LEARNING LEXICAL COLLOCATIONS WITH CONCORDANCING AND SCAFFOLDING

(1)

LEARNING LEXICAL COLLOCATIONS WITH

CONCORDANCING AND SCAFFOLDING

THANH HA LE

s1938851

MA in Applied Linguistics

Faculty of Arts

University of Groningen

Supervisor: Dr. Sake Jager

Second reader: Dr. Wander Lowie

(2)

Acknowledgements

I would like to express my deepest gratitude to Dr. Sake Jager, my supervisor, who supported and encouraged me generously throughout this study. Without his expertise and guidance, my thesis would not have been possible.

My sincere thanks also go to my teachers at Department of Applied Linguistics, University of Groningen, for their lectures in MA thesis course, which have always been a source of academic guidance for me over the writing of this thesis.

My appreciation is also extended to my colleagues in Vietnam and my friends in the Netherlands for their kind support and warm encouragement.

(3)

Abstract

(4)

INTRODUCTION

‘You shall know a word by the company it keeps’ Firth (1957:179). It has been generally observed that word meaning can be ‘governed’ by the meaning of other words which tend to co-occur in its environment. There are always restrictions on how words can go together in a language construction (Hammer, 1991; Hill & Lewis, 1997; Nation, 1990; Lewis, 1993, 2000). There are syntax rules which work on word classes rather than individual words and can be explicitly stated. However, there are other restrictions which apply to individual words only. Conventionally, we would rather say a quick glance but a fast car, a quick meal but fast food, not the other way round. It is typical to say a complete disaster rather than a full or a whole

disaster. These restrictions cannot be expressed in terms of rules, but they can be identified as recurrent patterns in any language in their own right. ‘The restrictions on how words can be used together’ (Richards et al., 1992:62) or ‘the way words naturally co-occur’ (Lewis, 2000:132) can be referred to as collocation.

(7)

2 expressions made by non-native speakers are ‘not because of faulty grammar but a lack of collocations’ (Hill, 2000:49). An immediate pedagogical implication is that collocation should be given ‘the same kind of status in our methodology as other aspects of language such as pronunciation, intonation, stress, and grammar’ (Hill, 2000: 59). Teaching collocation should be given ‘a top priority in every language course’ (Lewis, 2000:8)

As the importance of collocations in vocabulary teaching and learning is increasingly recognized, a wide range of classroom activities and resources for teaching and learning collocations has been introduced in the language classroom. With the development of technology, learners are no longer confined to classroom language but can now have access to language corpora or vast databases of authentic texts stored on computers or on the Internet. A concordancing tool can help them obtain samples of naturally-occuring texts from these language corpora. Quite a number of studies have suggested using concordancers to study collocation patterns in authentic texts (Hoey, 2000; Wang, 2002, Chang & Sun, 2009). However, Woolard (2000) points out that the learners could feel ‘overwhelmed by the sheer amount or density of information’. Therefore, some form of training or guidance is needed to provide learners with skills to work with concordancers and language corpora, as Lewis (2000) argues. The argument of the present study is that learners should be provided some type of assistance without which they would not be able to work with corpora and concordancers by themselves. This type of assistance or ‘scaffolding’ can be withdrawn once learners can achieve the task successfully and independently.

(8)

3 but do not naturally go together in English. Given the fact that there are limited applications of authentic resources and technology-supported tools in the teaching situation at her school, the study was an attempt to introduce Collins Wordbanks Online (CWB), an online concordancer and assist her learners in using this concordancer by scaffolding them while they were first working with the concordancer to develop their knowledge of English lexical collocations. The ultimate aims of the study are to raise the learners’ awareness of the importance of learning English lexical collocations and to increase learners’ independence in their own learning of lexical collocations. The study seeks to answer the three following research questions:

1. What is the effect of concordancing on the learners’ performance in lexical collocation tests? i.e. Do the learners perform better in lexical collocation tests when they have access to CWB than when they do not have?

2. What is the effect of teacher’s scaffolding on the learners’ performance in lexical collocation tests? i.e. Do the learners who receive teacher’s scaffolding when working with CWB perform better than those who do not receive teacher’s scaffolding?

(9)

4

CHAPTER I: THEORETICAL BACKGROUND

The theoretical background will be discussed under three subheadings corresponding to three major areas of interest of the study: (1) Collocation, (2) Corpora and concordancing in language learning, and (3) Scaffolding. These three key areas are considered by the study to play an important role in promoting learners’ independent learning of lexical collocations.

I. 1. Collocation

Collocation is a central aspect in the Lexical Approach which was first developed by Michael Lewis in 1993 (Lewis, 1993). It attempts to combine both structure and vocabulary and suggests that language consists largely of ‘chunks’ rather than individual words (Lewis, 1993:95). ‘Language consists of grammaticalized lexis, not lexicalized grammar’, as Lewis (1993) argues. In fact, lexical phrases in any language offer far more communicative and expressive power than grammatical structures can. According to Lewis (1997, 2000), conventional combinations are part of the native speaker’s knowledge of their mother tongue and they can use them naturally with unconscious ease. Schmitt (2000) further supports the Lexical Approach by arguing that the mind stores and processes lexical chunks as individual wholes rather than combinations of single words. As he believes, due to the mind’s limited short-term capacity, it is much more efficient for the brain to recall a ‘chunk’ of language rather than single words and treat the ‘chunk’ as one piece of information.

(10)

5 Language Teaching and Applied Linguistics, 2002:87). Within this conceptual framework, there are two major types of collocations: lexical collocations (recurrent combinations of two lexical content words/ open class words commonly found together) and grammatical collocations (recurrent combinations of one lexical content word/open class word and a grammar function word/ closed class word) (Benson et al., 1986; Hill, 2000; Lewis, 2000; Wu et al., 2010). Lexical collocations can be recurrent word combinations of verb + noun (commit a crime), adjective + noun (an acquired taste), noun + verb (disaster strikes), noun (unit) + noun (a surge of anger), adverb + adjective (bitterly disappointed), verb + adverb (sincerely wish). Grammatical collocations are those words which naturally co-occur together between, for e.g., a noun + preposition (advisor to), a noun + to infinitive/ that clause (a pleasure to, an agreement that), a preposition + noun (in advance), adjective + preposition (proud of), adjective + to infinitive/ that clause (ready to, aware that), or different verb structures (keep on doing sth, remind sb to do

sth).

