Effects of Corpus-Aided Language Learning in the EFL (Grammar Classroom: A Case

Study of Students' learning Attitudes and Teachers' Perceptions in Taiwan

Author(s): MING HUEI LIN
Source: TESOL Quarterly , DECEMBER 2016, Vol. 50, No. 4 (DECEMBER 2016), pp. 871-893
Published by: Teachers of English to Speakers of Other Languages, Inc. (TESOL)

Stable URL: https://www.jstor.org/stable/44984721

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

is collaborating with JSTOR to digitize, preserve and extend access to TESOL Quarterly

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
 Effects of Corpus-Aided Language
Learning in the EFL Grammar
Classroom: A Case Study of Students9
Learning Attitudes and Teachers 9
Perceptions in Taiwan
MING HUEI UN
Tamkang University
New Taipei City , Taiwan

This study employed a blended approach to form an extensive assess

ment of the pedagogical suitability of data-driven learning (DDL
Taiwan's EFL grammar classrooms. On the one hand, the stu
quantitatively investigated the effects of DDL compared with that o
a traditional deductive approach on the learning motivation and self
efficacy of first-year college students majoring in English. On
other, it qualitatively examined a group of early-career teache
(EÇTs') hands-on experience of teaching DDL to these studen
The research results via i-tests show that only those who receiv
DDL treatment have enhanced learning attitudes in general. No
significant differences, however, were found between the effects o
these treatments in a multivariate analysis test of covariance.
contrast, qualitative inquiry shows that, despite facing techni
difficulties and increased workload, the ECTs found their DDL teach-
ing experience innovative and interesting, believed in its effectiveness
in grammar learning, and judged it to transform Taiwanese students'
grammar learning patterns from passivity to active engagement. This
article concludes with suggestions for future DDL applications and
investigation in the EFL grammar classrooms of Taiwan.
doi: 10.1 002/tesq. 250

Corpus-aided
past decade
past decade
(e.g., Campoy,
language Cubillo,
(e.g., Campoy,
Belles-Fortuno,
learning &Cubillo,
Gea-Valor,
has won Belles-Fortuno, many advocates & Gea-Valor, over the
2010; Flowerdew, 2011; O'Keefe, McCarthy, & Carter, 2007). Of all the
applications, data-driven learning (DDL; Johns, 1991) is likely to be the
most notable one (L. S. Huang, 2011), whose characteristics and under-
lying effectiveness have been widely discussed (Boulton, 2010; Gavioli &

TESOL QUARTERLY Vol. 50, No. 4, December 2016 871

© 2015 TESOL International Association

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
Aston, 2001). In one treatment centered on the DDL approach,
students are encouraged to take on the role of researchers and observe,
speculate, and identify/ generalize the contextualized linguistic data pre-
sented in the format of a concordance (Johns, 1991; Johns, Lee, &
Wang, 2008) while their teacher serves mostly to coordinate the
students' research (Johns, 1991). The language investigating process of
the DDL approach encapsulates the nature of discovery learning (cf.
Boulton, 2009; Gilquin & Granger, 2010), which is believed to promote
language learners' grammatical consciousness (cf. Schmidt, 1990) and
provide them with incentives to involve themselves in discussion (Gavioli
& Aston, 2001; Gilquin & Granger, 2010). The discovery learning exer-
cises centered on DDL-based materials are believed to correspond to
current principles of language learning (Gavioli & Aston, 2001) in devel-
oping learner autonomy when students observe authentic linguistic fea-
tures and make their own generalizations (Gavioli, 2009; L. S. Huang,
2011). Additionally, DDL activities reflect the noticing hypothesis dis-
cussed in second language acquisition (cf. Johns et al., 2008), in which
"learners' acquisition of linguistic input is more likely to increase when
their attention is drawn to salient linguistic features" (Flowerdew, 2012,
p. 216). Together, the various advantages may reinforce one another,
leading students' language development and learning attitudes into the
virtuous cycle envisaged in contemporary motivation theory (Dörnyei,
2001; Dörnyei & Ushioda, 2011).
With its appealing features and potential benefits, corpus use has
also yielded much empirical evidence about the effectiveness of using
it in language teaching contexts. Although the reported empirical
results have not yet proved consistent, the mixed findings appear ten-
tatively optimistic (Cresswell, 2007; C. Yoon, 2011). The great majority
of language learners have generally reacted positively to DDL-centered
treatments (Leńko-Szymańska, 2014; H. Yoon & Hirvela, 2004). Specifi-
cally, students enjoyed the process of discovering the linguistic rules
themselves (Phoocharoensil, 2012) and found this a useful and effec-
tive approach to language learning (Charles, 2014; Geluso &
Yamaguchi, 2014; Phoocharoensil, 2012). Students' positive attitudes
to the value of DDL are also reflected in their actual language skills,
with improved collocations (Rezaee, Marefat, & Saeedakhtar, 2014),
retention of vocabulary (Jalilifar, Mehrabi, & Mousavinia, 2014), com-
mand of grammar rules (Smart, 2014), and ability to self-correct and
revise linguistic errors when consulting a corpus (Todd, 2001; Tono,
Satake, & Miura, 2014). Students taught with DDL are found to
become able to apply learned linguistic features in both general com-
positions and academic writing (Z. Huang, 2014; H. Yoon, 2008).
Although the empirical evidence has apparently helped DDL gain a
foothold in language learning communities, some of the drawbacks in

872 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
using DDL still challenge its pedagogic suitability. Obvious examples
are the time needed before students can use and be familiar with a
corpus (Gilquin & Granger, 2010). This issue is often found interwo
ven with students' struggles to draw language features accurately from
the cut-off sentences, partly due to their unfamiliar vocabulary (Geluso
& Yamaguchi, 2014; H. Yoon & Hirvela, 2004). These challenges, a
the students of Kennedy and Miceli (2001) experienced, are at time
discouraging and frustrating. Such experiences may also explain why
certain students, chiefly those with low linguistic ability, have no confi
dence in formulating language patterns when faced with corpus mate
rials (e.g., Boulton, 2010). Although it may be premature to doubt th
suitability of DDL materials because of individual reports of student
feeling challenged and frustrated, more evidence also seems necessar
before students' positive reactions to DDL can be translated into
improved motivation (cf. Daskalovska, 2015; Gavioli & Aston, 2001;
Yang, Wong, & Yeh, 2013). Above all, DDL students' motivational
improvements have received little quantitative assessment.
Similarly, DDL students' self-efficacy levels, as far as I have been
able to find, have not been researched. Self-efficacy is defined as
"[one's] judgments of and confidence in [his or her] personal capabil-
ities to perform certain tasks" and "is believed to ... determine the
amount of effort and time that [one is] willing to devote to a task"
(Lin, 2014, p. 578). If so, it would be helpful to monitor the develop-
ment of students' self-efficacy in learning. Doing this would further
clarify some of the negative perceptions or experiences discussed
above (i.e., concerns about the time and effort spent on DDL and
issues about frustration and confidence). Equally, the results of exam-
ining DDL students' self-efficacy might complement the study of learn-
ing motivation, in view of the positive correlation of self-efficacy and
motivation in particular (Pajares, 2003).
Besides assessing the pedagogic suitability of corpus use from lan-
guage students' perspectives, it is equally important to gain insights
into this subject from language teachers' perceptions. So far, however,
only a few publications have addressed or shared potential problems
reported by DDL teachers with hands-on experience (e.g., Gilquin &
Granger, 2010). What causes most concern is that even teachers some-
times find it time-consuming to prepare suitable corpus-based materi-
als (Gilquin 8c Granger, 2010; Hunston, 2002; Leńko-Szymańska,
2014). This concern may seem unavoidable when corpus-based materi-
als are mostly derived and screened from a mass of corpus data, but
perhaps, as Gilquin and Granger (2010) comment, teachers them-
selves are not yet used to applying and practicing with corpora in class.
Although some teachers have used corpora regularly in their studies
and academic activities, as Tribble (2012) cautions they are not

