THE RELATIONSHIP BETWEEN STATED BELIEFS, CLASSROOM PRACTICES,
AND READING RESEARCH STUDIES OF ORAL CORRECTIVE FEEDBACK
IN FOUR ESL TEACHERS
By
Nobuhiro Kamiya

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Second Language Studies
2012

ABSTRACT
THE RELATIONSHIP BETWEEN STATED BELIEFS, CLASSROOM PRACTICES,
AND READING RESEARCH STUDIES OF ORAL CORRECTIVE FEEDBACK
IN FOUR ESL TEACHERS
By
Nobuhiro Kamiya
A call for a bridge between research and practice is nothing new, yet it seems to be an
impossible task to accomplish. Despite the dedicated efforts made by numerous researchers in
the past, the world is still largely separate between researchers and teachers. One piece of
evidence is that research studies seem to have only a little impact on classroom practices.
However, there is a paucity of studies on this issue in the field of second language teaching.
Therefore, the present study was conducted in the hope of filling this gap, precisely, in order to
investigate how ESL teachers would respond when they read research studies.
Four teachers who were teaching a speaking and listening class at an intensive English
program at a large Midwestern university in the US and their students participated in the present
study. These teachers with varying degrees of teaching experience read three research studies on
oral corrective feedback and a PowerPoint presentation that included a taxonomy of oral
corrective feedback and a summary of each of these three studies. Teachers’ teaching
philosophy, stated beliefs and classroom practices of oral corrective feedback were investigated
through three classroom observations, two interviews, and two stimulated recalls over a
semester. The classroom data were analyzed using AS-units, and a content analysis was
conducted for the data of interviews and stimulated recalls.
The results show that, regardless of their teaching experience, teachers’ classroom
practices of oral corrective feedback remained largely intact. In regards to their stated beliefs on

oral corrective feedback, one of the teachers, who was an MA TESOL student with limited
teaching experience, showed a gradual formation of her stated beliefs on oral corrective feedback
after reading the research studies. However, her stated beliefs were still unstable and vulnerable
in the sense that they showed a radical shift caused by another reading in one of the classes that
she was taking in her program. In contrast, the other three full-time teachers with abundant
teaching experience had already established firm stated beliefs on oral corrective feedback, and
reading the research studies did not have any influence on their practices. However, their
consciousness of oral corrective feedback was raised. What was common among the four
teachers was that, faced with contrasting findings and claims proposed by the three research
studies, they were highly selective with which study they identified; they were basically
choosing the study whose findings or claims corresponded with their stated beliefs on oral
corrective feedback prior to reading.
Finally, the participation in the present study, especially the stimulated recalls, seemed
to have provided the teachers with opportunities to reflect on their stated beliefs and classroom
practices of oral corrective feedback. In this regard, although research studies may not have a
direct impact on the stated beliefs and classroom practices of oral corrective feedback of
language teachers, especially those with ample teaching experience, it can be concluded that
research studies still have an invaluable role to play in their professional development.

ACKNOWLEDGMENTS

My first sincere gratitude should be given to Dr. Shawn Loewen, who has been the
chair of my committee and supported me not only for the completion of this dissertation but also
throughout my career as a PhD student. My second gratitude is dedicated to Dr. Debra Friedman,
who has served as the co-chair and provided me, who was unfamiliar with qualitative research,
with extensive feedback on my dissertation. In addition, I would like to gratefully acknowledge
the invaluable assistance of my two other committee members, Dr. Susan Gass and Dr. Senta
Goertler, who especially helped me to refine my research design. I am also greatly indebted to all
the participants, especially the four teachers who generously agreed to spend their precious time
for the sake of my research in their busy schedules. Furthermore, my special thanks go to my
wife, who has given me encouragement, and my son and daughter, who have given me smiles.
Finally, I would like to thank the graduate school at Michigan State University which kindly
offered me the graduate school dissertation completion fellowship.

iv

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................................vii
CHAPTER 1
INTRODUCTION .................................................................................................................1
CHAPTER 2
LITERATURE REVIEW ......................................................................................................10
2.1 Corrective Feedback ....................................................................................................10
2.2 The Stated Beliefs and Classroom Practices of L2 Teachers ......................................17
2.2.1 Teachers’ Stated Beliefs .......................................................................................17
2.2.2 The Relationship Between Stated Beliefs and Classroom Practices ....................19
2.2.3 Can Teachers Change? ..........................................................................................28
2.2.3.1 The Effect of Teacher Education ...................................................................32
2.2.3.2 The Effect of Teaching Experience ...............................................................48
2.2.3.3 The Effect of Research Studies ......................................................................54
2.3 Conclusion ...................................................................................................................58
CHAPTER 3
METHOD ..............................................................................................................................61
3.1 Setting ..........................................................................................................................61
3.2 Participants...................................................................................................................62
3.2.1 Cecile ....................................................................................................................64
3.2.2 John .......................................................................................................................65
3.2.3 Jim .........................................................................................................................65
3.2.4 Tom .......................................................................................................................66
3.3. Procedures ...................................................................................................................66
3.3.1 Reading the Three Research Studies.....................................................................66
3.3.2 The Schedule of Data Collection ..........................................................................68
3.3.3 Classroom Observations .......................................................................................70
3.3.4 Interviews..............................................................................................................71
3.3.5 Stimulated Recalls ................................................................................................72
3.4 Analyses .......................................................................................................................74
3.4.1 The Classroom Data..............................................................................................74
3.4.2 The Interview and Stimulated Recall Data ...........................................................77
3.4.3 The Procedures of Analyses..................................................................................78
CHAPTER 4
THE TEACHERS’ TEACHING PHILOSOPHY, STATED BELIEFS AND CLASSROOM
PRACTICES OF CF ............................................................................................................81
4.1 Cecile ...........................................................................................................................81
4.2 John ..............................................................................................................................93
4.3 Jim ................................................................................................................................108

v

4.4 Tom ..............................................................................................................................116
4.5 Discussion ....................................................................................................................127
4.5.1 Stated Beliefs on CF .............................................................................................127
4.5.2 Classroom Practices of CF ....................................................................................130
4.5.3 The Relationship Between Stated Beliefs and Classroom Practices of CF ..........132
CHAPTER 5
The TEACHERS’ RESPONSES TO THE READING OF THE RESEARCH STUDIES ...134
5.1 Cecile ...........................................................................................................................136
5.2 John ..............................................................................................................................147
5.3 Jim ................................................................................................................................163
5.4 Tom ..............................................................................................................................174
5.5 Discussion ....................................................................................................................190
5.5.1 The Responses to the Three Research Studies ......................................................190
5.5.2 Stated Beliefs on CF .............................................................................................195
CHAPTER 6
THE EFFECT OF THE READING ON TEACHERS’ STATED BELIEFS AND CLASSROOM
PRACTICES OF CF ..............................................................................................................198
6.1 Cecile ...........................................................................................................................198
6.2 John ..............................................................................................................................210
6.3 Jim ................................................................................................................................224
6.4 Tom ..............................................................................................................................234
6.5 Discussion ....................................................................................................................244
CHAPTER 7
DISCUSSIONS AND CONCLUSIONS ...............................................................................248
7.1 Discussions Overarching the Whole Study..................................................................248
7.1.1 Responses to Each Type of Research Study .........................................................248
7.1.2 Stated Beliefs on CF .............................................................................................250
7.1.3 Classroom Practices of CF ....................................................................................252
7.1.4 Participating in Research as an Opportunity for Self-Reflection..........................253
7.2 Limitations and Ideas for Future Research ..................................................................256
7.3 Recommendations to Stakeholders ..............................................................................258
7.3.1 Researchers ...........................................................................................................259
7.3.2 Pre-Service Teacher Educators .............................................................................261
7.3.3 In-Service Teachers Educators .............................................................................262
7.3.4 In-Service Teachers ..............................................................................................263
7.4 Conclusions ..................................................................................................................264
APPENDICES .......................................................................................................................269
REFERENCES ......................................................................................................................285

vi

LIST OF TABLES

TABLE 1 The extracted results of Li (2010) .........................................................................14
TABLE 2 Brief descriptions of each teacher .........................................................................64
TABLE 3 The procedures for data collection ........................................................................69
TABLE 4 The actual dates of the data collection ..................................................................69
TABLE 5 Cecile’s use of corrective feedback at Observation 1 ...........................................89
TABLE 6 John’s use of corrective feedback at Observation 1 ..............................................99
TABLE 7 Jim’s use of corrective feedback at Observation 1 ...............................................112
TABLE 8 Tom’s use of corrective feedback at Observation 1 ..............................................117
TABLE 9 The reading that the teachers did ..........................................................................135
TABLE 10 Cecile’s use of corrective feedback at three observations ..................................199
TABLE 11 John’s use of corrective feedback at three observations .....................................211
TABLE 12 Jim’s use of corrective feedback at three observations .......................................225
TABLE 13 Tom’s use of corrective feedback at three observations .....................................235

vii

Chapter 1: Introduction

Over the years, the dominant conception of the relationship between research and
classroom practices has been one of implied transmission. There has been an
entrenched, hierarchical, and unidirectional assumption that interpretations developed
and explanations posited through research can – and should – influence in some way
what teachers understand, and therefore what they do, in their classroom practice. Yet,
as we know, this does not happen, which revives the basic question: What is the
relationship between research and teaching? (Freeman, 1996a, p. 89)