(11)

6 the nature of collocations, the role of collocation knowledge in facilitating memory and thinking, in L2 acquisition and production. Collocations need to be acquired because they can also be a source of mother tongue interference for L2 learners, as in the researcher’s situation. Her students tend to collocate words in English the same way as they think they can in their native Vietnamese.

(12)

7 Combinations (Benson, Benson & Ilson, 1997), LTP Dictionary of Selected Collocations (Hill & Lewis, 1997), and Oxford Collocation Dictionary for Students of English (2009). Dictionaries of this type can give a ‘much more comprehensive account’ of possible collocates of a word (Woolard, 2000). Furthermore, vast databases of authentic texts are now made available to teachers and learners thanks to technology and the development of the Internet. Empowered by a concordancer, today’s learners can explore a corpus database of authentic language, exploit a large number of examples of words in context, and draw on collocation patterns from these examples by themselves just at the click of a mouse. In the next part of this chapter, discussion will be held on how this approach can facilitate learning collocations.

I. 2. Corpora and concordancing in language learning

(13)

8 rather than traditional ‘drill & kill’ exercises’; (3) ‘A focus on learner-centred activities’; (4) ‘A focus on the use and exploitation of tools rather than ready-made or off-the-shelf learnware’. The raw data of DDL come from language corpora which can be made accessible by using a concordancer. A concordancer selects, organizes and indexes examples of a given word or phrase used in contexts. It displays the typical patterns in which the selected word or phrase is used in the format of key-word-in-context (KWIC) concordances. Thus, a concordance is a list of occurrences of the given key word or phrase which are extracted from a corpus and formatted in separate lines. The following concordance of the key word offer is generated from Collins Wordbanks Online (http://wordbanks.harpercollins.co.uk/), an online concordancer.

Figure 1: Concordance of offer

(14)

9 texts in computer-readable format. Language corpora can vary from general corpora, textbook corpora, multi-lingual corpora to learner corpora, from subject-specific or genre-specific corpora. The use of these corpora depends on the type of activities, the level of teachers’ control, the facilities available (Mishan, 2004). Free online corpora such as British National Corpus (http://www.natcorp.ox.ac.uk/), Collins Wordbanks (http://wordbanks.harpercollins.co.uk/), Hong Kong Virtual Language Centre (http://vlc.polyu.edu.hk/) are commonly suggested by L2 teachers for their students’ consultation (Hadley, 2002; Mishan, 2004). Such corpora of general language are repositories of both spoken and written authentic English from a wide variety of sources such as newspapers, magazines, websites, journals, books, TV broadcasts and radio (Hadley, 2002).

Concordancing, which is accessing a corpus to learn the patterns in which a word or phrase is typically used through a concordancer, has been gradually introduced into the language classroom. Both quantitative and qualitative-oriented research has been conducted so far to measure the effectiveness of using concordances in the classroom, as Sripicharn (2004) observes. A language corpus coupled with a concordancer is ‘an extremely powerful data-driven learning tool’ (Papp, 2004) which can promote the use of authentic materials; support the development of learners’ inductive learning habits, language learning autonomy, and learners’ responsibility for their own learning; and facilitate the cultivation of learners’ explicit knowledge of the target language, their critical language awareness, and other transferable skills (Johns, 1994; Lamy & Mortensen, 2000; Papp, 2004; Stevens, 1995).

(15)

10 computer-assisted concordancers allows exploiting abundant examples of genuine language in use for classroom and self-study activities. These authentic texts can ‘increase learner participation, enable learners to self-discover language features and become aware of words and spoken language’ (Murdoch, 1999). Concerning learners’ learning habits, the change from deductive to inductive approach has also led to the change of the teacher and the learner’s role in the language classroom. The teacher has become ‘a coordinator of research or facilitator’ while the learner now ‘learns how to learn through exercises that involve the observation and interpretation of patterns of use’ (Bernardini, 2004). These ‘transferable skills’ will help them improve their abilities to deal with authentic language and foster their sense of independence as a language learner and their learning skills in general.

As discussed earlier, collocation is among the information about the language which concondances can reveal. Sinclair (1991) further argues that collocation is one major feature of concordances which cannot presented completely and sufficiently in a dictionary. Collocation or ‘the company that a word keeps’ is now made easier than ever thanks to technology. This implies that the most immediate and most potential use of concordances is intended for teaching and learning collocations. In fact, concordancing makes massive examples of words in context readily available to learners. This facilitates learners in developing the skill of ‘noticing’ which is vitally important when it comes to learning collocations (Lewis, 2000:155). ‘Noticing examples in context without formal practice helps turn input into intake’ (Lewis, 2000:199). As argued by Chambers (2005), learners themselves already show positive reactions to concordancing as the result of their growing awareness of descriptive language.

(16)

11 collocations rather than lexical collocations. Hadley (2002) even regards concordancing as a method of studying grammar patterns. Wang (2002), Sun & Wang (2003) and Chang & Sun (2009) attempt to examine the effect of concordancing on learning collocations but their research focus is on grammatical collocations (i.e. verb+prep) rather than lexical collocations. Wu et al. (2010) admit that ‘little research has been done’ on the computer-assisted acquisition of lexical collocations. Wei (1999) suggests that more emphasis should be placed on lexical collocations as this approach would encompass a wide range of lexical words. Lexical collocations need more pedagogical attention as they have more expressive power than grammatical collocations but are ‘more problematic’ and less likely to be found in dictionaries (Čeh 2005). Given these considerations, it is the focus of the present study to examine the effect of concordancing in learning lexical collocations.

(17)

12 ‘massive output’ of concordances (Lewis; 2000; Wang, 2002). Training should involve helping students become more accustomed with naturally-occurring authentic texts and ignore unnecessary information (Lewis, 2000). There has been evidence from the literature that scaffolding can be introduced as a form of support to learners in dealing with corpus data (Wang, 2002; Chang & Sun, 2009). The present study put forward a hypothesis that scaffolding could also be provided as a form of teacher’s support in concordancing. Thus, it is within the scope of this study to investigate the effect of scaffolding on such concordancing. The next part of this section will provide theoretical background on scaffolding which will guide this study to developing the types of scaffolding needed to scaffold concordancing for learning lexical collocations.

I. 3. Scaffolding

(18)

13 The responsibility for task completion gradually shifts from teacher to the learner and is defined by Bruner (1983) as the handover stage.