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 873

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
therefore certain to handle DDL well in class. Notwithstanding a few
teaching demos and the pedagogic discussions in the literature (e.g.,
Campoy et al., 2010; Hunston, 2002; Johns, 1991), training programs
still seem to be needed for in-service (or preservice) teachers or
researchers who want to take a corpus-aided approach (Breyer, 2009;
Jalilifar et al., 2014) before putting DDL into practice.
The DDL approach seems hitherto supported by much theoretical
and empirical evidence, as reviewed above. However, as discussed,
quite a few aspects still require further investigation before this
approach can be generally assumed to have value in language class-
rooms. The pressing research question at this point is this: Judging
from learning and teaching perceptions, is the pedagogical suitability
of DDL-supported treatments in EFL grammar classrooms in Taiwan
justified, in particular when compared with that of a traditional deduc-
tive approach (TDA)?
To answer this research question, the present study performed a
blended assessment in an experiment involving three university-level
English grammar classes in Taiwan, where English is mostly learned as
a foreign language (EFL). The classes were taught in turn by three
groups of early career teachers (ECTs) of TESOL with either blends of
DDL and TDA or pure TDA. I investigated the roles of both the
students and the ECTs to answer the research question more compre-
hensively. Specifically, I wanted first to find whether, after the experi-
ment, the three classes had separately improved their learning
motivation and self-efficacy regarding grammar. Among the three
classes the same variables were then compared statistically to deter-
mine whether any significant differences could be found to suggest the
efficacy of corpus use. Finally, I conducted in-depth individual inter-
views and focus group interviews with the ECTs, and the qualitative
accounts generated from them were used to portray the nature of the
ECTs' experience in teaching with DDL in contrast to that with TDA.

METHOD

The general research design for the current research is

in schematic form in Figure 1 and then explained.

Participants

The undergraduate students. The experimental site was a private

university in Northern Taiwan. Recruited for the project was a
convenient sample from three intact English writing classes (Classes

874 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
 English Department
at a private university in northern Taiwan

Undergraduate program: Postgraduate TE SOL program:

Three writing classes (52 students altogether) A class about Form-Focused Instruction (14 ECTs)

Cla
+ V
Week 1: Weeks 1 & 2: DDL training
An entry questionnaire Basic theories, applications, research papers about
corpora & language teaching // DDL teaching
demo

Week
Class
Class
Class
♦ ♦
Week 6/Pr
Class A: T
Class B: T
Class C: T
♦ *
Week 8/Pr
Class A: T
Class B: Tr
Class C: Tr
+ *

Week 10: Week 10:

An exit questionnaire Interviews

FIGURE 1. Diagrammatic presentation of the resea

A, B, and C) taught by three different teach

I All of the participants were first-year Englis
semester. Class A had 16 students; Class B, 17; and Class C, 19. Gender
distribution in each class was similar: more females than males, in an
average ratio of 2 (male): 5 (female). The participants were mostly
aged between 18 and 19.
To participate in this experiment, every student signed a consent
form. The writing classes in which they had enrolled were part of a
two-credit required course for freshman English majors, and the three
classes had the same writing and grammar textbooks and a consistent
curriculum with the same teaching goals (improving their paragraph-
writing skills and grammar ability) .
Before taking part, the students had learned English for about 8 to
10 years, and their English proficiency levels spanned mostly the B1
and B2 levels in the Common European Frameioork of Reference for Lan-
guages. None of these students had been taught English grammar with

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 875

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
DDL before. An entry questionnaire was conducted to examine all the
students' entry behaviors in terms of their perceptions regarding gram-
mar learning, with results as presented in the Results section below.

The ECTs. The ECTs chosen for this project were an intact class of
14 graduate students who took the elective course Form-Focused
Instruction (FFI) taught by me in the TESOL program of the experi-
mental site. The aim of the course was to acquaint ECTs with various
types of form-focused teaching theories, approaches, or methods and
also to train their ability to teach English language forms (grammar in
particular) to EFL students and students of English as a second Ian-
guage (ESL).
Before consenting to join the experiment, the ECTs had been
informed of the major tasks to complete for this project: three teach-
ing practicums requiring them to teach separate groups using pure
TDA and blends of DDL and TD A. However, because the 14 ECTs
were at different years of study and had diverse course timetables, they
were allowed to choose group members with whom they could cooper-
ate relatively easily, according to their circumstances. As a result, the
ECTs formed three groups, called A, B, and C. Group A had six ECTs
(two males and four females), all second-year master's students, and
Group B had seven (one male and six females), all first-year master's
students. Group C had only one ECT (a second-year MA student) who
volunteered to form Group C alone, given her personal circumstances
and weekly timetable.
Before this project, the ECTs had all had experience, ranging from
6 months to 3 years, teaching Taiwanese students general English skills
at private language schools or at the experimental site. However, none
of them had ever been taught with DDL, learned about this approach,
or used a similar treatment with their students.

Arrangements of the Treatment

To meet the goal of the project and the syllabuses and course
requirements of both the graduate and undergraduate courses, the
three teaching practicums were thus arranged to take place in Weeks
3, 6, and 8 of the 18-week semester. These 3 weeks were also when the
first three grammar units were scheduled for learning by the under-
graduates. The grammar taught to them focused on the concepts of
passives; relative clauses; and some phrases indicating contrast, results,
and purpose. The weeks between the three practicum weeks were
scheduled for lectures on paragraph-writing skills by the original teach-
ers of the classes.