There is a gap between research and practice in the realm of applied linguistics as can
be seen in the trend that more attendees choose to participate in only one of the two major annual
conferences in our field, the American Association for Applied Linguistics (AAAL) and the
Teaching of English to Speakers of Other Languages (TESOL; Belcher, 2007), and teachers
expect researchers to bridge the gap between theory and practice (McNamara, 2002). It is also
one of the fields’ important missions to seek better second language (L2) teaching (Ellis, 1997;
Mitchell & Myles, 1998; Pica, 1994) as can be seen in the submission guidelines in TESOL
Quarterly (TQ), which states “TQ is committed to publishing manuscripts that contribute to
bridging theory and practice in our profession” (TESOL International Association, 2012).
Fortunately, previous studies have shown that the engagement with research is
beneficial to teachers (S. Borg, 2009a). Accordingly, researchers have been conducting a whole
host of studies in order to fulfill this duty, and often provide pedagogical implications, which can
be admittedly difficult to write, yet useful (Berliner, 1987). This issue has been a recurrent topic

1

in many professional discussions, and efforts have been made to close the gap thus far for
decades (Allison & Carey, 2007; Belcher, 2007; S. Borg, 1998a; Clark, 1988; Clarke, 1994;
Crandall, 2000; Ellis, 2005; Hall & Loucks, 1982; Kramsch, 1995; Krashen, 1983; Lightbown,
1985, 2000; Markee, 1997; Ortega, 2005, 2012; Shkedi, 1998; Stenhouse, 1981; Stewart, 2006).
In some countries, this movement to connect theory and practice was led at the national level; in
New Zealand (Erlam, 2008; Erlam, Sakui, & Ellis, 2006), and in the UK, whose initiative was
called the evidence-based practice (McNamara, 2002). Moreover, it has been a recurrent topic in
recent major conferences (Ellis, 2009; Fogle, Jacobsen, Friginal, Grieve, & Uccelli, 2010; Harris,
Hellermann, Keck, Olsher, & Santos, 2011; Leung, Hawkins, & Canagarajah, 2010). For
example, Harris et al. (2011) presented a paper titled as “Bridging the research-practice divide
through collaborative research on video-based classroom discourse” at AAAL. Also, Ortega
(2005, 2012) has proposed reevaluation of research by its social utility (i.e., “its potential for
positive impact on societal and educational problems,” 2005, p. 430), educational relevance, and
the moral ends. However, despite these efforts by a group of researchers, unfortunately, research
studies seem to have only little impact on classroom practices not only in L2 teaching but also
school education in general (Eisner, 1984; Stark et al., 1989, as cited in Kagan, 1992a; but see
Hall & Loucks, 1982, for a type of research to solve this problem). For example, among the 19
English as a Second Language (ESL) teachers interviewed in Crookes and Arakaki (1999), only
four acknowledged using research journals to obtain ideas that they could utilize for their
teaching. Chaudron (1988) lamented that “[t]heories and claims about language teaching
methods … have rarely been based on actual research in language classrooms” (p. xv). Some
researchers state that researchers write pedagogical implications “to ostentatiously link the
research to practice,” (Han, 2007, p. 387) which is “often more pretentious than genuine” (p.

2

387), and “premature, far-fetched, excessive” (Magnan, 2007, p. 402). Lightbown (1985) stated
that many of the pedagogical implications of second language acquisition (SLA) research are
“(a) premature, (b) based on research which was extremely narrow in scope, (c) based on
overinterpretations of the data, (d) based largely on intuition, or (e) all of the above” (p. 180).
In the first place, there has been doubt whether or not SLA research should contribute
to teaching (Ellis, 1994; Han, 2007). Such a notion in the field is nothing new; it has been around
for a long time. For example, Tarone, Swain, and Fathman (1976) concluded that SLA research
did not provide any reliable guidelines for teaching. Also, Lightbown (1985) stated that
“[s]econd-language acquisition research does not tell teachers what to teach, and what it says
about how to teach they have already figured out” (p. 182). Krashen (1983) distinguished
between theoretical SLA research and applied SLA research, and claimed that neither one of
them should be applied to practices directly. Hatch (1978) argued that applying the results
obtained from research to pedagogy often required “incredible leaps in logic” (p. 123). Ellis
(1997) even went so far as to state that SLA research may not need to have a direct relationship
with L2 teaching. M. Kennedy (1997) ironically claimed that “the main thing we have learned
from educational research is that we have not learned much from educational research” (p. 4),
and ascribed it to four factors: “(a) The research itself is not sufficiently persuasive or
authoritative…, (b) The research has not been relevant to practice…, (c) Ideas from research
have not been accessible to teachers…, [and] (d) The education system itself is intractable and
unable to change, or it is conversely inherently unstable, overly susceptible to fads, and
consequently unable to engage in systematic change” (p. 4; see also McDonough & McDonough,
1990, for an additional three reasons; see Shkedi, 1998, for reasons for the education field in
general).

3

Related to the first point, in many topics of SLA, researchers have not yet reached any
consensus; only a handful of findings can be presented to L2 teachers conclusively (Bolitho,
1991; Ellis, 1994, 1997, 2005; Han, 2007; but see Long, 1990, for an opposing opinion). Thus, as
Corder (1980, as cited in Ellis, 1994) pointed out, teachers simply cannot wait for those definite
answers to come out, and they must choose to teach in the way that they think the best.
Related to the second point, while teachers are to develop practical knowledge
(Oakeshott, 1962), also called perceptual knowledge (Johnson, 1996a), which is “the procedural
knowledge an individual practitioner has derived from experiences of teaching and learning
languages” (Basturkmen, Loewen, & Ellis, 2004, p. 247), researchers are to advance technical
knowledge (Oakeshott, 1962), also called conceptual knowledge (Johnson, 1996a), which is “the
body of explicit ideas derived by a profession from deep reflection or empirical investigation”
(Basturkmen et al., 2004, p. 246). According to some researchers (Crandall, 2000; Lightbown,
1985; J. Richards & Nunan, 1990), the former is fostered in teacher training, defined as
“approaches that view teacher preparation as familiarizing student teachers with techniques and
skills to apply in the classroom” (J. Richards & Nunan, 1990, xi) whereas the latter occurs in
teacher education (or teacher development in Freeman, 1989), defined as “approaches that
involve teachers in developing theories of teaching, understanding the nature of teacher decision
making, and strategies for critical self-awareness and self-evaluation” (J. Richards & Nunan,
1990, xi; see also Ellis, 1994, and Lightbown, 1985, who both claim that SLA research has a role
only in the latter). Ellis (1998) claims that technical knowledge “cannot easily be applied
off-the-shelf in the kind of rapid decision making needed in day-to-day living” (p. 40; also see
Eraut, 1994, for the same argument). Teachers may presume that “the research knowledge per se
does not articulate easily and cogently into classroom practice, much current knowledge in SLA

4

may be of limited use and applicability to practicing teachers” (Freeman & Johnson, 1998, p.
411) because, according to Markee (1997), “[SLA theory and research] do not address the
real-life concerns of teachers and policy-makers” (p. 81). Similar claims have been made by
several other researchers (Clark, 1988, Clarke, 1994; McDonough & McDonough, 1990). To
support this view, S. Borg (2009a) conducted a survey with 505 English teachers in 13 countries
followed by interviews and additional e-mail questionnaires with selected participants. Among
11 choices, “the results give teachers ideas they can use” was chosen as the third most important
characteristic of good-quality research. Furthermore, one reason why teachers did not read
research studies was that they are “too theoretical and of limited practical use” (p. 370) and of
“limited generalizability” (p. 370). The major reason why the 47 experienced teachers of various
subjects in Israel interviewed in Shkedi (1998) refrained from reading research articles was that
they lacked practicality. S. Borg (2007a) states “[teachers] are unable to see what published
research means for their classroom practice” (p. 744; partial data were also reported in S. Borg,
2007a, 2007b).
Another point to mention related to the second point is that whereas research is
conducted to discover “general truths” (Ellis, 2005, p. 43), teaching needs to be “contingent and
local” (p. 43). Thus, Ellis claims that “research findings based on one particular context of
learning may have little relevance to the particular classroom contexts that individual teachers
find themselves working in” (p. 43).
Finally, related to the third point, namely, a lack of accessibility (Ellis, 1997), research
studies are often inaccessible to teachers “conceptually, stylistically, and sometimes even
physically” (S. Borg, 1998a, p. 274; see also Bartels, 2003). In the first place, they are usually
difficult to comprehend without understanding terminology, which Bolitho (1991) deplored as

5

researchers “blind[ing] an audience with science” (p. 26), and consequently, teachers are simply
“turned off” (Markee, 1997, p. 80). In the questionnaire conducted by McDonough and
McDonough (1990), an English as a Foreign Language (EFL) teacher in Spain wrote “[research]
is often couched in difficult language, conducted in situations that are not immediately familiar
or relevant, presented using complex statistical devices, and hidden in publications of low
circulation (p. 107). Eykyn (1987) also bemoaned the use of jargons from the point of view of L2
teachers. Lack of understanding the language used in research studies and unavailability of
research studies were also mentioned in Shkedi (1998) as reasons why teachers avoid reading
research (see Stenhouse, 1981, for possible solutions for this problem). Furthermore, teachers are
usually unfamiliar with statistics, which is the main type of evidence in quantitative research.
None of the 13 teachers interviewed in Zeuli (1994) understood the statistics used in the research
study that they read. Lastly, it is needless to say that lack of time is also a major reason why
teachers refrain from reading research studies because they are usually occupied in their busy
schedule (S. Borg, 2007a; Crookes & Arakaki, 1999).
These points together often lead L2 teachers to avoid reading research studies. In S.
Borg (2007a), among 50 English teachers, 30.6% answered that they read research studies rarely,
55.1% answered sometimes, and 14.3% answered often. Among the 47 teachers interviewed in
Shkedi (1998), only three teachers stated that they read research studies. In the study conducted
by J. Richards, Gallo, and Renandya (2001) with 112 L2 teachers mostly from Southeast Asian
countries, their beliefs, the changes of their beliefs, and the sources of the changes were
investigated by open-ended questionnaires. Among the results, related to the present study, it was
found that research consisted of only 8.9 percent of all the sources of the changes in their beliefs.
Finally, Pennington and Urmston (1998) conducted a questionnaire with 40 first-year students