Scaffolding is closely related to Zone of Proximal Development (ZPD) which is a key term in Socio-Cultural Theory (SCT). Within the SCT framework, Zone of Proximal Development (ZPD) refers to ‘the distance between the actual developmental level as determined by independent problem solving and the level of potential development as determined through problem solving under adult guidance or in collaboration with more capable peers’ (Vygotsky, 1978:86). A learner’s ZPD is the distance between what (s)he is capable of doing unassisted and what (s)he is capable of doing with guidance or support from his teacher or another learner. As for second language acquisition, Ohta (1995:96) defines the ZPD as the ‘difference between the L2 learner’s developmental level as determined by independent language use, and the higher level of potential development as determined by how language is used in collaboration with a more capable interlocutor’. Mitchell & Myles (1998:146) view the ZPD as ‘the domain of knowledge and skill where the learner is not yet capable of independent functioning, but can achieve the desired outcome given relevant scaffolded help’. Lantolf (2002:105) describes the ZPD as ‘the site where future development is negotiated by the expert and the novice and where assistance is offered, appropriated, refused and withheld’.

(19)

14 language is within the learner’s ZPD, the right kind of scaffolding will ideally enable the learner to progress to their potential level extending his/her current level. Scaffolding is beneficial only if the skill to be acquired or the task to be achieved is within the learner’s ZPD. Tasks or skills beyond the ZPD are beyond the learner’s reach and not yet achievable. This implies the important role of language teachers to ‘foresee’ their learners’ potential level of language to provide the right level of scaffolding needed at different time in the course of their second language development.

Scaffolding cannot be used interchangeably for all forms of teacher support within the classroom. There have been some attempts to make a distinction between scaffolding and help (e.g., Maybin et al., 1992; Hammond & Gibbons, 2001; Sharpe, 2001; Wells, 1999). They propose that scaffolding can be distinguished from other forms of support in the following aspects: (1) it is temporary support; (2) it enables the learners to complete the task which would not be possible without support; (3) it is intended to bring the learner to a new level of competence which will enable them eventually to carry out such a task on their own; (4) and it results in the learners having achieved their potential level of competence.

(20)

15 language learners in a new task by providing models in spoken and/or written context, probing questions to elaborate on their learnt knowledge. Long & Sato (1984) and Sato (1986) employ the collaborative discourse of conversational scaffolding to develop learners’ syntactic structures in early language acquisition. In general, scaffolding has been used in early L1 acquisition studies as task-specific explicit strategies rather than only instructions to help learners realize their ZPD.

(21)

16 analogies drawing on students’ experiences; (4) schema building: integrating new information into ‘pre-existing structures of meaning’; (5) re-presenting text: making new language appropriate by transforming linguistic constructions from one genre into another; and (6) developing metacognition: conducting ‘reciprocal teaching’ (see Palincsar & Brown, 1984), think-alouds and self-assessment.

There have been attempts to generalize features of scaffolding in L2 classroom settings (e.g., Van Lier, 2004; Walqui, 2006). Walqui (2006) refined and added up to Van Lier’s descriptions of scaffolding (Van Lier, 2004). ‘Pedagogical scaffolding’, as termed by Walqui (2006), have six following central features:

1. Continuity: tasks are repeated, with variations and connected to one another.

2. Contextual support: exploration is encouraged in a supportive environment with access to means and goals.

3. Intersubjectivity: mutual engagement and rapport are established.

4. Contingency: task procedures are adjusted depending on learners’ actions; contributions and utterances are oriented towards each other and may be co-constructed.

5. Handover/takeover: There is an increasing role for the learner as skills and confidence increase; the learner takes over increasing parts of the action.

6. Flow: Skills and challenges are in balance; participants are focused on the task and are ‘in tune’ with each other.

(Walqui, 2006:165) u

In addition, Applebee (1986) put forward 5 five criteria for effective scaffolding as follows: 1. Student ownership of the learning event: the student has their own role in the task.

2. Appropriateness of the instructional task: the tasks should build upon the knowledge and skills the student already acquire, but should be difficult enough to allow them to expand their knowledge and skills.

3. A structured learning environment: the student is presented with useful strategies and approaches to the task.

4. Shared responsibility: Tasks are accomplished jointly by both the teacher and the learner, and the teacher takes more collaborative than evaluative role.

(22)

17 These generalized features and criteria for effective scaffolding can be regarded as theoretical considerations for the development of the right type(s) of scaffolding needed to support learners in learning a specific target language form or skill in a particular classroom context.

A large body of empirical research in L2 has examined different scaffolding types and processes in different teaching contexts to provide empirical evidence for the significance of scaffolding in L2 acquisition. Although these studies employ different types and/ or processes of scaffolding, they tend to focus exclusively on how L2 teachers provide guided assistance to learners. Aljaafreh & Lantolf (1994) investigate how negative feedback is effectively negotiated in tutor-learner interactions aimed at correcting grammatical errors in written essays. They develop a 13-point regulatory scale of scaffolding, starting from more implicit to more explicit feedback depending on the learner’s ability to notice and benefit from the feedback. Ohta (2000) shows how learners were able to produce utterances using scaffolding in the forms of prompts, co-constructions, and recasts. Hakamäki (2005) investigates the use of scaffolded assistance of an EFL teacher during a classroom interaction, using the same process of scaffolding proposed by Wood et al. (1976). Li (2008) examines how scaffolding may facilitate Chinese EFL learners’ learning grammatical forms at the tertiary level, using seven scaffolding ‘functions’ modified from Wood et al. (1976). Its findings show that scaffolding functions such as evaluation, confirmation request can be of great benefit for students’ learning in the feedback stage.

(23)

18 scaffolding strategies as ‘offering explanations, clarifying and verifying student understanding, linking with student background knowledge, and negotiating meaning’. Woodward-Kron (2007) explores the mode of one-to-one consultation in teacher-student interactions to scaffold students’ writing. Nguyen & Le (2009) enhance learners’ academic writing skills by using a scaffolded approach including teacher’s modeling, teacher-student conferencing. Chang & Sun (2009) examine the use of scaffolding and web concordancers in assisting EFL learners’ proof reading their writing. Three types of scaffolding prompts (i.e., procedural, elaborative, and reflective scaffolding prompts) were used to guide learners to the proofreading process. These scaffolding prompts were found to produce significantly positive effects on the learners’ proof reading performance.