876 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
 Before the teaching practicums started, I had used the first 2 weeks'
FFI (3 hours per week) to acquaint the ECTs with DDL and assign take-
home reading about corpus use. Specifically, the ECTs were given essen-
tial knowledge of corpus linguistics and its application in educational
contexts in the first 2 hours and a book chapter about corpora and lan-
guage teaching by Hunston (2002) in the third. ECTs were then
involved in a 1-hour class discussion about a research paper presenting
two samples of DDL lessons and materials (Johns, 1991). In the last
2 hours, I demonstrated how to create adequate DDL materials and
how to teach with them. After the training, the three groups were ran-
domly assigned the sequences of their practicums.
To gain experience in teaching grammar using DDL, with particular
reference to TDA, the ECTs had to use three different treatments for
the different classes: a pure TDA treatment (Treatment A) to Class A,
a blend of 40% DDL and 60% TDA (Treatment B) to Class B, and
another blend of 60% DDL and 40% TDA (Treatment C) to Class C.
This arrangement exposed each class to only one treatment, so as to
make it feasible to examine whether these treatments alone had differ-
ent effects on the undergraduate participants' learning motivation and
self-efficacy. It should be explained, however, that the proportions in
the different treatments were defined by the percentage distributions
of the material content to be taught using TDA or DDL. For instance,
if a lesson with Treatment B contained 10 grammatical items or con-
cepts, 4 of them (40%) had to be taught with TDA and the remaining
6 (60%) with DDL; the split was decided by each group.
The 20% difference between Treatment B and Treatment C may
seem insignificant at first glance. However, it should be noted that
such an arrangement was concerned with the premise that, as dis-
cussed above, DDL has generally been regarded as a time-consuming
approach for both teachers and students. Additionally, in this project
DDL was totally foreign to both the ECTs and the undergraduate par-
ticipants. Given these circumstances, the minor difference in ratio was
highly likely to result in different perceptions from both parties (ECTs
and students), meriting further discussion. For similar reasons, no
100% DDL or other relatively high proportion of DDL instruction
(e.g., 70%, 80%, 90%) was considered for this project.

Creation of DDL Materials and Typical Lessons of DDL

and TDA

Whereas with Treatment A the ECTs went through the sched

grammar units as presented in the grammar textbook, with Treatm
B and C the ECTs had to create their own DDL materials for the

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 877

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
grammar units that were designated for teaching, because DDL is a
method requiring sufficient authentic examples to illustrate a particu-
lar grammatical concept. Hence, the ECTs compiled examples
(concordance lines) from either the Corpus of Contemporary Ameri-
can English (corpus.byu.edu/coca) or the British National Corpus
(corpus.byu.edu). Both databases are free to access online. The former
contains 450 million English words collected in the early 21st century;
the latter comprises 100 million words collected in the late 20th
century.
Figure 2 illustrates the DDL material that the members of Group A
created to teach their students to distinguish the uses of prepositions
in passive sentences once the students had successfully identified the
typical passive pattern of Subject be p.p. by object.
As Figure 2 shows, a typical DDL session taught by Group A usually
started with a general question (e.g., Question 15 in Figure 2), guiding
students to observe any differences (or rules) in the given concor-
dance lines. The students were then allowed a few minutes to think on
their own or discuss with peers any phenomenon they found. Next
they were encouraged to share their observations with the class. Ide
ally, the ECTs tended to verbally agree with students' correct observa
tion of the rules (e.g., That is a correct observation, you're rìght , Does
anyone else agree}) without providing explanations. Following these, the
ECTs would often double-check the students' understanding, asking
them, for example, to create sentences that illustrated rules they ha
discovered themselves (e.g., Question 16 in Figure 2).
In the DDL process, when the students had difficulties identifying
the designated linguistic patterns, the ECTs would give more detaile
guides to finding the target rule. Sample questions included the
following: "Can you first underline the main verbs of those sentences?"

Question 15. Please observe the two sets of sentences below. What differences do you see?

1 swathe of high-level business is tackled by Mrs. Thatcher in ad hoc groups

2 this is not one of the books mentioned by Mr. Taylor, perhaps because the
3 already have a first degree. It is offered by universities, polytechnics and
4 coatings plant. # Hosted It was hosted by Egon Münk, managing director of
5 obenefit. The Branch has been approached by the Institute of Chartered Account

6 which can be opened with a small key

7 the door to development is locked with the key on the inside
8 The drainage ditches were dug with spades
9 they might have been cut with a circular saw
10

Question 16. Please create a sentenc

FIGURE 2. Example of t

878 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
"Now, please identify what are on the left-hand side of the verbs."
"Please circle all the prepositions." "Did you see any similarities or dif-
ferences in terms of word forms?"
In contrast, a typical TDA session started by explaining the target
rules and then presenting sample sentences illustrating them. This was
also the order followed by the grammar textbook. For instance, in pre-
senting the idea of by + agents and with + instruments in passive sen-
tences, the ECTs would first follow the explanation of the textbook
(e.g., when you describe how something was done, the thing you use
is called the instrument) and then present the example sentence with
emphasis in bold (e.g., The box was opened with a knife).

Data Collection

Two major instruments were used to collect the data for ana
The first was a questionnaire created to investigate each class's l
ing experience with a TDA or a blend of DDL and TDA. The sec
consisted mainly of focus group interviews and a one-on-one interv
intended to explore the nature of the ECTs' teaching experience

The questionnaire. I adapted and revised the questionnair

designed by Lin (2014), producing a 5-point Likert-scaled quest
naire of 16 questions (see Appendix A in the supplementary on
data). Administered to the grammar students before the experi
and again after it, the revised questionnaire was designed to exa
four factors, ranging from self-efficacy in learning grammar
motivation to learn grammar (see Table 1). A pilot study involv
participants, different from those in the main study, was conducte
examine its validity and reliability, following its revision. The r
show a very strong Cronbach's a at .92 for the whole question
and a high Cronbach's a (above .80) for each underlying factor

TABLE 1

Validity and Reliability and Analysis Resulta of the Questionnaire in the Pilot Study

Cronbach's Constructive
Variables Questions alpha validity

Factor 1: Self-efficacy in learning grammar 1-4 .86 44.98

Factor 2: Self-efficacy in applying learned 5-7 .80 11.34
grammar in writing
Factor 3: Self-efficacy in identifying 8-12 .89 8.67
learned grammar
Factor 4: Learning motivation to learn grammar 13-16 .84 7.65
Questionnaire as a whole 1-16 .92 72.64

EFFECTS OF CORPUSAIDED LANGUAGE LEARNING 879

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
Table 1). Additionally, the questionnaire was found to have overall sat-
isfactory constructive validity (72.64%). These results suggest that the
questionnaire was highly valid and reliable.