6

and 48 third-year students, who were about to graduate from a 3-year BA program in TESOL in
Hong Kong. Relevant to the present study, to the item which asked them to list the components
most helpful for their professional development as teachers, “reading articles” was consistently
ranked at the bottom in both of these student groups. Such a tendency creates “a world in which
teachers talk to teachers about techniques, and researchers and theoreticians talk to each other
about research and theory” (McDonough & McDonough, 1990, p. 103; see also Clarke, 1994 and
Crookes & Arakaki, 1999) as each of the communities establishes a unique discourse that is
incomprehensible to each other (Ellis, 1997; Kramsch, 1995; Markee, 1997). Clarke (1994)
states that this creates “a situation in which one group tends to do the teaching, while another
group does the speculation about how the teaching should be done” (p. 13). Allwright (2005)
calls it “the damaging split between researchers and teachers” (p. 27). Accordingly, teachers may
“dismiss researchers as ivory tower oddities” (Lightbown, 2000, p. 453; “an ivory tower”
appears also in Crookes & Arakaki, 1999, p. 16). Owing to such a discrepancy, Krashen (1989)
even stated that theorists have lost the respect of L2 teachers. As an instance, lack of trust in
research studies was one of the factors that prevents teachers from reading articles (Shkedi,
1998). A teacher in Crookes and Arakaki (1999) sarcastically remarked, “If you get the
professors from [an ESL teacher education program] and bring them … here and give them a
week to teach, they will probably miserably fail” (p. 16). In McNamara (2002), a teacher left
such a comment: “the wrong people [i.e,, researchers] coming up with the wrong suggestions,
poorly researched and then left to teachers on the ground to implement” (p. 17), “[research is] a
boring waste of time – involving loads of paperwork” (p. 20), and “people [i.e., researchers]
being paid to invent new names for tried and tested practices and then presenting them as
innovations” (p. 20). Still, as Ellis (1997) claimed, although SLA research may not have an

7

effect on teachers’ behaviors directly, it has a potential to indirectly influence their cognitions
and personal theories “either by helping them to make explicit their existing principles and
assumptions, thereby opening these up to reflection, or by helping them to construct new
principles” (p. 82). He calls this transformation rather than transmission. In an expansive
literature review for L2 teachers, Ellis (2005) concludes:
[A] literature review such as this is not to prescribe or proscribe what teachers should
do to ensure effective learning in their classrooms but to stimulate reflection on the
complex phenomenon of instructed language learning and a willingness to experiment
with new approaches in accordance with their local conditions. (p. 44)

Thus, there is a place for research in L2 teaching. But what exactly should its role be?
The present study tackles this issue; namely, how do L2 teachers respond to SLA research
studies?
The field of SLA has dealt with a variety of topics, some of which seem to be more
relevant to teaching than others. It should be noted that topics considered important in SLA may
not be considered so in the field of pedagogy. For example, Brindley (1990, as cited in Ellis,
1997) asked adult students in his postgraduate TESOL class in Australia to name the topic that
they thought the most relevant to their concerns among all the topics covered in the class. They
chose the developmental sequences in psycholinguistics as the bottom, which is “generally
considered of central importance by SLA researchers” (Ellis, 1997, p. 77).
The present study investigated L2 teachers’ stated beliefs and classroom practices of
oral corrective feedback (henceforth, CF1; viz., oral error correction). CF was chosen for the

1

CF can also take the form of writing. However, the present study focused on oral CF.
8

following reasons: (a) the topic is easy to apply to pedagogy (Ellis, 1997; Vásquez & Harvey,
2010); (b) L2 teachers are familiar with the topic because most make use of it in their classrooms
(Y. Sheen, 2004) and generally believe that it is beneficial (Leeman, 2007); (c) there has been an
accumulation of both descriptive and experimental studies on the topic; and (d) despite these
benefits, it is “an area in which SLA researchers and instructors historically have not seen eye to
eye” (Leeman, 2007, p. 113).

Therefore, CF is used to refer to oral CF.
9

Chapter 2: Literature Review

2.1 Corrective Feedback
In the history of language teaching, a host of teaching methods and approaches has
come into vogue, some of which have seen better days. Regardless of which methods or
approaches are adopted, perhaps with the exception of the Grammar Translation Method, one of
the common objectives in L2 teaching has been to enable the learners to be able to communicate
in the target language through cultivating their communicative competence (Hymes, 1971). As
the culmination of such a movement, Communicative Language Teaching (CLT) was introduced
between the late 1970s and early 1980s (Larsen-Freeman, 2000).
In the strong form of CLT (Howatt, 1984), the focus of the lesson is exclusively
meaning; developing fluency has priority, and accuracy is largely ignored as long as the meaning
can be correctly conveyed. In contrast, in the weak form of CLT, there is an occasional shift of
attention to form, called Focus on Form (FonF; Long, 1991, 1996), which “overtly draws
students’ attention to linguistic elements as they arise incidentally in lessons whose overriding
focus is on meaning or communication” (Long, 1991, pp. 45-46).
There are many ways to categorize FonF, one of which is to divide it between
proactive and reactive (Long & Robinson, 1998). The former, often called preemptive FonF, is
defined as FonF in which “attention is given to linguistic items identified as problematic,
although no immediate error has been produced” (Loewen, 2011a, p. 579) whereas the latter,
commonly known as CF, is defined as FonF in which “attention to form can occur sporadically
in response to any errors produced by the student” (p. 579).2 To date, CF has been the focus in

2

These definitions are for extensive FonF. Intensive FonF has different definitions.
10

many more previous studies than proactive FonF although L2 teachers may provide both of them
in the classroom (Alcón, 2007; Ellis, Basturkmen, & Loewen, 2001a; Williams, 1999; Zhao &
Bitchener, 2007). Because CF is the focus of the present study, only the reactive FonF
(henceforth; FonF) will be discussed.
FonF is one of the topics that has been extensively investigated in the field of SLA
partly because it is important both theoretically and pedagogically. Accordingly, a whole host of
descriptive studies have been conducted thus far. One of the seminal studies was conducted by
Lyster and Ranta (1997) who analyzed FonF provided in six French immersion classrooms
(Grades 4 and 6) in the Montreal area. They categorized FonF into six types, and used uptake,
defined as “student responses to corrective feedback” (p. 40), as the measurement of
effectiveness. The results showed that recasts, a term which was first introduced to the SLA field
and defined by Long (1996) as “utterances that rephrase a [learner’s] utterance by changing one
or more sentence components (subject, verb, or object) while still referring to its central
meanings” (p. 434), were used the most despite their ineffectiveness in promoting uptake (but
see Long, 2007, for a different view). Using the same database, Lyster (1998a) shows that FonF
was provided for a variety of errors: grammatical errors (46%), lexical errors (24%),
phonological errors (19%), and unsolicited L1 use (11%). In another study, Y. Sheen (2004)
compared the data from four settings: (a) Canadian immersion, (b) Canada ESL, (c) New
Zealand ESL, and (d) Korean EFL, and showed that recasts were used the most frequently
among all kinds of FonF, regardless of context. Numerous previous studies have shown that
FonF is commonly provided in L2 classrooms for many kinds of errors across many educational
contexts (e.g., ESL in Loewen, 2003; EFL in Yorozuya, 1999; Japanese as a foreign language in
Yoshida, 2010).