(24)

19 collocations. In the following part of this section, the present study will present theoretical justifications for its selection of specific types of scaffolding for the experiment of the study. Scaffolding has it own role in problem-solving tasks. There has been mounting evidence from literature that scaffolding can facilitate learners’ problem-solving processes; and three following types of scaffolding have been proposed as the guidance throughout learners’ problem-solving processes(Chang & Sun, 2009; Davis & Linn, 2000; Ge & Land, 2004; King, 1991, 1992; Lin & Lehman, 1999; King & Rosenshine, 1993; Scardamalia et al., 1984):

1. procedural scaffolding: instructions to help students follow a problem-solving procedure to complete a specific task;

2. elaborative scaffolding: instructions to facilitate students in activating their reasoning processes and elaborating on their learnt knowledge and experiences; and

3. reflective scaffolding: instructions to promote students’ metacognition and self-reflection. In terms of students’ metacognition, Walqui (2006:176) further argues that scaffolding can promote the development of metacognition skills needed for problem-solving in the following aspects: applying learned strategies while engaging in activity; choosing the most effective strategy for the particular activity; monitoring, evaluating and adjusting performance during activity; planning for future performance based on evaluation of past performance.

(25)

20 1. elicit “learners’ self-explanation, self-questioning, self-monitoring, and self-reflection

during their learning processes”;

2. “guide students in their knowledge construction, knowledge integration, and knowledge representation during their work on complex learning tasks”;

3. “link their arguments or explanations with their existing knowledge”; and

4. “make students’ thinking more apparent and explicit’, ‘better able to recognize areas in which their own understanding is lacking and to engage in knowledge integration”

(26)

21 scaffolding in problem-solving processes, this study establishes a hypothesized effect of scaffolding on concordancing in learning collocations.

(27)

22

CHAPTER II:

RESEARCH METHODOLOGY

The preceding chapter presented the theoretical underpinning for this study, defined the research problems, annotated the findings of relevant literature related to concordancing and scaffolding. In this chapter, the research methodology is presented with a working design of a concordancing tool and a scaffolding scale for the study. Then, the chapter describes the subjects, the materials developed for the study, procedures for data collection and data analysis.

II. 1. Research design

This is a longitudinal study which was conducted in 5 weeks to examine the effect of concordancing and scaffolding on the development of learners’ knowledge about lexical collocations. Learners were assessed in terms of their overall performance in lexical collocation tests and in terms of their ability (1) to identify collocations or mis-collocations, (2) to provide correction with suggested answers, (3) to provide their own correction. Its main instruments are four tests on lexical collocations: a pre-test, test 1/pre-test 2, test 2, and a delayed post-test.

(28)

23 the experiments, which include practice tasks for collocation awareness raising, for concordancing, and for scaffolding.

II. 1. 1. Concordancing tool

(29)

24

II. 1. 2. Scaffolding scale and scaffolding types

A scale of teacher’s instructions was especially designed by the study to scaffold learners’ concordancing. As Woolard (2000:33) suggests, when it comes to learning collocations, it is ‘essential’ for teachers to equip students with search skills for them to ‘discover’ collocations by themselves. Therefore the argument of this study is that scaffolding learners’ concordancing should be primarily aimed at developing their corpus search skills to ‘dig up’ collocation patterns from concordances.

With these perspectives, the study developed a scale of scaffolding of its own from level 5 to 0 (level 5 is when the students received most support from the teacher and level 0 is when the students received least support from the teacher). For each level, different types of scaffolding were assigned with specific scaffolding prompts provided to students. This scale of scaffolding was gradually reduced from level 5 to level 0 as the students were making more correct answers and therefore were receiving less support. This scale corresponds with the following search procedure:

• Orientation,

• Identifying collocation/ mis-collocation, • Selecting the right keyword,

• Analysing concordance output, • Collecting possible collocates, and • Deciding on the best collocate.

(30)

25 progress. Three following types of scaffolding were employed in developing this scale of scaffolding: procedural, elaborative, and reflective scaffolding. Procedural scaffolding was given at the beginning of each scaffolding session. Then elaborative scaffolding and reflective scaffolding were provided while the students were going through the practice tasks on sentence-by-sentence basis.

Search procedure (step 1 to 5)

Description of scaffolding (from level 5 of most support to level 0 of least support)

(0)

Orientation Elicit from the students the specific procedure of concordancing that they are going to follow.

Reinforce students’ understandings about basic grammar categories of noun, verb, adj.

(1) Identifying collocation/ mis-collocation

Ask students to list word combination(s) of verb+noun/ adj+noun in the sentence and check their frequency in CWB.

Confirm collocation/ mis-collocation. (2) Selecting the

right keyword

Direct students’ attention to mis-collocation.

Draw on students’ prior knowledge to identify the keyword, to separate the keyword (the noun) from the mis-collocation, and to generate concordances of the key noun.

(3) Analysing concordance output

Ask students to draw on their linguistic resources to identify relevant patterns of a verb + the key noun / an adj + the key noun and skip irrelevant patterns.

Confirm collocation with Word Sketch function (a function in CWB). (4) Collecting

possible collocates

Elicit from the students possible collocates of the keyword.

Ask students to check their frequency and to confirm results in Word

Sketch (5) Deciding on the

best collocate

Ask students to draw on the meaning of the mis-collocate to select the best collocate among the synonym(s).

Ask students to check and make sure the best collocate can fit in terms of meaning and form in the context.

(31)

26 Procedural scaffolding was needed to engage students in the concordancing process, to establish the specific search procedure, and more importantly, to develop a line of reasoning which is the basis for students to extend their own reasoning during concordancing later. From this starting point, elaborative scaffolding and reflective scaffolding were provided during concordancing to encourage students to jointly construct ideas with teacher, to follow the line of reasoning being developed, to elaborate and to reflect on their own prior linguistic knowledge and their learning experiences. After two scaffolding sessions, students were supposed to improve their own understandings of the concordancing procedure and the reasoning process so that they can independently draw on concordances to develop their knowledge of lexical collocations.

The following are some examples of scaffolding prompts which the teacher (the researcher of the present study) provided to her students at each stage.

Search procedure Scaffolding type and prompts

(5)

Orientation Procedural prompts: defining goal and search procedure

We need to …, and then to ….

How can we…? Why do we…?

It is best then to…

What are the parts of speech of the words in this sentence? (don’t count grammar words)

For example, … is a noun while … is a verb/ adj. (4) Identifying

collocation/ mis-collocation

Reflective scaffolding: monitoring and evaluating

Do you agree or disagree with….?

Is there anything wrong with this combination? Does this sound natural to say…?

We can see from the frequency check that …. (3) Selecting the

right keyword

Elaborative scaffolding: elaborating thoughts, eliciting explanations

What we have found is that … does not sound natural. What is being discussed in this sentence? Why?