Interviews. To learn about the relatively wide range of feelings and

ideeis among the ECTs about their teaching experience in this project,
Groups A and B gave separate focus group interviews. Such interviews
may also shed light on differences and/or similarities in perception
between groups of individuals. Group C had only one ECT, so she was
interviewed by herself, because her experience of teaching alone with
a relatively novel approach might also provide valuable insights into
the results.
The focus group interviews lasted at least 30 minutes, and each was
digitally voice-recorded. The records were then transcribed verbatim
for data analysis. To facilitate the collection of data, a group of open-
ended questions (see Appendix B in the supplementary online data)
was used, aimed at comparing the ECTs' hands-on experience of DDL
with that of TDA. Follow-up questions were also used for clarification
when required.

Data Analysis

Quantitative analysis. To address the study's main objective, a num-

ber of analytical methods were employed on the collected data. First,
multivariate analysis of variance (MANOVA) was used to examine
whether before the project started the three writing classes had similar
perceptions of grammar learning. Second, I used a group of paired-
sample i-tests to examine whether the three classes significantly
improved their learning attitudes after the project ended. After this, he
conducted multivariate analysis of covariance (MANCOVA) on the
scores of the overall exit questionnaire and its four underlying factors
to gain insight into the specific differences between the three classes
after the experiment. The overall entry questionnaire scores were trea-
ted as the covariate for conducting the main MANCOVA analysis.

Qualitative analysis. In this research the qualitative approach

adopted to analyze the textual accounts collected from the ECTs'
interviews was the classic phenomenological method of analysis pro-
vided by Moustakas (1994). As Lin, Groom, and Lin (2013, p. 132)
conclude, this method sets out a stepwise procedure that explicitly
allowed the research to move from specific observations of individual
ECTs' experiences to generalized statements about the collective
understandings of these experiences. Thus, I began by producing an

880 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
exhaustive list of every interviewee comment or statement relevant to
the experience (for an example, see Appendix C in the supplementary
online data). I then reduced and eliminated ambiguous statements
from the list and sorted the remainder into groups of invariant con-
stituents , that is, statements expressing single perceptions or perspec-
tives (see Appendix D in the supplementary online data) . The clusters
formed by specific sets of invariant constituents together then served
as a basis for creating the textural and structural descriptions of the par-
ticipants' experiences. Whereas the textural descriptions were focused
on the participants' thoughts and feelings pertaining to these experi-
ences, the structural descriptions accounted for the particular manner
in which these feelings and thoughts were connected with the target
experience. Finally, I combined these two layers of interpretations to
develop a final account of the meanings and essence of the ECTs'
experience as a whole.

QUANTITATIVE RESULTS

Entry Behaviors Between the Three Classes

As Table 2 shows, the MANOVA of the entry questionnaire scores

reveals no statistically significant differences among the three groups,
in that the multivariate F(8, 92) value is only .84 at p = .430, with a
very small effect size (partial eta2 = .081). This result indicates that,
before the project, the three writing classes had similar levels of atti-
tude to learning and using grammar skills, guaranteeing the subse-
quent statistical tests as valid.

Learning Attitudes of the Individual Classes After the

Treatments

Table 3 shows comprehensive comparisons of the attitudes to gram-

mar learning of the writing classes before and after the project.
Clearly, not all the items examined evinced a statistical ¿-value; only
Class C was found to experience enhanced learning attitudes overall
after receiving Treatment C, with a significant difference found for
TABLE 2

MANOVA Results for Questionnaire Scores in the Three Classes

Item Effect Value F Hypothesis df Error df Sig. Partial eta2

Classes Wilks's Lambda .84 1.02 8 92 .430 .081

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 881

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
TABLE 3

Paired-Sample ¿-Tests for the Questionnaire Entry and Exit Scores of the Classes

Item Test N Mean SD df t p r

A Overall questionnaire entry 16 56.25 4.93 15 2.02 .062 .46

exit 16 59.06 7.24
Factor 1 entry 16 14.31 1.89 15 0.14 .889 .04
exit 16 14.25 2.62
Factor 2 entry 16 9.56 1.03 15 2.21 .043 .50
exit 16 10.44 1.26
Factor 3 entry 16 19 2.42 15 0.52 .608 .13
exit 16 19.31 2.89
Factor 4 entry 16 13.38 1.71 15 0.14 .167 .04
exit 16 13.94 2.11
B Overall questionnaire entry 17 55.76 9.82 16 1.36 .191 .32
exit 17 57.53 7.59
Factor 1 entry 17 13.47 2.7 16 0.62 .545 .15
exit 17 13.18 2.56
Factor 2 entry 17 9.88 2.18 16 1.64 .120 .38
exit 17 10.41 1.84
Factor 3 entry 17 17.71 3.8 16 1.02 .324 .25
exit 17 18.53 2.37
Factor 4 entry 17 14.71 2.54 16 0.77 .456 .19
exit 17 14.41 2.6
C Overall questionnaire entry 19 54.74 7.24 18 2.20 .041 .46
exit 19 56.42 8.15
Factor 1 entry 19 13.26 1.66 18 0.37 .716 .09
exit 19 13.37 2.17
Factor 2 entry 19 9.47 1.74 18 0.83 .420 .19
exit 19 9.79 1.99
Factor 3 entry 19 17.79 3.17 18 1.13 .276 .26
exit 19 18.37 3.06
Factor 4 entry 19 14.21 2.76 18 0.65 .525 .15
exit 19 13.84 2.19

the overall questionnaire test scores (t

medium effect size r (.46). Specifically, t
had an overall effect on improving stud
self-efficacy in grammar acquisition. A
Class B revealed the same stronger learn
Class C did, Treatment A was found af
exclusive effect on self-efficacy in terms o
in general), with a statistically signif
¿(15) = -2.21, p = .043, and a large effec

Differences Between the Classes After the Treatments

No significant differences were found between the classes in t

various questionnaire scores on exit (multivariate F = .859, p = .
partial eta2 = .089; see Table 4). This result indicates that there wa

882 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
TABLE 4

MANCOVA for the Exit Questionnaire Scores in the Three Classes, Using the Entry
tionnaire Scores as Covariate

Item Test Value F Hypothesis df Error df Sig. Partial etc ł

Class Wilks's Lambda .830 0.859 10 88 .574 .089

difference between the effects of the treatments on the three classes'

learning attitudes when the difference between the classes in the entry
questionnaire scores was statistically controlled for.

QUALITATIVE RESULTS

This section first presents the interview lengths for each group and
the pseudonym created for each ECT to facilitate discussion (see
Table 5). For reasons of space, however, this section presents only the
main characteristics of the final analysis of the ECTs' overall DDL
teaching experience as a whole.