11

As FonF has accumulated a number of descriptive studies, researchers have also
conducted experimental studies to investigate the effectiveness of FonF using measures other
than uptake. Saturated with myriads of experimental studies, recently some researchers have
started to synthesize the previous studies in the form of meta-analysis.
One consensus that all the meta-analyses conducted thus far have reached is that CF is
more effective than no CF, at least to some extent (Li, 2010; Lyster & Saito, 2010; Mackey &
Goo, 2007; Russell & Spada, 2006). For example, Russell and Spada analyzed 15 studies on CF
and written corrective feedback for grammar published between 1988 and 2003, and found that
CF was more effective than no CF (d=1.16); nevertheless, the results regarding each type of CF
often seem to be contradicting between some meta-analyses as shown in the following
paragraphs.
One of the seminal (but not the oldest) and influential meta-analyses in SLA was
conducted by Norris and Ortega (2000), which analyzed 49 unique sample studies published
between 1980 and 1998. The effectiveness of L2 instruction was defined in light of test scores.
With regard to the present study, they analyzed the effectiveness of two types of CF: recasts and
metalinguistic feedback, the former of which is commonly regarded as an implicit type of CF
(Ellis, Loewen, & Erlam, 2006, 2009; Li, 2010; Lyster, 1998a; Lyster & Ranta, 1997). The latter
is defined as “comments, information, or questions related to the well-formedness of the
student’s utterance, without explicitly providing the correct form” (Lyster & Ranta, 1997, p. 47),
and commonly regarded as an explicit type of CF (Ellis et al., 2006, 2009; Li, 2010; Lyster,
1998a; Lyster & Ranta, 1997). The results showed that metalinguistic feedback was slightly
more effective than recasts (d=0.96, d=0.81, respectively), being congruous with the overall
finding in Norris and Ortega (2000) in that the explicit type of instruction was more effective

12

than the implicit type of instruction. However, the researchers cautioned that the stronger
effectiveness of the explicit type of instruction may be attributable to the fact that most of the
tests used for measurement in the studies included in their meta-analysis were advantageous to
the explicit types of instruction because those were likely to measure explicit knowledge, which
led Doughty (2003) to say that “the case for explicit instruction has been overstated” (p. 274).
The work of Norris and Ortega (2000) was followed by a series of meta-analyses in
SLA with different criteria for the inclusion and exclusion of studies. Mackey and Goo (2007)
conducted a meta-analysis with 28 unique sample studies published between 1990 and June
2006. They categorized CF into three kinds: recasts, negotiation, and metalinguistic feedback.
Negotiation includes clarification requests, comprehension checks, and confirmation checks.
Among many analyses, the results regarding CF showed that recasts were more effective than
metalinguistic feedback (d=0.96, d=0.47, respectively), which clearly contradicts Norris and
Ortega (2000).
It was necessary to wait for ten years after Norris and Ortega (2000) in order to see a
meta-analysis that included PhD dissertations in an attempt to avoid publication bias for the first
time. Li (2010) chose 22 published studies and 11 PhD dissertations of CF and investigated the
effectiveness of the five kinds of CF: recasts, negotiation, clarification, explicit correction, and
metalinguistic feedback. Explicit correction is defined as CF that provides the correct form while
“clearly [indicating] that what the student had said was incorrect” (Lyster & Ranta, 1997, p. 46).
In addition, he categorized the former three types as implicit CF and the latter two types as
explicit CF. The timing for the posttest was categorized into three kinds depending upon how
many days passed since the end of the treatment: immediate (less than 7 days), short-term
delayed (8-29 days), and long-term delayed (more than 30 days). The extracted results are

13

summarized in Table 1.
Table 1
The Extracted Results of Li (2010)
Immediate
Short-term delayed
Long-term delayed
Recasts
0.506
0.439
0.533
Clarification
0.531
Explicit correction
0.877
Metalinguistic feedback
0.581
0.518
Implicit CF
0.542
0.444
0.544
Explicit CF
0.693
0.608
0.440
Note. The values represent the effect sizes (Cohen’s d) in the fixed-effects model.

Overall, the results showed that explicit CF was more effective than implicit CF.
Especially comparing recasts and explicit correction, the difference in d was 0.371 in the
immediate posttest. Nevertheless, what is noteworthy is that these numbers reversed in the
long-term delayed posttest. Li (2010) speculated that implicit CF may be more effective to
develop implicit knowledge; therefore, its effectiveness is more durable than explicit CF which
may have a direct effect on explicit knowledge. This shows that explicit CF may be more
beneficial in a short term, but in a long run, implicit CF may be more effective.
As can be seen from the contradicting results of these meta-analyses, there seems to be
no consensus as to which type of CF is effective to what extent. Among all the kinds of CF,
recasts have been the target of the most empirical studies thus far, and its effectiveness varies
greatly across the meta-analyses. There are two possible reasons. One is that, as stated above,
although recasts are commonly regarded as an implicit type of CF, “recasts themselves range
from more implicit to more explicit” (Lyster & Saito, 2010, p. 278). Therefore, it is
understandable that the results of the implicit type of instruction do not accord with those of
recasts. It is unknown how explicit or implicit recasts were in each study; thus, the hypothesis
that the explicit types of CF is more effective than the implicit types of CF, which was claimed

14

in Norris and Ortega (2000), cannot be substantiated yet. Another reason is that both laboratory
studies and classroom studies were included in the meta-analyses. It is generally claimed that
participants are more likely to notice CF in the laboratory setting more than in the classroom
setting. Therefore, even when CF is, or is classified as, implicit by researchers, it may be explicit
to participants; thus, its effectiveness becomes as strong as that of explicit CF. Indirectly
supporting this hypothesis, some of the meta-analyses discussed above compared laboratory
studies and classroom studies, and showed that CF in the lab is more effective than that in the
class (Mackey & Goo, 2007, d=0.96, d=0.57; Li, 2010, d=1.09, d=0.472, respectively).3 Thus, in
order to truly understand the effectiveness of each CF in L2 classroom, a meta-analysis should
include only classroom studies.
Such a study was conducted by Lyster and Saito (2010), who analyzed 15 published
studies, restricting the studies to only quasi-experimental classroom studies. They categorized
CF into three kinds: recasts, explicit correction, and prompts. Prompts include clarification
request, repetition of errors, elicitation, and metalinguistic clues. The results showed that,
regarding between-group contrasts, recasts were the least effective (d=0.53) and both explicit
correction and prompts had large effect sizes (d=0.84 and d=0.83, respectively). Regarding
within-group contrasts, explicit correction was the least effective (d=0.60), followed by recasts
(d=0.70), and only prompts had a large effect (d=1.14). In conclusion, although these three kinds
of CF were all effective to some extent, the effectiveness of explicit correction was weaker than
that of recasts. Furthermore, prompts were the most effective type of CF, whose explicitness
3

Russell and Spada (2006) found that CF in the classroom studies was slightly more effective
than that in the laboratory studies (d=1.12, d=0.93, respectively). However, the number of
studies was relatively small (n=9, n=6, respectively) compared to other studies (Mackey & Goo,
2007, k=25, k=27; Li, 2010, k=11, k=14, respectively). In addition, their effect sizes were rather
close to each other. Therefore, in the present study, the claim made by Mackey and Goo (2007)
and Li (2010) was adopted.
15

varies greatly depending upon what type of prompts is actually provided (see Lyster & Saito,
2010, p. 278 for the classification of prompts). These two findings are another factor to cast
doubt on the superiority of the explicit types of CF.
Regardless of their explicitness, one common feature of prompts is that they provide
learners with an opportunity for modified output. Therefore, Lyster and Saito (2010) showed that
in the classroom setting it may be important to provide learners with a chance to utter the correct
form. However, this notion was in contradiction to Mackey and Goo (2007), which showed that
the studies without opportunities for modified output (either intentionally or unintentionally) had
a larger effect size than the studies with them. This study included both laboratory studies and
classroom studies. Therefore, it still may be true that modified output is crucial in the classroom
setting; however, above all, the conclusion derived from a meta-analysis of only 15 studies
should be interpreted with great caution. Thus, the necessity of modified output is still
controversial.
As can be seen above, at least regarding CF, even with meta-analyses, researchers have
not reached any consensus yet. In addition, a minority of researchers have cast doubt on the
effectiveness of CF (R. Sheen, 2005), and some even state that L2 teachers should abandon the
use of FonF altogether in their classrooms (e.g., Krashen, 1981; Schwartz, 1993; Truscott, 1996,
1999). Truscott (1999) listed a number of reasons for his claim that especially FonF for
grammatical errors must be abolished, from both teachers’ and students’ perspectives. For the
former, teachers have difficulty in (a) understanding the error, (b) presenting the correction, (c)
being consistent, (d) tailoring the correction to the student, and (e) maintaining a communicative
focus. For the latter, students have difficulty in (a) noticing and recognizing the correction, (b)
taking the correction seriously, (c) understanding and accepting the correction, and (d)

16

incorporating the correction. Unfortunately, most of his arguments were anecdotal. He even cited
some of the previous studies that showed the ineffectiveness of CF and also criticized some
studies that showed the effectiveness of CF. Nevertheless, Truscott’s argument seems not
supported by most of the research studies published after his paper.
In conclusion, regarding the benefits of CF, previous studies show conflicting results,
and researchers claim different opinions. Then, what are L2 teachers supposed to take out of
this? This is one of the main topics of the present study.