The topic is the keyword, so the keyword should be a …(noun)… How does …(the noun)… go with other words?

(32)

27 concordance

output

explanations, monitoring, evaluating

In what way is … related to…?

How can …be used to combine with …(the noun)…? Another combination could be…

From Word Sketch function, we know that this…. is a collocation (1) Collecting

possible collocates

Elaborative scaffolding: elaborating thoughts, eliciting explanations

Which one is used to describe …(the noun)… ? Why is it? Which can happen to …(the noun)…? Why is it?

An example of this is …

From Word Sketch function, we can see these are collocations of …(the noun)…

(0) Deciding on the best collocate

Elaborative and reflective scaffolding: inducing reasoning, prompting for justifications, monitoring and evaluating

Which one is close in meaning to …? This is a collocate, but not a synonym of….. What is a new example of...?

How are ….. and … alike?

Which one is the best… and why?

Compare …. and … with regard to form and meaning.

From Concordance function, we can see that … is a collocation.

II. 2. Subjects

The study was conducted on 20 Vietnamese students of intermediate level of English at Faculty of English, Hanoi National University of Education (HNUE), Vietnam. All of these students belonged to Group A (Intake 58) and were in their second semester of second year, English major. They were placed in the same group as they scored within the same range in the university entrance examination organized by Vietnam Ministry of Education and Training in July 2009. These students already took a fundamental grammar course in their first year and generally have basic computer skills.

II. 3. Materials

(33)

28 Regarding the materials for concordancing and scaffolding sessions, the first set of material was developed to raise student’s awareness of the importance of English collocations. It consists of a concrete definition of collocation, examples of different types of lexical collocations, and follow-up exercises (see Appendix A). The second set of materials was designed to guide students to concordancing (see Appendix B). This set is adapted from CWB User Guide (http://wordbanks.harpercollins.co.uk/Docs/Help/guide.html) to provide students with basic understandings about how to use CWB followed by four practice exercises with CWB concordancing. The concordancing exercises are mainly concerned with how to generate a concordance of a word or phrase and how interpret Concordance and Word Sketch output. The third set of materials was used for teacher’s two scaffolding sessions (see Appendix C). It was also used as concordancing practice tasks for the control group. It comprises two practice tasks in which students have to identify (mis) collocations and provide their own correction of mis-collocations.

(34)

29 not (cf. Lewis 2000 and Hargreaves 2000 argument about testing collocations). Woolard (2000) further supports this approach by arguing that using mis-collocations is an effective method of raising learners’ awareness of collocations. In addition, learners are also tested on their ability to use collocations productively by providing their own correction of collocation errors.

Regarding the test items, each test consists of the same number of ‘verb+noun’ and ‘adjective+noun’ collocations as test items, the same number of collocation errors, and the same number of multiple choice questions. All the tests were restricted to the types of ‘verb+noun’ and ‘adjective+noun’ collocations since empirical research shows that these two types of collocations are those which learners are most likely to make lexical collocation mistakes with (Yan, 2010). In fact, ‘verb+noun’ and ‘adjective+noun’ structure were reported to occupy about 50% and 25% of all lexical collocation mistakes in students’ writing. Besides, for students of intermediate level, it would be more appropriate to choose the two types rather than longer strings of collocates. The restriction to these two types of collocation is also for the matter of consistency in test design, test result analyses and discussion of findings. The lexical words in each structure were checked against the Word List in the students’ coursebook New Cutting Edge Intermediate (Cunningham & Moor, 2005) to ensure that the test items are of the right level of difficulty to the participants. Then the structures were checked against with CWB to ensure that their occurrences in the corpus were significant (as for collocations) or non-significant (as for mis-collocations). This method was adopted in accordance with Gyllstad’s (2007) approach to selecting and creating collocation test items. As a final step of test development, the tests were proof-read by a well-educated native speaker of English.

(35)

30 cover 3 major areas: (1) the subjects’ experiences in computer-assisted language learning, (2) Their attitude toward using CWB as a concordancer (for both groups), and (3) Their attitude toward teacher’s scaffolding (exclusively for the experimental group). The first area was included to evaluate the subjects’ familiarity with computer-assisted language learning for further discussions later about their attitudes and their performance in the tests. The subjects’ attitudes were measured on a five-point Likert-type scale (where 1 is strongly disagree to 5 is strongly agree). An open-ended question was placed at the end of the questionnaire to ask for the subjects’ opinions about the advantages and disadvantages of CWB concordancing (for both groups) and of teacher’s scaffolding (for the experimental group only).

II. 4. Procedures

The procedures of the study with allocated time slots can be summarized as follows:

W Session Control group (N=10) Experimental group (N=10)

1 1 (30m) 2 (25m) 3 (60m) 4 (60m)

Training on collocation awareness

Pre-test 1: without CWB

concordancing

Practice task 1: how to use CWB

Practice task 2: how to use CWB (cont.)

Training on collocation awareness

Pre-test 1: without CWB

concordancing

Practice task 1: how to use CWB Practice task 2: how to use CWB (cont.) 2 5 (25m)

6 (45m) 7 (45m)

Post-test 1/ Pre-test 2: with CWB

concordancing

Practice task 3: CWB concordancing (cont.)

Practice task 4: CWB concordancing (cont.)

Post-test 1/ Pre-test 2: with CWB

concordancing

Practice task 1: CWB concordancing and teacher’s scaffolding

Practice task 2: CWB concordancing and teacher’s scaffolding (cont.) 3 8 (25m)

9 (45m)

Post-test 2: with CWB concordancing

Survey questionnaire on the use of CWB

Post-test 2: with CWB concordancing

(36)

31 First, a brief training session was conducted for the students to become accustomed to collocations, to raise their awareness of the importance of collocations, to develop their ability to recognize collocation patterns as well as collocation errors. Then the pre-test was administered to both two groups without access to CWB. After that, in the next two sessions, the students were explained about key terms such as corpus, concordance, and concordancing. They were also given instructions on how to subscribe to CWB and use its major functions Concordance and

Word Sketch. Then the students had to do concordancing practice exercises.

Subsequently, the post-test 1 was administered to the students to measure whether both groups performed better when they were allowed to use CWB for concordancing. After the post-test 1, the control group kept practicing and working with CWB for the next two sessions and they were given the answers to the exercises at the end of each session. Meanwhile, the experimental group received teacher’s scaffolding in these two sessions and scaffolding prompts were provided in each step of the concordancing procedure. The post-test 2 was then conducted on two groups to measure their performance after they received different treatments. As a follow-up step, two survey questionnaires were administered to two corresponding groups. Finally, the delayed post-test was delivered to the subjects two weeks later to assess whether CWB concordancing and teacher’s scaffolding still had a lasting effect on their performance. The delayed post-test was sent to the subjects via email while other preceding tests were paper-based and conducted in class.