Final Account of the ECTs' DDL Experience as a Whole

DDL as a new, fun, and great experience to the ECTs. Whereas

TDA was consistently described by all the three groups as a "tradi-
tional" and "boring" approach that Taiwanese students were "way
too familiar with" and in which the ECTs had to "do most of the

TABLE 5
An Overview of the Interviews

Interview duration
Group Pseudonyms Gender (min/sec)
Group A ECT 1 Male 38'44
ECT 2 Male
ECT 3 Female
ECT 4 Female
ECT 5 Female
ECT 6 Female
Group B ECT 7 Female 45 '45
ECT 8 Female
ECT 9 Female
ECT 10 Male
ECT 11 Female
ECT 12 Female
ECT 13 Female

Group C ECT 14 Female 32'20

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 883

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
talking" in class, DDL was generally welcomed and favored as a "fresh,"
"interesting," "cool," "natural," and "flexible" method. Such over-
whelmingly positive perceptions were probably ascribable to the excite-
ment of finding it "great . . . fun ... to teach with concordance lines"
(ECT 14), in that this extraordinary feature not only "allows nonnative
speakers to observe how native speakers actually use grammar" (ECT
11) but also provides grammar teaching with "greater diversity" in
course design (ECTs 5 and 14).

DDL as engaging students in discussion, increasing class interac-

tions, and causing students to think. In addition to describing how
uninteresting TDA was, the ECTs also criticized it because the students
who were taught with TDA in this project were mostly "unresponsive"
(ECT 11), "just sat there and listen [ed]" (ECT 14), and "passively
waited to be fed" grammar rules (Group A). Although it may not seem
uncommon for such passive learning patterns to emerge from TDA
sessions in general, such strong comments are more likely to have
resulted from the comparison with the new experience of DDL. Dur-
ing the DDL sessions the students were judged to be "more focused"
and "engaged" in class (ECTs 6 and 14); they even "actively partici-
pated in peer discussion" (ECT 2). This change further led to "in-
creased interactions" between the ECTs and their students, as ECTs 10
and 14 remarked. The mutually reinforcing benefits in turn were
believed by both Groups A and B to have successfully diverted the
students to a "learning-through-thinking" mode. This achievement was
endorsed by ECTs 4, 5, and 6 as its most valuable contribution to
grammar instruction.

Successful DDL materials coming at a price - Time and

effort. Although the advantages of DDL were highly approved by the
ECTs, producing successful DDL materials was predictably regarded as
a "laborious" task. First, as Group B agreed, in the words of ECT 10,
"searching for the (suitable) data is troublesome and time-consum-
ing

after trying various sets of syntax queries

only about this, but about the time needed
vocabulary in these samples was too difficul
were concerned lest too many unfamiliar w
Similar issues were also reported by Groups
mer highlighted spending the whole weeken
on creating their first batch of DDL mater
with Treatment A, the latter specifically re
on designing DDL materials for roughly 40
tive clauses with Treatment B.

884 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
 What questions to ask: A challenge to meaningful DDL treat-
ments. Producing appropriate concordance lines beforehand was not
the only aspect worrying the ECTs. They also highlighted the chal-
lenge of asking adequate guiding questions to make a meaningful
DDL treatment. For one thing, both Groups A and B struggled with
the use of detailed questions. As ECTs 2 and 8 consistently illustrated,
"if we ask our students, for example, to observe the words on the
right-hand side of the node word and then identify the word forms of
those on the other, isn't it like way too obvious?" More general guid-
ing questions (e.g., observe the table and tell me what you notice), however,
troubled both groups, whose students often seemed "confused" about
what to expect from a DDL table or how to start working with one.
Group C also felt that "the presence of the teacher in a DDL treat-
ment is very important." The students seemed to get lost easily unless
they were "walked through" a DDL task.

DDL changing future EFL grammar instruction. Despite their con-

cerns about coping with DDL, all three groups believed that DDL
treatments are worth the trouble and were ready to integrate this
approach in future grammar classes. While most of the ECTs simply
agreed in conclusion that DDL "is worth the effort because students
do learn from it," ECT 10 elaborated reasons:

I think [DDL] is quite worth it because they [students] are really using
their brains to think and learn

pening between us [ECTs] and students

munication in my teaching, rather than a one-

Agreeing on the overall contributions of

showed great interest in applying a blend o
future classrooms, rather than pure TDA, as th

Successful DDL conditions: Considering lan

ciency and grammar types. Although DDL
support in these and other EFL grammar c
felt that its use probably benefits relatively ad
level learners more than low achievers. As ECTs 8 and 10 observed,
"those students who had no clear idea of word forms or basic sentence
structures were unable to tell what was going on between the words in
the sample sentences or why certain words looked certain ways, such
as p. p. [past participle]." Other group members agreed, in particular
ECT 11, who added that students chosen to receive DDL treatments
"must have certain background knowledge about grammar." ECT 1
echoed this idea that some of the students had no clue what to do
with the concordance lines because they lacked "syntactic concepts."

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 885

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
 In addition to concern over the challenge that low achievers might
sense in a DDL-centered activity, all groups warned against including
certain areas of grammar in DDL, for cost-effectiveness considerations.
Consistently, the three groups suggested that "simple grammar con-
cepts" (e.g., however) or "fixed phrases" (e.g., so that, in order to) should
be taught with TDA because such simple concepts are almost immedi-
ately understandable and easily explicable with TDA. As ECT 14 explic-
itly stated, "it would seem pointless to spend so much time on this,"
particularly when there is "a lot (of grammar) to cover in the limited
time" (ECT 1). In contrast, these comments may suggest that the ECTs
found the laborious DDL treatments more worth using, if not more
effective, with other relatively complex grammar patterns, such as pas-
sives or relative clauses.

DISCUSSION

To shed light on the pedagogical suitability of DDL-cen

ties in the specific EFL grammar-learning context of Taiwan
quantitatively examined the effects of blends of DDL and
students' attitudes to learning grammar. An in-depth
approach was also conducted to investigate exclusively th
spectives on and perceptions of the use of DDL in gra
rooms. The blended research approach generated a varie
results for discussion.