2.2 The Stated Beliefs and Classroom Practices of L2 Teachers
2.2.1 Teachers’ Stated Beliefs
Beliefs of teachers have been a focus of many studies in education; however, despite
its long history, there does not seem to exist a consensus regarding how beliefs can be defined,
causing difficulty in studying them (M. Borg, 2001; S. Borg, 2003a; Pajares, 1992; Woods,
1996). One reason may be that beliefs cannot easily be amenable to empirical investigation,
which often leads beliefs to be regarded as a proper topic for other disciplines, such as
philosophy and religion (Pajares, 1992). Therefore, educational researchers have no choice but to
resort to teachers’ stated beliefs expressed in such ways as interviews and questionnaires. For
this reason, rather than beliefs, the present study revolves around stated beliefs, which are
defined as “statements [L2] teachers made about their ideas, thoughts, and knowledge that are
expressed as evaluations of what ‘should be done’, ‘should be the case’, and ‘is preferable’
(Basturkmen et al., 2004, p. 244). It is important to acknowledge that beliefs and stated beliefs
may not always match because teachers may hold some beliefs unconsciously or they do not
want to express some of their beliefs, which are not congruent to external norm, such as from

17

society, their current teaching paradigm, and even the interviewers, whose norms interviewees
tend to follow when they express their beliefs (Kagan, 1990; Pajares, 1992; Woods, 1996).
Even teachers without any previous experience of teaching are neither “blank/ clean
slates” (Angelova, 2005, p. 31; Feiman-Nemser & Melnick, 1992, p. 4; Kumaravadivelu, 2001,
p. 552), “tabula rasa” (Akbari & Dadvand, 2011, p. 1; Freeman, 1992, p. 4; Rankin & Becker,
2006, p. 355), nor “empty vessels” (Freeman & Johnson, 1998, p. 401). Among all the factors,
previous studies showed that prior language learning experience has a great impact on not only
inexperienced teachers’ stated beliefs of teaching but also their actual classroom practices either
positively or negatively (for an early review of education in general, see Kagan, 1992a; for
reviews of L2 teachers, see S. Borg, 2009b, and Legutke & Ditfurth, 2009) because, by the time
they become teachers, they have been in many classrooms as students, watching a variety of
teachers, a process which Lortie (1975) called the “apprenticeship of observation” (p. 61).
Previous studies have investigated the relationship between teachers’ stated beliefs
about teaching and their learning background among teachers of multiple subjects or subjects
other than L2 teaching (Calderhead & Robson, 1991; Goodman, 1988; Holt-Reynolds, 1992;
Hsiao-Ching, 2000; Nespor, 1987; Powell, 1994), of ESL (Bailey et al., 1996; M. Borg, 2005;
Eisenstein Ebsworth, & Schweers, 1997; Farrell, 1999; Golombek, 1998; Johnson, 1994; Mok,
1994; Numrich, 1996), of EFL (S. Borg, 1999a; Duff & Uchida, 1997), and of a Japanese as a
foreign language classroom (Sato & Kleinsasser, 1999). In addition to the review of studies by
Freeman (1992) and Kagan (1992b), all of these studies showed that the teachers’ stated beliefs
about teaching were strongly influenced by their past learning experience, and in the case of L2
teachers, by their formal L2 and often L1 learning experiences (also see S. Borg, 2003b; Busch,
2010). Especially related to the present study, Numrich (1996) analyzed a series of teaching

18

reflective diaries written by 26 novice ESL teachers in the US enrolled in a master’s program in
TESOL while taking a 10-week practicum. Among these teachers, seven of them refrained from
correcting students’ errors mostly owing to their own negative experience of being corrected in
the classroom when they were learning an L2 (e.g., being humiliated). M. Kennedy (1990) states
that “[t]eachers acquire seemingly indelible imprints of teaching from their own experiences as
students and these imprints are tremendously difficult to shake” (p. 8). In addition to these past
learning experiences, Duff and Uchida (1997) listed two more components that affected
teachers’ stated beliefs: past teaching experiences and cross-cultural experiences, the former of
which was also shown to influence teachers’ stated beliefs in Nespor (1987). Therefore, the truth
is likely that a variety of teachers’ past experiences plays a role in forming their stated beliefs.

2.2.2 The Relationship Between Stated Beliefs and Classroom Practices
Previous research shows that beliefs have a strong effect on behaviors (for a review of
education in general, see Fang, 1996; Kagan, 1990; Pajares, 1992; for mathematics education,
see Brown & Cooney, 1982); likewise, in language teaching, there is a robust relationship
between practices and cognition: beliefs, assumptions, knowledge (BAK; Woods, 1996), and
thoughts of L2 teachers (Bailey, 1996; S. Borg, 1999a, 2009b; Pajares, 1992; J. Richards &
Lockhart, 1994; Woods, 1996). In his extensive literature review of beliefs, Pajares (1992)
believes in the power of beliefs and claims that “[f]ew would argue that the beliefs teachers hold
influence their perceptions and judgments, which, in turn, affect their behavior in the classroom”
(p. 307). Previous studies show that teachers’ stated beliefs and classroom practices agree with
each other, some of which will be discussed in the following paragraphs.
In Hsiao-Ching (2000), stated beliefs and the classroom practices of a biology teacher

19

were congruent in two points. First, the teacher believed that it was important for her to ask
many questions in her class in order to help students become independent science thinkers and
also to interest them in the content. Following this belief, she, in fact, asked numerous questions
in her class. Second, she believed that boys and girls had different learning styles and class
participation styles, stating that boys were generally better than girls in multiple facets in science
class. This belief affected her behavior in that she interacted with boys much more frequently
than girls.
Focusing on L2 teachers, Bailey (1996) used classroom observations and interviews
with stimulated recalls (Gass & Mackey, 2000) in order to investigate how six experienced
teachers in an Intensive ESL program in the US decided to deviate from their lesson plans. The
results showed that the six decision-making principles that teachers were using were heavily
influenced by their own beliefs about teaching, which were (a) serve the common good, (b) teach
to the moment, (c) further the lesson, (d) accommodate students’ learning styles, (e) promote
students’ involvement, and (f) distribute the wealth.
Johnson (1992a) investigated the stated beliefs of 30 ESL teachers, using A
Multidimensional TESL Theoretical Orientation Profile, which consisted of the Ideal
Instructional Protocol (Kinzer, 1988, as cited in Johnson, 1992a), the lesson plan analysis task
(Kinzer, 1988, as cited in Johnson, 1992a), and a Beliefs Inventory (Leu & Kinzer, 1991, as cited
in Johnson, 1992a). The study showed that “the majority of ESL teachers (60%) possess clearly
defined theoretical beliefs which consistently reflect one particular methodological approach
toward second language teaching” (p. 93). In a subsequent study, then, among these participants,
she selected three experienced teachers (4-12 years of teaching) at the secondary level (Grades
9-12), and analyzed the relationship between their stated beliefs and classroom practices. These

20

three teachers were chosen based on the results of her first study, so that each of them had a
different theoretical orientation of L2 teaching: skill-based, rule-based, and function-based. She
observed their reading and writing classes eight times, and analyzed the data using instructional
units defined as the instructional events that are featured by “one set of instructional activities
used to accomplish an instructional objective with a single set of materials” (Martin-Rehrmann
& Leu, 1987, p. 10, as cited in Johnson, 1992a, p. 96). The results showed that over 50% of their
instructional units were congruous to their theoretical orientations (skill-based: 54.5%,
rule-based: 69.1%, function-based: 93.5%). This finding was further corroborated by the
comments made by the three teachers at the post-observation interview.
Brumfit, Mitchell, and Hooper (1996; Mitchell, Brumfit, & Hooper, 1994a, 1994b)
investigated the stated beliefs and classroom practices about the role of knowledge about
language (KAL), which was defined as having five categories: (a) Language as System, (b)
Language Learning and Development, (c) Styles and Genres of Language, (d), Social and
Regional Variation, and (e) Language Change through Time. The teachers were teaching either
contemporary English or foreign languages (French, German, and Spanish) in Grade 7 or 9 in
five secondary schools in the UK. The data were collected through classroom observation lasting
eight weeks, and interviews. The results showed that there were many discrepancies between the
English teachers and foreign language teachers. The foreign language teachers regarded explicit
presentation of KAL important because of its “direct contribution … to the development of
pupils’ target language proficiency” (Brumfit et al., 1996, p. 77). To support this notion, the
teachers mainly focused on the first category of KAL: Language as System, and tended to teach
KAL explicitly in the classroom, dealing with sentence-based grammar and vocabulary work. In
contrast, the English teachers believed that such an approach was only marginally helpful for the

21

development of learners’ linguistic ability, and they were “concerned with features of whole
texts, and rarely paid attention to the specific detail of sentence structure” (p. 79). Accordingly,
they seldom taught KAL explicitly in the classroom; instead, they tended to focus on three of the
other four areas of KAL: Language Learning and Development, Styles and Genres of Language,
and Social and Regional Variation. Despite such differences, what was common between these
two groups of teachers was that, as shown above, their stated beliefs and classroom practices on
KAL were congruent to each other.
However, the relationship between stated beliefs and classroom practices can also
differ in each teacher (Almarza, 1996; Kumaravadivelu, 1993; J. Richards, Ho, & Giblin, 1996),
and also in case of student teachers, it may be superficial in the sense that teachers may behave
merely in accordance with what has been taught in the program and “may shed once they do not
feel they have to conform to certain standards” (Almarza, 1996, p. 72). Moreover, some studies
show that teachers’ stated beliefs and practices do not always match (Basturkmen et al., 2004; M.
Borg, 2005; S. Borg, 1999b, 2009b; Karavas-Doukas, 1996; Phipps & S. Borg, 2007), which
leads Pajares (1992) to claim that stated beliefs can be unreliable in representing reality.
Johnson (1996b) recounted the practicum experience that a MA TESOL student had by
weekly classroom observations, coupled with interviews before and after each class, stimulated
recalls and written journals. In the first half of the practicum, the teacher was forced to teach in a
way that she did not wish to follow, facing “the gap between her vision and the reality of
teaching” (p. 34) due to the time constraints and the curriculum of the program for which she
worked. She expressed her dilemma:
I don’t like it when I see myself teaching this way. I want it to be more student centered not teacher - centered, but sometimes it’s just easier to stand up there and tell

22

them what they need to know. This is not my vision of good teaching but sometimes I
find myself doing it anyway. (p. 37)