(37)

32

II. 5. Design and analyses

(38)

33

CHAPTER III:

RESULTS AND DISCUSSION

In this chapter, the results and further discussion will be presented on the effect of CWB concordancing and teacher’s scaffolding on the students’ overall performance in lexical collocation tests and three sub-areas of performance: (1) identification of collocations/ mis-collocations, (2) selection of correct collocation with options provided, (3) correction of collocation errors with no options provided. Besides, students’ attitudes toward concordancing and teacher’s scaffolding were also measured.

III. 1. Results

III. 1. 1. Effects of CWB concordancing

Following is the results of the pre-test and post-test 1 scores of all the students when they were placed in two different conditions: without CWB concordancing and with CWB concordancing.

Table 1: Effect of concordancing on students’ overall performance

Treatment N Mean SD t-value df p

Without CWB concordancing 20 9.35 2.39 With CWB concordancing 20 15.65 2.18

-11.20 19 .00

A paired-samples T-test performed on the pre-test and post-test 1 scores shows that the students performed significantly better when they had access to CWB compared to when they had no concordancing tool, t(19)=-11.20, p<.05. There was a significant improvement of all the students from the mean score of 9.35 in the pre-test to 15.65 in the post-test 1.

(39)

34 experienced the same line of development with their score progressing from the pre-test (M=9.30, SD=2.00) to post-test 1 (M=15.30, SD=1.83).

Apart from the students’ overall performance, statistical analysis also presents the students’ performance in three sub-areas as follows:

Figure 2:Effect of concordancing on students’ performance in three sub-areas

(40)

35 provided (df= 2, F=21.65, p<.05). The students’ scores in all the three aspects were significantly improved when they were concordancing with CWB than with no concordancing tool.

Table 2: Effect of concordancing on identification of (mis) collocations

Group Mean SD N Control group 7.10 1.91 10 Experimental group 7.60 1.08 10 Pretest Total 7.35 1.53 20 Control group 11.10 .88 10 Experimental group 10.90 .73 10 Posttest1 Total 11.00 .80 20

With regard to students’ development in their ability to identify (mis) collocations, both groups improved their score extensively with access to CWB. As visually displayed, the mean score for both groups progressed from 7.35 in the pre-test to 11.00 in the post-test 1. The control group improved their mean score from 7.10 to 11.10 and the experimental group advanced from 7.60 to 10.90 in the pre-test and post-test 1 respectively.

Table 3: Effect of concordancing on selection of collocations with guided answers

Group Mean SD N Control group 1.00 1.05 10 Experimental group .70 .82 10 Pretest Total .85 .93 20 Control group 3.00 1.49 10 Experimental group 2.80 1.48 10 Posttest1 Total 2.90 1.45 20

(41)

36 3.00 (as for the control group) and 0.70 to 2.80 (as for experimental group). There was a large standard of deviation in the post-test 1 scores of both groups. This reflects a wide range of scores in the students’ test. 60% of the control group students and 50% of the experimental group students obtained the maximum score for this part but others just scattered along the score range.

Table 4: Effect of concordancing on self-correction

Group Mean SD N Control group 1.30 .95 10 Experimental group .70 .82 10 Pretest Total 1.00 .99 20 Control group 1.90 .99 10 Experimental group 1.60 .95 10 Posttest1 Total 1.75 .97 20

Relating to the students’ development in their ability to provide their own correction, the students also performed significantly better when they were facilitated with CWB. The mean score for all the students progressed from 1.00 in the pre-test to 1.75 in the post-test 1. The control group improved from 1.30 to 1.90 and the experimental group from 0.70 to 1.60 in the pre-test and post-test 1 respectively.

(42)

37

III. 1. 2. Effects of teacher’s scaffolding

The following are the results of Repeated Measures ANOVA performed on the scores of the three tests (post-test 1/ pre-test 2, post-test 2, delayed post-test) of both groups to measure the effect of teacher’s scaffolding on their overall performance.

Figure 3:Effect of scaffolding on students’ overall performance

(43)

38

Table 5: Effect of scaffolding on students’ overall performance

Group Mean SD N

Control group 16.00 2.54 10

Posttest1/

Pre-test 2 _{Experimental group} _15.30 _1.83 ₁₀

Control group 17.00 2.11 10 Posttest2 Experimental group 19.80 .79 10 Control group 17.20 1.55 10 Delayedposttest Experimental group 18.80 1.55 10

The experimental group improved their score significantly from 15.30 in the post-test 1 to 19.80 in the post-test 2 whereas the control group score was averaged at 16.00 and 17.00 in the two respective tests. After 2 weeks, the experimental group performed did not perform as well in the delayed post-test (M=18.80) while the control group performed moderately better (M=17.20) compared to post-test 2. However, the experimental group performance remained significantly higher than that of the control group in the delayed post-test.

Regarding student’s ability to identify (mis) collocations separately, scaffolding did not make a significant difference on the students’ performance (df=2, F=.26, p=.78). The experimental group did not perform significantly better than the control group in this area. Compared to their score in the post-test 1 (M=10.90), the scaffolded group scored marginally higher in the post-test 1 (M=11.40) and in the delayed post-test (M=11.20). The non-scaffolded group was able to improve their score from M=11.10 in the post-test 1 to M=11.30 in the post-test 2 and maintained their mean score at this level in the delayed post-test. For both groups, their score stayed within a limited range from 11.20 to 11.40 in the post-test 2 and delayed post-test.

(44)

39 performances of the two groups was found not statistically significant. Observations during the tests and retrospective interviews after the tests with scaffolded group show that scaffolded students used CWB to check the occurrences and frequency of word combinations rather than single words. These are among the search skills they learned from the scaffolding sessions. The result was that scaffolded students obtained the maximum mean score of 4 in both post-test 2 and delayed test. As for the control group, they maintained their mean score at 3.40 in both post-test 2 and delayed post-post-test due to the effect of concordancing rather than scaffolding. Therefore it can be assumed that although scaffolding did not have significant effect on the performance of the experimental group compared to that of the control group, scaffolding facilitated the experimental group students in adopting the search skills of checking occurrences and frequency of word combinations.