First, it is interesting that only Treatment C was found to signifi-

cantly improve the students' overall learning motivation and self-
efficacy in grammar, but such an effect was not detected with the
other treatments. To begin with, this improvement in itself corre-
sponds to many previous studies reporting favorable comments from
their students on the use of DDL (e.g., Leńko-Szymańska, 2014;
Phoocharoensil, 2012; H. Yoon & Hirvela, 2004). In particular, it
echoes the strong belief held by DDL language learners that such an
approach is useful and effective (Charles, 2014; Geluso & Yamaguchi,
2014; Phoocharoensil, 2012). In addition, the fact that only Treatment
C had such an overall effect justifies the researcher's method design,
recalling that the use of DDL in what may seem trivial differences of
proportion (20% variance) in this project did, as predicted, generate
different results, but in a positive way with regard to the DDL students.
To be specific, this suggests that greater exposure to DDL-centered
activities not only is more likely to cultivate improved learning atti-
tudes in general, but presents DDL as a potentially suitable approach
even to EFL students who are often typified as traditional learners,
such as the Taiwanese undergraduate students of the present study

886 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
(cf. Leel, 2011; Lin, 2009; Tamney & Chiang, 2002). In this regard, it
is worth adding that this finding and interpretation echo the success-
ful research of Tono et al. (2014). They found the pedagogical use of
corpora effective also with Japanese EFL students, whose conventional
educational context generally resembles that of Taiwan. The consis-
tency between the findings of this study and those of Tono et al. may
to some extent moderate the warning by Kihçkaya (2015) against the
use of DDL discovery learning in certain educational contexts.
Although optimistic implications still call for caution, given that the
improvement of Class C was nonsignificant in a statistical comparison
with those of the other classes, the overall positive perceptions of and
perspectives on the use of DDL in the grammar classroom from the
ECiy accounts may explain why certain previous researchers (e.g.,
Cresswell, 2007; C. Yoon, 2011), as well as the present researcher,
would still tend to conclude that DDL use is a promising approach.
First of all, not only did the ECTs warmly endorse their teaching expe-
rience with DDL, but they observed several positive changes in their
DDL grammar classrooms, from increased classroom interactions
between teachers and students to the proactive and engaging learning
models of the DDL students. The ECTs also found that their DDL stu-
dents were stimulated to learn through thinking, an observation in
line with the reflection of those in the study of Oghigian and Chuj o
(2010), who commented that "doing [DDL activities] gave them time
to think about the process and to consider the target vocabulary and
grammar more carefully" (p. 212). Finally, although in this project
DDL was described as a laborious approach, as it is in the literature
(cf. Gilquin & Granger, 2010; Hunston, 2002; Leńko-Szymańska,
2014), its overall effectiveness is believed by the ECTs to be worth the
trouble; it has earned itself a good chance of being incorporated into
their future teaching.
Although the discussion thus far has increased the legitimacy of rec-
ommending a DDL-integrated teaching approach in the EFL univer-
sity-level learning context in Taiwan, several issues still call for
attention before safely putting DDL into practice. The first subject to
address is the exclusively significant gain in self-efficacy in applying
grammar for the non-DDL class. This effectiveness is most probably
ascribable to the exhaustively instructional pattern entailed in the nat-
ure of TDA, with teachers first explaining linguistic rules and features
and then demonstrating how these contribute to example sentences.
Although treatments as such are often criticized as undesirable learn-
ing mechanisms, students being mostly trained to passively imitate
what teachers show them (cf. Leel, 2011; Lin, 2009), it is perhaps
through this repetitive reinforcement - or perhaps continuous assur-
ance - from the authority teacher of the way the target grammatical

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 887

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
patterns are applied that learners can better develop self-confidence in
practicing the learned items. This result, along with the interpretation,
may further indicate an indispensable need to continue considering
such practices in future grammar treatments. Such a drill may appear
especially meaningful when this is weighed: No clear-cut effect has yet
been identified as such with either of the DDL-integrated treatments
used in this project or elsewhere.
Second, as the ECTs observed, a DDL approach may benefit inter-
mediate- to advanced-level learners more than those with limited lin-
guistic abilities, because corpus analysis requires certain intellectu
capacities (cf. Bennett, 2010; Mauranen, 2004). This observation corr
sponds to the report of Chen (2011, p. 71) and echoes the uncertain
raised by Wang, Yang, and Song (2010, p. 347). Whereas the preser
vice teachers in Chen's report were concerned about the ability of lo
level students to specify parts-of-speech tags in corpus-based activities,
Wang et al. question the overall efficiency of corpus-based treatmen
with lower level learners. The ECTs' concern about student proficien
levels may also explain why most of the current literature tends to rec
ommend DDL for advanced learners (cf. Granath, 2009) and why mo
DDL investigations have hitherto been conducted on such students
(cf. Boulton, 2008). This is not to suggest, however, that corpus u
should be avoided when teaching low-level EFL learners, in particula
when increasing numbers of studies have been conducted to resolv
this issue and some have shown positive results with low-level students
or beginning learners (e.g., Boulton, 2010; Oghigian & Chujo, 2010;
H. Yoon & Hirvela, 2004). Instead, more empirical evidence is neede
to clarify the general pedagogic impact of DDL on students at diffe
ent levels (cf. Hirata, Hirata, & Thompson, 2013).
In addition to student proficiency levels, the best questions to ask
students not only are an influential factor in the success of DDL trea
ments, as identified by these ECTs, but directly concern the fund
mental nature and meaning of DDL: discovery learning. First, t
ECTs, who agreed on the usefulness of step-wise questions, consistently
highlighted the importance of guiding students through their DD
tasks instead of asking general questions. This suggestion is consiste
with the advice of Bennett (2010) on the use of explicit questions, i
particular with lower level learners. Although such a practice may seem
helpful, it raises a fundamental question: When DDL students a
expected to take on the role of "language detectives" (Johns, 1997,
101), how many clues and hints (i.e., questions or highlighted key
words) should they receive and in what ways should they be present
for the student detectives to successfully solve cases on their own? The
crucial question is probably this: Would the joy of cracking a case b
lost if the way was made too obvious or too much evidence already

888 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
pointed to a fixed answer? In other words, how much is left for discov-
ery in this scenario and how many incentives remain in these discov-
ery-learning activities? For future DDL treatments, these are questions
worth considering.

CONCLUSION

Given the overall multiple DDL effects found in this st

DDL a highly recommendable approach in university-leve
mar classrooms in Taiwan. However, it must not be for
although the ECTs in this project considered the extr
demanded for DDL treatments worth the effort, Gro
shared such a load as a team. Whether or not most of them will con-
tinue to hold similarly optimistic beliefs about DDL when they tea
on their own - as is most likely - requires further investigation. Th
issue is particularly worth examining when it involves teachers in
long-term course of DDL-centered grammar instruction, rather th
over the short term.