In M. Borg (2005), the ESL teacher showed two discrepancies between her stated
belief and her classroom practices. One is that, although she expressed strong objection against
the didactic approach, her lecturing was so lengthy that her mentor even criticized her for that
point. Related to the present study, the other difference was observed in her use of CF in the
sense that, despite the fact that she stated that the teacher’s CF needed to be careful and sensitive
in order to avoid giving the students a feeling of embarrassment, she simply said “no” to her
students’ errors.
Kumaravadivelu (1993) analyzed whether or not two teachers of English who believed
in CLT were following the principals of CLT through classroom observation using five
macrostrategies. The results showed that whereas one teacher was a successful implementer of
CLT, the other one was not. Nunan (1987) conducted a similar study of CLT with five English
teachers, all of whom showed a somewhat traditional teaching style. These contrasting results
may be ascribed to the fact that each teacher had a different degree of understanding CLT.
Such a phenomenon was also described in Karavas-Doukas (1996) who developed a
Likert scale questionnaire consisting of 24 items (12 favorable ones and 12 unfavorable ones) to
investigate the stated beliefs toward the communicative approach in teaching English. After
establishing its reliability, 101 Greek secondary school teachers who were teaching English took
the questionnaire, and the classes of 14 of those teachers were observed in order to see whether
their classroom practices were congruent to their stated beliefs. The results showed that, although
all of the teachers except one favored the communicative approach, their classroom practices

23

often deviated from it, and exhibited an eclectic or traditional teaching method. The author stated
that this discrepancy was mainly derived from teachers’ lack of understanding of the principles
of the communicative approach, which led them to answer many of the items inconsistently (i.e.,
they answered to two questions related to the same concept differently).
Basturkmen et al. (2004) used class observation, interviews, cued response scenarios,
and stimulated recalls in order to analyze the stated beliefs and classroom practices of incidental
FonF (i.e., CF) of three ESL teachers who were male native speakers of English. The results
showed that their stated beliefs and classroom practices were not congruous to each other to a
large extent. What was interesting in this study was that one of the teachers had only one year of
teaching experience whereas the other two teachers had 11-15 years of teaching experience. It
should be noted that the inconsistencies between the stated beliefs and the classroom practices
were seen more in the inexperienced teacher than in the experienced teachers. The researchers
claimed that this might be ascribed to the proceduralization of technical knowledge, and
speculated that such discrepancies between stated beliefs and the classroom practices may
disappear as the teacher becomes more experienced. This study shows the necessity of taking
teaching experience into account in the study of the relationship between stated beliefs and
classroom practices, or at least, the danger of generalizing results across different teacher
populations.
Relevant to the present study, although two EFL teachers in Turkey described in
Phipps and S. Borg (2007) stated that they believed in explicit CF, in the classrooms one of them
refrained from correcting learners’ errors because of “a lack of confidence, and concerns about
learners’ feelings” (p. 17). Also, the other teacher, although she followed her beliefs, did not feel
entirely confident about her decision because she was concerned about how her students would

24

react.
There are some conceivable reasons for the discrepancy between teachers’ stated
beliefs and practices. One reason is that beliefs and stated beliefs do not always conform to each
other; therefore, even if teachers’ practices are congruent with their beliefs, they may not accord
with their stated beliefs. In addition, S. Borg (2003a) argues that a change in teachers’ cognition
may not directly lead to a change in their behaviors because contextual factors, such as “the
social, psychological and environmental realities of the school and classroom” (p. 94), may
“hinder language teachers’ ability to adopt practices which reflect their beliefs” (p. 94). As
Phipps and S. Borg (2007) mentioned, a change in classroom practices may happen without a
change in stated beliefs, and vice versa. Some of the previous studies on this issue are shown in
the following paragraphs.
Starting with a study on a mathematics teacher, Brown and Cooney (1982) reported the
dilemma of the teacher, who did not put proofs in his tests due to the poor performance of his
students despite his belief that proofs were an essential component in his teaching. Focusing on
L2 teachers, in Lamie (2004), English teachers in a Japanese high school needed to abandon their
desire to adopt CLT for such outside factors as prescribed textbooks and the university entrance
examination that their students needed to take. A teacher in S. Borg (1998b), despite his
reluctance, incorporated explicit work into his teaching for his students’ sake. Similarly, among
the four EFL teachers observed in Duff and Uchida (1997), granting their students’ requests
three of them reluctantly changed their teaching styles to adopt more interactive activities and/or
decrease their control over the students in class. However, their stated beliefs seemed not to
change in any way. Collie Graden (1996) found that, in most cases, the six secondary foreign
language teachers’ stated beliefs of reading and classroom practices in their reading instructions

25

were congruent; still, there were some inconsistencies mainly derived from “the day-to-day
necessity of planning activities for students who did not or could not perform according to the
teachers’ expectations” (p. 390). Tsui (1996) reported an EFL writing teacher who reverted to a
product writing approach from a process writing approach that he preferred for several reasons.
However, the truth is likely that multiple factors are interwoven, so it may be impractical to
single out one specific reason for the discrepancies between teachers’ stated beliefs and
classroom practices (for more possible reasons, see Fang, 1996, and Lamie 2004).
All in all, it is natural to see that some of teachers’ stated beliefs match their practices
whereas others do not. For example, all of the four teachers in Duff and Uchida (1997) showed
some congruence as well as incongruence between their stated beliefs and classroom practices.
For another example, in Burns (1996) which analyzed the stated beliefs and the classroom
practices of six experienced ESL teachers in Australia through classroom observations and
ethnographic interviews including stimulated recalls, the detailed account of one teacher was
described. The teacher believed in the natural communication whose topics were directly related
to the learners’ life in the frame of the learner-centered style of teaching in order to give them
confidence. Her class was indeed communicative in which the students talked about their daily
activities and she believed that the interaction gave confidence to the learners; however, the
interactions, which were mostly lead by the teacher, not between the students, were “more
suggestive and traditional” (p. 167) rather than natural. Thus, some parts of her stated beliefs
matched her classroom practices whereas other parts did not (However, see Chapter 7.1.4 for her
change).
Woods (1996) showed a case of a teacher who, despite her disfavor, repeatedly made
evaluative feedback, such as “good” in response to students’ answers because that behavior had

26

already become automatic in her teaching, which she found difficult to get rid of. However, in
contrast, some of her other beliefs matched her classroom behaviors. Furthermore, Woods
described another teacher in detail whose stated beliefs matched her classroom behaviors. In
conclusion, it is legitimate to say that no teacher can show perfect agreement or disagreement
between stated beliefs and classroom practices.
Finally, there is one more caveat in this issue. In Phipps and S. Borg (2009) which
investigated the tensions between the stated beliefs and classroom practices in grammar teaching
of three experienced EFL teachers in Turkey through 18 months of interviews and classroom
observations, all of the teachers showed some inconsistencies between their stated beliefs and
classroom practices derived from their students’ expectations and preference and for the sake of
classroom management. However, further analyses show that their classroom practices were
actually congruous to their core beliefs, which are defined as the beliefs that have strong
connections or communication with other beliefs (Rokeach, 1968), which had been formed
mainly by their personal experiences. Furthermore, Smith (1996) analyzed the decision making
of nine experienced teachers in adult ESL classes by classroom observations, stimulated recalls,
and an interview. She particularly focused on whether each teacher was concerned with the
product or the process of the language that their students used. The results showed that those
teachers who prioritized accuracy (i.e., product) designed a curriculum and class emphasizing
language code. On the other hand, those teachers who prioritized fluency (i.e., process)
developed curricula emphasizing student interaction and meaningful communication. However,
these two concepts are quite broad and as the researcher admitted, the teachers’ theoretical
beliefs were eclectic rather than dichotomous. This means that if their stated beliefs were
compared with their classroom practices more narrowly, many discrepancies would be expected

27

to emerge. Therefore, these studies suggest that whether stated beliefs and classroom practices
match may depend upon which facets of stated beliefs are used for comparison.

2.2.3 Can Teachers Change?
Whether or not they match, each teacher has their own stated beliefs and classroom
practices. The next question would be whether these stated beliefs and classroom practices can
be changed through some kind of experience, a possibility on which some researchers have cast
doubt (Brown & Cooney, 1982; Buchmann, 1987, 1991; Buchmann & Schwille, 1983; Kagan,
1992a; Lasley, 1980; Lortie, 1975; Pajares, 1992; Richardson, 1996; Tatto, 1998) even when
change is “logical or necessary” (Pajares, 1992, p. 317). One reason is that “attempts to influence
teacher’s behaviour will have an impact only in areas where the input is valued and salient to the
individual, and where it is congruent with, and interpretable within, the teacher’s own world of
thought and action” (Pennington, 1996, p. 340). Berliner (1987) also claims that “a teacher may
at any time be quite justified in choosing to modify or ignore certain findings” (p. 30).
Pennington (1996) further discusses the difference between input and intake, and only accessible
input, which was defined as “those types of information which teachers are prepared to attend to”
(p. 340) can become intake, which is likely to facilitate a change in teachers. Another reason is
that teachers feel comfortable with their beliefs, forming their “self,” so changing their beliefs
means changing themselves, which they tend to resist (Pajares, 1992). Pajares (1992) further
argues that the beliefs that teachers acquire newly are most vulnerable, but “[w]ith time and use,
they become robust, and individuals hold on to beliefs based on incorrect or incomplete
knowledge even after scientifically correct explanations are presented to them” (p. 317), a point
also mentioned in Nisbett and Ross (1980). Furthermore, “[t]o change means to chance the