(45)

40

Figure 4: Effect of scaffolding on students’ seff-correction

Descriptive statistics also revealed that the scaffolded group performed significantly better than the non-scaffolded group in both the post-test 2 and delayed post-test.

Table 6: Effect of scaffolding on students’ self-correction

Group Mean SD N

Control group 1.90 .99 10

Posttest1/

Pre-test 2 _{Experimental group} _1.60 _.97 ₁₀

(46)

41 The experimental group was found to make greater progress than the control group in self-correction compared to the other two areas. In the post-test 1, the experimental group score was averaged at 1.60 while the control group score was at 1.90. In the post-test 2, the scaffolded group progressed to 4.40 while the mean score of the non-scaffolded group was 2.30. Although in the delayed post-test, the scaffolded group performed less well than in the post-test 2 with the mean score of 3.60, they were rated higher than the control group whose mean score was 2.40.

III. 1. 3. Students’ experiences in computer-assisted English language learning

The survey questionnaires administered to both groups show that all the students have access to computer at home or at university. The frequency level of using the Internet and computer for English language learning was averaged at M=3.1 and M=3.2 for the control group and experimental group respectively. This is based on 5-point frequency scale from 1 as ‘never’ to 5 as ‘always’. 50% of the control group and 40% of the experimental group used an online dictionary as one of the major online resources for their English learning. 100% of the control group and 90% of the experimental group were offered online learning resources by their teachers but none of them received teachers’ in-class support in using these online resources. No students in either group had ever used a concordancing tool and/ or received teachers’ in-class support in using a concordancing tool before this study was conducted.

III. 1. 4. Students’ attitudes toward CWB concordancing

Table 7: Descriptive statistics on students’ attitudes toward CWB

Control group N=10 Exp. group N=10 Descriptions Mean SD Mean SD

Q1. I enjoy using CWB as support for learning

collocations. 3.60 .67 4.40 .52

(47)

42 Q3. I think CWB is useful in searching for

collocations. 3.10 .88 4.70 .48

Q4. I think CWB is useful in helping me identify

(mis) collocations. 4.40 .52 4.70 .48

Q5. I think I will use CWB as a concordancing tool in learning collocations later in the future.

3.40 .52 4.00 .00

The students’ attitudes were measured on a 5-point Likert scale from (1) strongly disagree, (2) disagree, (3) neutral, (4) agree, to (5) strongly agree. In general, both groups showed positive attitudes toward concordancing (Mean>3).

(48)

43

Figure 5: Students' attitudes toward CWB concordancing

Repeated Measure ANOVA analysis reveals that the interaction between the groups and their attitudes was significant (df=4, F=6.58, p<.05). The experimental group improved their attitudes toward concordancing more than the control group.

(49)

44 delayed post-test. In fact, they all performed better in these three tests with CWB compared to their pre-test score without CWB.

The chart also reveals that the students in both groups agreed least on Question 2 and Question 3. In Question 2, the scaffolded students expressed made more favourable attitudes toward CWB in terms of its ability to support learners than the non-scaffolded group. In Question 3, they also found CWB was easier to use and to search for collocations than the control group. Scaffolding was found to improve the experimental group’s perceptions about CWB concordancing.

In the open-ended question, the students in both groups perceived a number of advantages and disadvantages of using CWB summarized in the following table:

Table 8: Advantages and disadvantages of CWB

Responses Control group N=10 Exp. group N=10 Advantages

Large number of examples of words in context Large number of examples of word combinations Display of frequency of word combinations Support in identifying (mis) collocations

Support in providing correction for mis-collocations

40% 50% 10% 60% 20% 30% 60% 50% 70% 60% Disadvantages

Difficulty to select the right keyword

Too many new words in concordance output Time-consuming

Slow Internet connection to CWB

Internet connection needed to access CWB Difficulty to find correction for mis-collocations

(50)

45 Regarding advantages of CWB, students in both groups share common opinions that CWB can display of large number of examples. As a concordancing tool, CWB can generate concordances of words in context and word combinations. Both groups also have similar views about CWB’s support in identifying collocation patterns and errors. However, they have different attitudes toward CWB frequency check and its support in providing replacement for mis-collocations. The scaffolded students took more advantage of CWB frequency check than the non-scaffolded students. The scaffolded students also found CWB more supportive in helping them to find a replacement for mis-collocations from the concordance output.

On the other hand, one disadvantage of CWB perceived by both groups is that concordance output exposes them to unfamiliar vocabulary. In their opinion, concordancing with CWB was time-consuming due to the slow Internet connection speed at the experiment site and due to unfamiliar vocabulary of the concordance output. Another disadvantage of CWB was that the students need to have Internet connection to access CWB but Internet connection was not always readily available to them. The non-scaffolded students found it more difficult to select the right key word and decide on a synonym to replace a mis-collocation than the scaffolded students.

III. 1. 5. Students’ attitudes toward teacher’s scaffolding

Table 9: Experimental group’s attitudes toward teacher’s scaffolding

Exp. group Descriptions

Mean SD

Q6. I find teacher’s support in my CWB searches useful. 4.60 .52 Q7. I can follow all the steps in the teacher’s support. 4.60 .52 Q8. I was given hands-on practice while using CWB searches. 4.80 .42 Q9. I received immediate support and feedback from teacher while using

(51)

46 Q10. I will recommend using CWB as support for learning English

collocations to other students. 4.10 .32

As displayed, experimental group showed positive attitudes toward scaffolding. For Question 6, they were reported to find teacher’s support during their CWB searches useful (M=4.60, SD=.52). For Question 7 and 8, they could follow all the steps in teacher’s support (M=4.60, SD=.52) and had hands-on practice for each step (M=4.80, SD=.42). For Question 9 and 10, their attitudes changed downward to M=4.30 and M=4.10 respectively but remained positive toward scaffolding (M>4). Retrospective interviews after the survey questionnaire reveal that the students would prefer more immediate support from the teacher. Besides, teacher’s guidance on corpus search skills should be given in a smaller group. In their opinion, concordancing takes time and search skills are needed during concordancing so they could not readily agree strongly that they would recommend CWB to other students.