Finally, the design of this project opens up a few opportunities for

further research. First of all, in this study student learning motivation
and self-efficacy were both examined, but learning preferences, which
may have a corresponding effect with either DDL or TDA, were not
considered. Cross-examining learning styles with different blends of
DDL and TDA may best identify the ideal practice in given educa-
tional contexts, in turn verifying whether certain important claims
made in this study are valid. One such claim is that treatments with
larger amounts of DDL will generate more desirable effects on gram-
mar students than treatments without. Similarly, this article includes
no discussion of actual performance by DDL or TDA students in gram-
mar acquisition, so future investigation can contribute by examining
this, notably linking it to students' learning attitudes and/or the
DDL/TDA teachers' hands-on experience. A subtle yet interesting
facet of this line of inquiry is to test whether the ECTs' favoring of
DDL to teach complex grammar (i.e., relative clauses or passives) over
simple phrases is reflected in students' grammar performance.
In addition to considering these aspects, the sample chosen for this
study came from a single site in northern Taiwan, and its size may
seem slightly insufficient to rigorous educational statisticians although,
as Lin (2014) states, writing or grammar classrooms are generally
small. Future DDL studies may thus consider recruiting larger samples
from different regions of Taiwan to further assess the effects of DDL
treatments on Taiwanese students. Likewise, in this project the post-
graduates investigated were early career student teachers who were

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 889

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
equipping themselves with both language education theories and prac-
tical teaching applications and may thus have adapted more readily to
different teaching practices. Hence, it may shed greater light on the
use of DDL in Taiwan's EFL contexts to examine the DDL experience
of in-service teachers in DDL-integrated sessions. Furthering the
understanding of DDL teachers' experience, future practitioners may
also consider using a pedagogic corpus instead of regular corpora like
those used in this study. This practice may ease the difficulties that the
ECTs reported in finding appropriate concordance lines (cf. Hunston,
2002; Willis, 1993, 2003). Last but not least, in this study the ECTs
were allocated only 6 hours of DDL training, which to some extent
may not have familiarized them with the skills necessary to teach confi-
dently with DDL. This may have resulted in some of the ECTs' com-
plaints about the difficulties they encountered in manipulating the
syntax queries to find suitable sentences for DDL material. Future
studies or practices are thus suggested to provide the DDL teachers
with more intensive or longer hours of training to improve their teach-
ing experience.

ACKNOWLEDGMENTS

This article was written with funding support from Taiwan's Ministry
and Technology (MOST 1 03-24 10-H-032-001) and the Key Research Fun
kang University - Development of International Distance Learning Cur
and Compilation of Foreign Language Teaching Materials. I also thank t
and anonymous reviewers of TESOL Quarterly for their many insightful
on earlier versions of this article.

THE AUTHOR

Dr. Ming Huei Lin is an assistant professor in the English Department

University, in Taiwan. He has wide-ranging interests in TESOL, applied
and educational technology.

REFERENCES

Bennett, G. R. (2010). Using corpora in the language learning classroom : Cor

tics for teachers. Ann Arbor: University of Michigan Press.
Boulton, A. (2008). But where's the proof? The need for empirical evid
data-driven learning. In M. Edwardes (Ed.), Proceedings of the BAAL a
ference 2007 (pp. 13-16). London, England: Scitsiugnil Press.
Boulton, A. (2009). Testing the limits of data-driven learning: Languag
ciency and training. ReCALL , 27, 37-51. doi:10.1017/S0958344009000
Boulton, A. (2010). Data-driven learning: Taking the computer out of t
tion. Language learning , 60 , 534-572. doi:10.1111/j. 1467-9922. 2010.00

890 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
Breyer, Y. (2009). Learning and teaching with corpora: Reflections by student
teachers. Computer Assisted Language Learning , 22, 153-172. doi: 10. 1080/
09588220902778328
Campoy, M. C., Cubillo, M. C. C., Belles-Fortuno, B., & Gea-Valor, M. L. (Eds.).
(2010). Corpus-based approaches to English language teaching. London, England:
Continuum.
Charles, M. (2014). Getting the corpus habit: EAP students' long-term use of
personal corpora. English for Specific Purposes, 35, 30-40. doi:10.1016/
j.esp.2013.1 1.004
Chen, H.J. H. (2011). Developing and evaluating a web-based collocation retrieval
tool for EFL students and teachers. Computer Assisted Language Learning, 24, 59-
76. doi: 1 0. 1 080/0958822 1 .20 1 0.526945
Cresswell, A. (2007). Getting to "know" connectors? Evaluating data-driven learn-
ing in a writing skills course. Language and Computers, 61, 267-287. doi: 10. 1163/
9789401203906.018
Daskalovska, N. (2015). Corpus-based versus traditional learning of collocations.
Computer Assisted Language Learning, 28, 130-144. doi: 10. 1080/
09588221.2013.803982
Dörnyei, Z. (2001). Motivational strategies in the language classroom. Cambridge, Eng-
land: Cambridge University Press.
Dörnyei, Z., 8c Ushioda, E. (2011). Teaching and researching motivation (2nd ed.).
Harlow, England: Pearson Education.
Flowerdew, L. (2011). Corpora and language education. New York, NY: Palgrave
Macmillan.
Flowerdew, L. (2012). Corpora in the classroom: An applied linguistic perspective.
In K. Hyland, C. M. Huat, 8c M. Handford (Eds.), Corpus applications in applied
linguistics (pp. 208-224) . New York, NY: Bloomsbury.
Gavioli, L. (2009). Corpus analysis and the achievement of learner autonomy in
interaction. In L. Lombardo (Ed.), Using corpora to learn about language and dis-
course (pp. 39-71). Bern, Switzerland: Peter Lang.
Gavioli, L., 8c Aston, G. (2001). Enriching reality: Language corpora in language
pedagogy. ELT Journal, 55, 238-246. doi:10.1093/elt/55.3.238
Geluso, J., 8c Yamaguchi, A. (2014). Discovering formulaic language through data-
driven learning: Student attitudes and efficacy. ReCALL, 26, 225-242.
doi: 1 0. 1 01 7/S09583440 1 4000044
Gilquin, G., 8c Granger, S. (2010). How can data-driven learning be used in lan-
guage teaching? In A. O'Keefe 8c M. McCarthy (Eds.), The Routledge handbook of
carpus linguistics (pp. 359-369). London, England: Routledge.
Granath, S. (2009). Who benefits from learning how to use corpora? In K. Aijmer
(Ed.), Corpora and language teaching (pp. 47-65). Amsterdam, the Netherlands:
John Benjamins.
Hirata, Y., Hirata, Y., 8c Thompson, P. (2013). Two different types of corpora:
Japanese students' perceptions. In J. Lam, K. C. Li, S. K. S. Cheung, 8c F. L.
Wang (Eds.), Knowledge sharing through technology (pp. 1-15). Berlin, Germany:
Springer.
Huang, L. S. (2011). Corpus-aided language learning. ELT Journal, 65, 481-484.
doi: 10.1 093/ elt/ccr03 1
Huang, Z. (2014). The effects of paper-based DDL on the acquisition of lexico-
grammatical patterns in L2 writing. ReCALL, 26, 163-183. doi: 10. 101 7/
S09583440 1 4000020
Hunston, S. (2002). Corpora in applied linguistics. Cambridge, England: Cambndge
University Press.

