used the corpus to address errors in their writing It also includes students' assessments of the difficulty levels associated with six specific error types that can be corrected using the corpus, alongside a comparison of proficiency levels in essays from week 1 to week 6 In contrast, the qualitative data derived from learning logs and structured interviews provides insights into EFL students' experiences with corpus consultation during essay revision and their reactions to using the corpus for self-correction in writing.
Participants
In the initial phase of the interview, the researcher developed a series of questions aimed at gathering participants' background information, which encompassed their academic year, chosen major(s), self-assessment of writing and computer skills, as well as their experience with corpus usage.
The research involved 10 English students from HUFL, comprising 3 third-year and 7 fourth-year participants from various majors, including English Pedagogy, Interpretation, Tourism, and English Teaching for Primary Schools These students were selected due to their completion of the Writing 5 module, which equipped them with essential essay writing skills Their familiarity with corpus tools positions them to assess its effectiveness in improving their essay quality Additionally, as they engage in numerous academic assignments, utilizing corpus may significantly aid in self-revision and enhance their overall academic performance.
Interpretation English Pedagogy Tourism English Teaching for Primary
Participants‟ self-evaluation of writing skills
Participants‟ self-evaluation of computer skills
Proficient Advanced Intermediate Preliminary Beginners' level
Advanced Good Average Basic Bad
The analysis of Figures 3.2 and 3.3 indicates that 70% of participants identified as intermediate-level writers, with the remainder considering themselves advanced or proficient Participants exhibited a range of computer skills from basic to advanced, ensuring they possessed adequate English proficiency and technical ability to utilize the corpus effectively Notably, only two out of ten participants had heard of corpus from their lecturers but had never used it, while the others were entirely unfamiliar with the concept Consequently, it can be concluded that participants lacked prior experience in using corpus tools to enhance their writing skills and language competence, allowing for a controlled evaluation of the study's effectiveness and contributions.
Procedure
To ensure effective use of corpora, the researcher organized training sessions for students prior to the study Due to varying schedules, the training was conducted in smaller groups, with two 90-minute sessions held within one week During these sessions, participants were introduced to the COCA and BNC corpora, gaining familiarity with their functions and applications.
In the initial training phase, the researcher provided participants with essential information about the two corpora used in the study: COCA and BNC This included guidance on accessing and exploring these resources, an overview of their interfaces, functions, and instructions for conducting basic search queries Participants were given handouts summarizing this information, and following the session, they received supplementary documents via email that included additional instructions on using COCA and BNC, along with basic search exercises to enhance their familiarity with the corpora.
In the second stage of the study, participants were presented with a sample essay containing highlighted errors, which the researcher used to demonstrate the revision process utilizing COCA and BNC This approach helped participants grasp the essential procedures for the study Throughout the training, participants were tasked with consulting the corpus to correct another sample essay with highlighted errors The researcher then provided feedback on their performance, offered specific guidance on searching within COCA and BNC, and assisted with any difficulties encountered during the error-correction process.
Due to Covid-19, students transitioned to online learning, prompting a researcher to conduct a study where participants wrote weekly essays for six weeks on assigned topics, including argumentative and opinion pieces Each Thursday, the researcher emailed essay prompts, requiring submissions by Saturday, while participants disabled Microsoft Word's auto-correct feature After reviewing the first drafts, the researcher highlighted errors without providing detailed feedback, allowing participants to independently consult two corpora for corrections The revised essays were submitted by Monday, with the researcher tracking errors in categories such as grammar and spelling Ultimately, 56 essays were collected, with most participants completing the assignments, although a few missed one or two.
To evaluate participants' use of the corpus for writing improvement, the researcher utilized Kim's (2009) error analysis method, categorizing corrections with three symbols: + (plus) for improved errors, - (minus) for worsened errors, and 0 (zero) for unchanged errors Corrections were manually assessed based on the researcher’s linguistic expertise, with a focus on determining if participants effectively revised their essays using the corpus This analysis aimed to identify the types of errors and the extent to which the corpus facilitated appropriate revisions.
The researcher was in charge of evaluating the CERF (Common European
The evaluation of students' compositions based on the Framework of Reference for Languages may be limited due to the researcher's status as a non-native speaker To enhance the assessment's accuracy, a Text Analyzer tool was incorporated, offering more compelling evidence of students' improvements in their essays.
The Text Analyzer tool evaluates text difficulty based on the CEFR level by comparing each word to a list of the 10,000 most frequently used English words An algorithm assesses the text's difficulty using word position and average sentence length This analysis allows researchers to identify the appropriate proficiency level for a text and compare the difficulty of different texts By examining essays from students before and after using the corpus, the study highlights the positive impact of corpus consultation on improving writing skills.
Learning logs are recognized as effective tools for enhancing students' cognitive awareness and monitoring their learning processes (Commander & Smith, 1996) These logs encourage reflection on specific academic skills rather than personal feelings about a subject In this study, participants maintained weekly learning logs, documenting errors from their second drafts in a three-column chart: the first column listed the errors, the second identified their types, and the third included corrections made with the help of a corpus Guiding questions prompted students to reflect on their experiences, the benefits of using the corpus for self-revision, and any challenges faced during the correction process The study also assessed improvements in students' writing competence and their attitudes towards using the corpus as a supplementary tool for self-correction.
The study utilized both quantitative methods and structured interviews to assess the effectiveness of corpus in error correction and to gather detailed student opinions on its use This dual approach allowed the researcher to explore students' attitudes regarding the effectiveness, benefits, and drawbacks of corpus, providing valuable insights into factors that may affect their motivation to enhance their writing skills All 10 participants engaged in interviews conducted via Zoom and Google Meet, lasting approximately 20-25 minutes each due to the constraints of the COVID-19 pandemic.
The interview was structured into three sections: the first gathered background information on participants, including their majors, writing and computer skills, and corpus usage experience; the second explored their opinions and experiences using corpus for essay self-revision; and the final section assessed their attitudes towards corpus To ensure accurate data collection, both audio recording and note-taking were utilized The interviews were transcribed for qualitative analysis, with the researcher summarizing responses and using direct quotations to enhance understanding of students' experiences and attitudes towards corpus.
This chapter details the methodology of the study, highlighting the use of mixed methods to enhance understanding of the research problems It describes the participants and the training procedure implemented over six weeks Additionally, it outlines the data collection and analysis processes, showcasing the development and application of various tools, such as error correction analysis, a text analyzer tool, learning logs, and interviews, to effectively gather and analyze the study's data.