28

possibility that students might learn less well than they do under current practices” (Guskey,
1986, p. 9), which may lead teachers to feel reluctant to change their classroom practices.
Regarding the factors that potentially effect change in teachers, while proposing
different ideas, previous studies agree on the notion that rather than a single factor, multiple
factors are interwoven. For example, by adducing the examples of four teachers who either
changed their use of pairwork in their classrooms or not, D. Kennedy (1999) claims that three
factors are likely to influence whether or not teachers are willing to change their classroom
practices: attitudes, subjective norms (influence of others), and perceived behavioural control
(constraints/abilities).
Focusing on teachers’ experience, S. Borg (1999a, 1999c) listed three points that have
large influence on teacher cognition: (a) schooling, especially language education, (b) teacher
education, and (c) classroom experience. Examining eight L2 experienced teachers through
classroom observations and interviews including stimulated recalls, Woods (1996) also found
that their beliefs are influenced by these three factors. As discussed above, (a) is usually dealt
with in the frame of how previous L2 learning experience has an effect on their beliefs and
classroom practices. Regarding (b) and (c), previous studies investigated how teachers actually
change through (b) coursework, practicum, and workshop in a program in college, and (c)
teaching experience, typically in their first year of teaching.
The three factors that Van Fleet (1979) claimed can influence teachers’ beliefs were, in
fact, identical to what S. Borg (1999a, 1999c) proposed, but Van Fleet used different names and
definitions: enculturation (i.e., experience as students), education (i.e., teaching experience), and
schooling (i.e., teacher education). Citing this list, Brown and Cooney (1982) claim that, because
enculturation and education both have a large impact on teachers’ beliefs, there is only scant

29

room where schooling can play a role as schooling is the only factor that does not take place in
the actual classroom whose environment is similar to where they are most likely to teach (e.g., a
university class where a professor lectures on educational theories is not likely to be a context
where most of the students will teach in their career).
Similarly, Rokeach (1968) explained the difficulty in changing beliefs from the
perspective of “connectedness: the more a given belief is functionally connected or in
communication with other beliefs, the more implications and consequences it has for other
beliefs and, therefore, the more central belief” (p. 5), thus, the less amenable to change. One
factor that can make a belief “connected” is that it is learned directly from experience, called
underived beliefs. As discussed above, owing to the apprenticeship of observation (Lortie, 1975),
many beliefs that teachers have come from their direct experiences as students when they were
still young. This may make their beliefs more “connected,” which also may be one reason why
teachers’ beliefs often resist changing. At the same time, however, this suggests that beliefs other
than the underived beliefs which teachers acquire through sources other than direct experience,
are more amenable to change. Thus, if a study focuses on those “unconnected” beliefs, finding
cases where beliefs change may not be difficult.
While claiming that a single belief may be resistant to change because each belief is
“part of an interwoven network” (Woods, 1996, p. 293), Woods showed that neither the stated
beliefs nor the classroom practices of L2 teachers was stable, but dynamic, although the degree
of the dynamics may differ in each teacher and the change is gradual rather than sudden. For
example, in a study conducted by Mattheoudakis (2007), the change of stated beliefs of 66
undergraduate students who were taking a 3-year pre-service teacher education program to
become EFL teachers in Greece was tracked, using the Beliefs About Language Learning

30

Inventory (BALLI) from pre- to post-course. The BALLI was developed by Horwitz (1988) and
has thirty-four belief items in five categories: “1) difficulty of language learning; 2) foreign
language aptitude; 3) the nature of language learning; 4) learning and communication strategies;
and 5) motivations and expectations” (p. 284). The results indicate that among the 15 items that
were discussed in the article, most items showed a change; however, it was gradual rather than
sudden, yielding a statistically significant difference between the first and fourth years, but not
between each adjacent years. Mattheoudakis (2007) states that “belief development is a
cumulative process and that changes are more likely to occur gradually as a result of
accumulated and integrated knowledge” (p. 1282).
From the teachers’ point of view, in Bailey (1992) which asked 61 L2 teachers in the
US and Japan about a positive change that they made in their own teaching, all of the teachers
identified some change, although the extent of change significantly differed among them. For
example, 12 teachers stated that they changed their class from teacher-centered to
student-centered whereas one teacher stated that she/he increased the frequency of reading aloud
by students. Bailey further commented that the changes observed were “slow, gradual,
incomplete, partial, ongoing, evolutionary” (p. 276), thus concluding that “teachers do not
always implement the researchers' desired changes within the time frame of a formal
experiment” (p. 276).
Assuming that stated beliefs and classroom practices of teachers can change, the next
question would be how they would change because “[t]eachers, like other learners, interpret new
content through their existing understandings and modify and reinterpret new ideas on the basis
of what they already know or believe” (M. Kennedy, 1991, p. 2; also see M. Kennedy, 1997).
Kagan (1992a) also states that the teachers “filter it through their own personal belief systems,

31

translating and absorbing it into their unique pedagogies” (p. 75). C. Kennedy (1999) enlisted
three strategies to change the behaviors of teachers: “[1] rational approaches
(information-giving); [2] coercive strategies (forcing someone to change); and [3] cognitive
strategies (enabling individuals to reflect on and evaluate what they are doing and engaging with
attitudes and beliefs)” (p. vii), and points out that, in order for teachers to successfully change
practices, their cognitions also need to change, implying that only the cognitive strategies have
potential to succeed. Then, the relationship between the change of beliefs and that of practices
should also be investigated. For example, is it possible to change their beliefs without changing
their classroom practices, and the latter without the former? Which change must occur first?
Referring to staff development programs, Guskey (1986) claims that a change in classroom
practices must precede a change in beliefs with a change in student learning outcomes being
interceding between them. A change in beliefs, though, may or may not bring about a change in
classroom practices. Also, a change in classroom practices may happen without a change in
stated beliefs (Phipps & S. Borg, 2007). Therefore, this issue seems to be more or less the
problem of the chicken or the egg, and its effect is bidirectional (Richardson, 1996).

2.2.3.1 The effect of teacher education. Crandall (2000) suggests the possibility of
changing L2 teachers’ beliefs and classroom practices through teacher education if “awareness of
that prior learning is developed in the teacher education program and opportunities for practical
experiences and conscious reflection upon those experiences are provided throughout the
program” (p. 35). However, previous studies show contradicting results as to whether or not
teachers change through teacher education (Kettle & Sellars, 1996), such as coursework,
practica, and workshops in a program in university. For example, each of the four pre-service

32

mathematics teachers that were described in Brown and Cooney (1982) showed a different
degree both quantitatively and qualitatively as to how the methods course influenced their
practices.
Similar results have been reported with language teachers, too. In the BALLI (Horwitz,
1988) that Mattheoudakis (2007) conducted, three items related to the role of the teacher were
discussed in the article. While two of them, “the role of the teacher is to help students learn how
to learn” and “the role of the teacher is to share knowledge” remained stable throughout the
study, the students gradually came to disagree with the statement “the role of the teacher is to
control the student,” from year 1 to year 4. In addition, there was a statistically significant
difference between those students who took the teaching practicum and those students who did
not. Ironically, it was the former who showed more support for the item, “the role of the teacher
is to control the student.” At the same time, though, Mattheoudakis concluded that the practicum
had only little influence on their stated beliefs as most of the items remained stable.
As another example of language teachers, S. Borg (2011) conducted questionnaires
and six interviews with English teachers who were taking an eight-week Delta (Diploma in
English language teaching to adults) program in the UK. The results show that among the six
participants, the stated beliefs of some of them changed whereas others did not change.
Finally, in M. Borg (2005), after the ESL teacher’s mentor advised her to change her
behavior of simply saying “no” to her students’ wrong answers, the teacher was successful in
making an “overnight transformation” (p. 13). In contrast, some of her other stated beliefs, such
as her antipathy toward the didactic teaching style, remained consistent throughout the program.
That is partly because knowledge of teaching will not simply be transmitted to,
followed, or realized by prospective teachers for their classroom practices; rather it will be

33

“processed and filtered through layers of experience and belief” (Rankin & Becker, 2006, p. 366)
as discussed in Johnson and Golombek (2002). A longitudinal study conducted by Freeman
(1993, 1996b) showed the impact of an in-service masters’ degree program on the cognitions of
four foreign language teachers; however, its influence on their classroom practices was
inconclusive because “[s]pecific aspects of the four teachers’ practice did change; however,
some patterns or routines also endured” (p. 495). S. Borg (2011) speculated that the reason for
such contradicting results can be ascribed to how the words, impact and change are
operationalized. In his study, even those teachers who claimed that their stated beliefs did not
change admitted that the program that they were taking stimulated their thoughts (e.g., they
became able to articulate their stated beliefs). However, their stated beliefs were basically intact.
In such a case, whether the program had an impact on their stated beliefs and whether there was
a change in this teacher depends on the interpretation of these terms of the researcher.
Brown and McGannon (1998, as cited in Peacock, 2001) investigated the stated beliefs
of 35 trainee teachers pursuing a Graduate Diploma in Education, before and after a three-week
teaching practicum using a survey that consisted of 12 statements taken from Lightbown and
Spada (1993). The second survey showed a few changes from the first survey, especially in CF.
Murray (2003) conducted three interviews with 10 English language teacher trainees
who were taking the CELTA (the Certificate in English Language Teaching to Adults) to
examine the development of their language awareness. Overall, every teacher showed changes to
some extent, among which, the ones relevant to the present study were they came to (a)
distinguish between more and less serious errors, (b) believe that errors were less important as
long as communication was successful, and (c) be cautious of identifying learners’ errors being
aware of language varieties.