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 891

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
Jalilifar, A., Mehrabi, iL, 8c Mousavinia, S. R. (2014). The effect of concordance-en-
riched instruction on the vocabulary learning and retention of Iranian EFL
learners. Procedia - Social and Behavioral Sciences , 98, 742-746. doi:10.1016/j.sb-
spro.2014.03.476
Johns, T. (1991). Should you be persuaded: Two samples of data-driven learning
materials. ELR Journal , 4 , 1-16.
Johns, T. (1997). Contexts: The background, development and trialling of a con-
cordance-based CALL program. In A. Wichmann, S. Fligelstone, T. McEnery, 8c
G. Knowles (Eds.), Teaching and language corpora (pp. 100-115). London, Eng-
land: Longman.
Johns, T., Lee, H.-C., 8c Wang, L. (2008). Integrating corpus-based CALL pro-
grams in teaching English through children's literature. Computer Assisted Lan-
guage Learning, 21, 483-506. doi: 10. 1080/09588220802448006
Kennedy, C., 8c Miceli, T. (2001). An evaluation of intermediate students'
approaches to corpus investigation. Language Learning and Technology, 5(3), 77-
90.

Kihçkaya, F. (2015). Computer-based grammar instruction in an EFL context:

Improving the effectiveness of teaching adverbial clauses. Computer Assisted Lan-
guage Learning, 28, 325-340. doi:10.1080/09588221.2013.818563
Leel, H. C. (2011). In defense of concordancing: An application of data-driven
learning in Taiwan. Procedia^Social and Behavioral Sciences, 12, 399-408.
doi:10.1016/j.sbspro. 201 1.02.049
Leńko-Szymańska, A. (2014). Is this enough? A qualitative evaluation of the effec-
tiveness of a teacher-training course on the use of corpora in language educa-
tion. ReCALL, 26, 260-278. doi:10.1017/S095834401400010X
Lin, M. H. (2009). Inductive and deductive approaches on grammar acquisition. Saar-
brucken, Germany: VDM Verlag Dr. Muller.
Lin, M. H. (2014). Effects of classroom blogging on ESL student writers: An
empirical reassessment. The Asia-Pacific Education Researcher, 23, 577-590.
doi: 10. 1007/s40299-01 3-01 31-8
Lin, M. H., Groom, N., 8c Lin, C. Y. (2013). Blog-assisted learning in the ESL writ-
ing classroom: A phenomenological analysis. Educational Technology and Society,
16(3), 130-139.
Mauranen, A. (2004). Spoken corpus for an ordinary learner. Inj. Sinclair (Ed.),
How to use corpora in language teaching (pp. 89-105). Amsterdam, the
Netherlands: John Benjamins.
Moustakas, C. (1994). Phenomenological research methods. Thousand Oaks, CA: Sage.
Oghigian, K, 8c Chujo, K (2010). An effective way to use corpus exercises to learn
grammar basics in English. Language Education in Asia, 1(1), 200-214.
doi: 1 0.5746/LEiA/ 1 0/Vl /Al 7/Oghigian_Chujo
O'Keefe, A., McCarthy, M., 8c Carter, R. (2007). From corpus to classroom: Language
use and language teaching. Cambridge, England: Cambridge University Press.
Pajares, F. (2003). Self-efficacy beliefs, motivation, and achievement in writing: A
review of the literature. Reading and Writing Quarterly , 19, 139-158. doi:10.1080/
10573560390143085
Phoocharoensil, S. (2012). Language corpora for EFL teachers: An exploration of
English grammar through concordance lines. Procedior-Social and Behavioral
Sciences, 64, 507-514. doi:10.1016/j.sbspro.2012. 11.060
Rezaee, A. A., Marefat, H., 8c Saeedakhtar, A. (2014). Symmetrical and asymmetri-
cal scaffolding of L2 collocations in the context of concordancing. Com-
puter Assisted Language Learning. Advance online publication, doi: 10. 1080/
09588221.2014.889712

892 TESOL QUARTERLY

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms
Schmidt, R. (1990). The role of consciousness in second language learning. Ap-
plied Linguistics, 11, 129-158. doi:10.1093/applin/ll. 2.129
Smart, J. (2014). The role of guided induction in paper-based data-driven learn-
ing. ReCALL, 26, 184-201. doi: 10.1 01 7/S095834401 4000081
Tamney, J. B., Sc Chiang, H. L. (2002). Modernization, globalization, and Confucianism
in Chinese societies. Westport, CT: Praeger.
Todd, R. W. (2001). Induction from self-selected concordances and self-correction.
System , 29, 91-102. doi: 10.1 01 6/S0346-25 IX (00) 00047-6
Tono, Y., Satake, Y., Sc Miura, A. (2014). The effects of using corpora on revision
tasks in L2 writing with coded error feedback. ReCALL, 26, 147-162.
doi: 10. 1 01 7/S095834401 400007X
Tribble, C. (2012, May). Teaching and language corpora: Quo vadis ? Paper presented
at the 10th Teaching and Language Corpora Conference, Warsaw, Poland.
Wang, B., Yang, M., Sc Song, Y. (2010). A study on a computer-based corpus
approach to college English writing. In Q. Zhou (Ed.), 2010 International Forum
on Information Technology and Applications (Vol. 3, pp. 344-347). Los Alamitos,
CA: IEEE Computer Society.
Willis, D. (1993). Grammar and lexis: Some pedagogical implications. In G. Fox,
M. Hoey, Sc J. M. Sinclair (Eds.), Techniques of description: Spoken and -written dis-
course (pp. 83-93). London, England: Routledge.
Willis, D. (2003). Rules, patterns and words: Grammar and lexis in English language
teaching Cambridge, England: Cambridge University Press.
Yang, Y. F., Wong, W. K., Sc Yeh, H. C. (2013). Learning to construct English (L2)
sentences in a bilingual corpus-based system. System, 41, 677-690. doi:10.1016/
j.system. 2013.07.014
Yoon, C. (2011). Concordancing in L2 writing class: An overview of research and
issues. Journal of English for Academic Purposes, 10, 130-139. doi: 10.1 01 6/j.-
jeap.201 1.03.003
Yoon, H. (2008). More than a linguistic reference: The influence of corpus
technology on L2 academic writing. Language Learning and Technology, 12(2),
31-48.
Yoon, H., Sc Hirvela, A. (2004). ESL student attitudes toward corpus use in L2
writing. Journal of Second Language Writing, 13, 257-283. doi:10.1016/
j.jslw.2004.06.002

Supporting Information

Additional Supporting Information may be found in the online ver-

sion of this article:

Appendix A. Learning Motivation and Self-efficacy Beliefs vis-à-vis

Grammar Learning
Appendix B. The main questions used for the interviews
Appendix C. Qualitative analysis: Example of listing interview state-
ments/comments
Appendix D. Qualitative analysis: Example of clusters of invariant
constituents

EFFECTS OF CORPUS-AIDED LANGUAGE LEARNING 893

This content downloaded from

47.72.1.113 on Thu, 19 Oct 2023 07:01:48 +00:00
All use subject to https://about.jstor.org/terms