Table 10: Advantages and disadvantages of scaffolding

Advantages

Guidance on the search procedure Guidance on keyword selection Use of frequency check

Use of Word Sketch function

Support in providing correction for mis-collocations. Guidance on checking meaning and form

Responses 40% 60% 30% 50% 40% 30% Disadvantages Time-consuming

Limited practice task questions

40% 40%

(52)

47 develop corpus search skills such as how to select the right keyword, to check for word frequency, to take advantage of the Word Sketch function in CWB, to identify the best collocate to correct a mis-collocation. Another advantage was that the students were guided through checking the best collocate in terms of meaning and form. However, they could also be aware that guiding through corpus search is a time-consuming process. Besides, they would prefer more practice task questions for them to apply corpus search skills.

III. 2. Discussion

The study sought to examine the effect of concordancing and teacher’s scaffolding on the students’ performance in lexical collocation tests and to measure the students’ attitudes toward concordancing and teacher’s scaffolding.

(53)

48 Relating to the effect of teacher’s scaffolding, the obtained results have suggested that scaffolding made significant improvement in their lexical collocation test scores. These results were consistent with findings of earlier research into the supporting role of scaffolding in teaching language aspects and skills, e.g., scaffolding writing (Woodward-Kron, 2007; Nguyen & Le, 2010), scaffolding reading (Clark & Graves, 2005), scaffolding learning grammar forms (Li, 2008), scaffolding grammatical collocation learning (Chang & Sun, 2009). However, the students’ scores for the three sub-areas of lexical collocation tests reveal that scaffolding did not have the same impact on the students’ three sub-areas of performance. Their test scores indicate that scaffolding did not make a significant difference in terms of the students’ ability to identify (mis) collocations and to select correct collocations among guided options. Scaffolding was indeed found to help the students perform significantly better in the area of providing their own correction of collocation errors in the tests. These findings should have immediate pedagogical implications which will be presented elaborately later in this section.

(54)

49 of new words in concordance output, providing a replacement for mis-collocations. This finding is a supporting evidence for the learners’ evaluation about the disadvantages of corpus searches observed O’Sullivan & Chambers (2006). Furthermore, both quantitative and qualitative data indicate that the scaffolded group increased their positive perceptions toward concordancing compared to the non-scaffolded group. They perceived more advantages and less disadvantages of CWB mentioned above. The scaffolded group was also found to be more motivated in using CWB. They showed more certainty about their ability to use CWB and its supporting role in searching for collocations.

In terms of the experimental group’s attitudes towards scaffolding, the students were generally found positive about teacher’s scaffolding for building search skills and procedure. They also showed their confidence in being able to follow the steps during the concordancing procedure and have hands-on practice in these steps. These findings suggest that scaffolding could reinforce the students’ feeling of certainty when they were working with CWB. However, qualitative data show that they were not fully able to internalize corpus search skills they have been scaffolded. This could account for their desire to be scaffolded in more concordancing practice tasks and receive more immediate support from the teacher.

So far the discussion has been aimed at addressing the research questions the study sought to answer. The rest of the discussion will further interpret the findings about the effects of concordancing and scaffolding on the students’ performance. The discussion will also touch upon pedagogical implications for collocation teaching/ learning practices.

(55)

50 generate concordances of any word or string of words. It gives students immediate exposure to a huge number of examples of words and their collocates in authentic contexts. Empowered by the frequency check function, CWB can facilitate the students in deciding whether co-occurrences of two words are significant to be a collocation. In this regard, concordancing with CWB can assist the students in inducing collocation patterns and therefore mis-collocation patterns as well. From the Lexical Approach, the ability to identify collocations is closely related to ‘one particular skill – noticing’ (Lewis, 2000:155). He claims that ‘collocation is mostly a matter of noticing and recording’ (Lewis, 2000:35). Noticing examples of the input language is ‘central to the acquisition of language’ (Lewis, 2000:163) and ‘crucial to expanding the learner’s mental lexicon’ (Hill et al., 2000: 117). This ability to ‘notice’ patterns is the ‘basis for the development of the independent learning strategies’, as claimed by Woolard (2000:32). Given access to a concordancer, the majority of both groups were able to recognize (mis) collocations from examples indexed in concordances whether they were being scaffolded or not. Scaffolding did not make a significant difference in the performance of the scaffolded compared to that of non-scaffolded group.

(56)

51 guided with corpus search skills and procedure. Evidently, when being scaffolded, the experimental group could follow the search procedure and acquired the skills to search for word combinations and check their frequency. As a result, they gained the maximum score for this area in subsequent tests.

(57)

52 Addressing the difficulty mentioned above, scaffolding in this study made a significant improvement in the area of self-correction. Scaffolding facilitated the experimental group in selecting the right key word and its possible collocates, in considering both meaning and form in context to decide on the best collocate among the synonyms. Being scaffolded, the experimental group could take advantage of the Concordance and Word Sketch function in CWB. These two functions coupled with the frequency check assisted them in finding synonyms and inducing the differences between synonyms through indexed examples. As a consequence, their scores in the area of self-correction were significantly improved compared to those of the non-scaffolded group. Scaffolding can be said to add up to the benefits of concordancing. Concordancing is indeed ‘a useful tool to employ in correction’ (Woolard, 2000:41) once scaffolding is given at the early stages of concordancing. As for learning collocations, being able to provide self-correction to collocation errors is also related to the ability to use collocations more productively. As observed by Chan & Liou (2005:233), a problem with online concordancing is that learners tend to ‘make more progress in receptive collocation tests than productive ones’. Therefore, the findings from this study suggest that scaffolding could help learners make more productive use of lexical collocations.

(58)

53 An empirical study conducted by Sun & Wang (2003) found the same supporting evidence for the inductive approach in learning collocations. The ‘inductive group’, as defined by Sun & Wang (2003), performed significantly better than the ‘deductive group’. In this regard, a concordancing tool can indeed help support learners’ inductive learning habits and promote their independence in learning collocations.

With regard to the learners’ learning level, it is evident from this study that this type of ‘discovery learning’ can be applicable to students of intermediate level. It was once argued that learners at lower than advanced level do not have enough linguistic resources and analytical skills to cope with authentic data (Holec, 1990; Hadley, 2002). However, the study findings show that concordancing had significant effect on the students’ performance at intermediate level, although some signs of uncertainty were shown in their attitudes toward concordancing. This suggests that data-driven learning can be relevant not only to advanced, sophisticated learners (Hadley, 2002) but can also be applicable to learners of lower level such as intermediate. This finding is in line with Yoon & Hirvela (2004) study in which corpus use was also found effective to learners of intermediate level. This implies that further research is needed to explore possibilities of using concordances in classroom for learners of low proficiency.

LEARNING LEXICAL COLLOCATIONS WITH CONCORDANCING AND SCAFFOLDING