34

In Pennington and Urmston (1998) described above, it was found that the third year
students were more realistic, pragmatic, diffident, and even cynical than their first year
counterparts. However, Pennington and Urmston ascribed the reason for this discrepancy to the
practicum which all of the students had to take in their third year and concluded that the program
“may not have provided satisfactory preparation for a career as a secondary English teacher in
the local community” (p. 34) as the change that they observed was in the opposite direction of
the intention of the program. In a follow-up study, Urmston (2003) conducted the same survey
with the first year students in the original study when they were about to graduate. Comparing
these two results statistically, the data likewise suggest that those students came to realize the
classroom realities. In addition, their responses to most of the items of the crucial aspects of
teaching largely remained stable, which further bolsters the hypothesis above by Pennington and
Urmston (1998) that the program did not have a large impact on their stated beliefs.
Kettle and Sellars (1996) described how two student teachers changed their practical
theories, including beliefs, in a year, while taking a subject called Professional Development in
their third year in a four-year bachelor program of primary education in Australia. The data were
collected through interviewing, card sorting techniques, and reflective journal writing. The
results show that both the coursework and the practicum had a large influence on their stated
beliefs about teaching.
Angelova (2005) incorporated 15-20 minutes mini-lessons teaching Bulgarian, a
language that the students were not familiar with, in the SLA class that she was teaching and
analyzed their effectiveness in promoting knowledge about language and language learning. The
students were 30 MA TESOL students in the US, and pre- and post-survey, reflective journals,
and classroom discussions were used for data collection. The qualitative results showed that the

35

mini-lessons had a large positive impact on the students’ knowledge about language and
language teaching, having influenced the students both cognitively and affectively. However,
although surveys were conducted twice at the beginning and end of the course, the author did not
present the quantitative results comparing these two surveys. Therefore, it is not clear what
aspects of their beliefs changed to what extent.
Akbari and Davand (2011) investigated the impact of obtaining a MA degree in light
of pedagogical thought units derived from the comments that the teachers made during
stimulated recall sessions conducted following a 75-minute classroom observation. The
participants were eight EFL teachers in Tehran, four of whom had a MA TESOL whereas other
four had a BA in English language and literature. The length of teaching experience was
controlled to be 2-3 years for both groups. The results showed that the MA group produced
almost twice as many pedagogical thought units as the BA group. In addition, related to the
present study, a chi-square analysis showed that the former also made significantly more
comments about Beliefs (teachers’ beliefs and knowledge about language, language learning, and
language teaching) than the latter. The researchers ascribed these differences to the MA group’s
“increased theoretical and methodological knowledge gained from the diverse graduate courses
they had undertaken” (p. 12). However, it should be noted that, although this may be one reason,
this may not be the only reason for such discrepancies.
Busch (2010) used the BALLI (Horwitz, 1988) to analyze how 381 pre-service
teachers who were taking an SLA class changed their stated beliefs. The class mainly dealt with
theory and research without teaching methodology, but the pre-service teachers were tutoring
ESL students concurrently. Among the findings, those relevant to the present study (i.e., CF)
showed that teachers’ answers to the item “If beginning students are permitted to make oral

36

errors in English, it will be difficult for them to speak correctly later on” showed a statistically
significant change from “somewhat disagree” to “disagree.” In addition, 36 teachers changed
from “agree” to “disagree,” some of whom commented that making errors was one of the natural
processes in SLA. However, it should be noted that this was a concept covered in class, the
researcher was also the teacher of the class, and their recounting was a part of their grade.
Therefore, it is possible that these teachers were adjusting their opinions to be geared toward
what they thought the researcher expected to hear.
A similar concern can be observed in MacDonald, Badger, and White (2001), who
investigated the change in the stated beliefs of 55 students pursuing a degree in TESOL (28 for
BA and 27 for MSc) in Scotland. The questionnaire from Lightbown and Spada (1993) was
conducted twice, at the beginning and end of the semester during which they took an SLA class.
The results showed that some of their stated beliefs changed, especially in such a way that they
abandoned the concept of behaviorism, which was one of the topics covered in the class. In
addition, the authors were the teachers of the class and the questionnaires were conducted in the
class. Thus, although not every topic covered in the class changed the students’ stated beliefs in
the way presented in the class, one cannot help but question the study’s validity, as the students
may simply have been responding to the questionnaire in accordance with the researcher’
expectation, a concern also raised by Kagan (1990).
In a questionnaire study conducted by Okazaki (1996) that used the BALLI (Horwitz,
1988) at the beginning and the end of a Japanese teacher education course, the stated beliefs of
the 25 undergraduate students remained largely stable except the items regarding CLT, which
was the focus of the course, many of which showed a statistically significant difference toward
favoring CLT. However, judging from the fact that this questionnaire was conducted and

37

collected within the class time, one can not deny the possibility that the students’ answers were
skewed toward favoring the objective of the course.
A similar tendency was also observed in Scott and Rodgers (1995), which used an
attitude survey to investigate how the stated beliefs of seven teachers of Spanish and French in
high school in the US changed through taking a nine-week writing workshop. The same survey
was conducted at the beginning and the end of the workshop. The results show that for all of the
20 items, the number of desired answers, defined as responses conforming to the principles
taught in the workshop, increased except one item, which consistently received the highest points
in both surveys. However, it should be remembered that this workshop was held by these two
researchers, and the survey was conducted during the workshop. In this case, the teachers were
all experienced ones. Therefore, one can expect less of the halo effect (Mackey & Gass, 2005)
than those studies above for pre-service teachers because the relationship between the
researchers and participants was only temporary in this study. Still, the possibility that the
teachers were responding in the way following the researchers’ intension is undeniable owing to
the hierarchy existing between these two groups, one in charge of the workshop and the other
one taking the workshop.
Almarza (1996) followed four foreign language student teachers enrolled in a
pre-service teacher education program for nine months in London, using interviews, journals,
classroom observations, and stimulated recalls. She analyzed the relationship between their
pre-training knowledge, knowledge covered in the program, and their classroom practices. The
results showed that although all four teachers followed the method taught in the program rather
homogeneously in their teaching, each of them responded to the contents covered in the program
quite differently. This is evidence for the possibility that, even if a change is observed in their

38

classroom practices, it may only be superficial. In fact, one participant seemed to revert to her
original idea once she finished her teaching practice. As discussed above in the studies of the
change in beliefs, the student teachers may simply have been tuning their teaching to fit what
they thought their observers wanted to see. This shows the importance of obtaining the data from
multiple perspectives over a long span of time when investigating the change of teachers.
The fact that each student teacher responds differently to a common program was also
reported in J. Richards et al. (1996). The authors reported on five trainee teachers taking a TESL
teacher training program focusing on CLT in Hong Kong called the UCLES/RSA Certificate in
TEFLA. The teachers did a teaching practicum, peer observation, discussion with course tutors,
group conferencing, and self-reports, and the data were collected from audio-recorded
discussions with course tutors and self-reports from the teachers. The authors identified five
points that emerged from the comments made by these student teachers: “[(a)] the trainee’s view
of their role in the classroom, [(b)] the discourse the trainees used to describe teaching, [(c)] the
factors trainees considered important in achieving continuity in a lesson, [(d)] the dimensions of
teaching the trainees found problematic, and [(e)] the trainee’s evaluation of their teaching” (p.
245). The results showed that each of the student teachers responded differently to each of these
five points, which also yielded a variety of classroom practices and their evaluations.
Nevertheless, the impact that the program had on the student teachers was strong, yielding a
change in their teaching.
Cabaroglu and Robers (2000) also showed that the response to a common program
differs in each student teacher. They conducted three interviews with 20 student teachers
attending a 36-week PGCE Secondary course in Modern Foreign Languages teaching in order to
analyze the change of their beliefs in teaching. They found that one teacher did not show any

39

change, two reversed their beliefs and the remaining 17 showed some change. However, it
should be noted that they operationalized change as movement or development in beliefs;
therefore, their definition was broader than other studies.
Such varying responses among student teachers to a common program were also
depicted in studies on teachers of subjects other than L2 teaching. A study conducted by
Hollingworth (1989) focused on the change of knowledge and beliefs of 14 elementary and
secondary (of math and science) pre-service teachers taking a 9-month graduate teacher
education program at University of California, Berkeley through interviews, their writing, and
classroom observations. This study found that the differences in their response were largely
affected by their stated beliefs prior to the entry to the program.
Bramald, Hardman, and Leat (1995) analyzed how the beliefs of 162 students
secondary teachers changed through taking a postgraduate certificate in education course in
England, using a questionnaire and interviews. The overall mean scores of the questionnaire
showed that the participants came to favor a more traditional teaching style; however, the
tendency seemed to differ depending upon their majors, some of which showed the opposing
change.
In addition to pre-service teachers, Pennington (1996; Cheung & Pennington, 1994)
shows the difference in reaction to a program could also be observed among experienced
teachers. In the project in which the author was also involved, eight incumbent secondary-level
English teachers conducted a collaborative action research about process writing as a thesis work
for a MA TESL degree that they were pursuing. She analyzed their responses to a questionnaire
conducted before and after the six-month training. Among all the findings, what is relevant to the
present study was that the two teachers who primarily had negative attitudes toward the program

40

tended to view research as the most valuable source more than other teachers who had positive
attitudes, possibly because their focus shifted away from classroom practices owing to their
unfavorable feelings toward the program.
In another study conducted with experienced teachers, all four EFL teachers described
in S. Borg (1999a) had an RSA Certificate in Teaching English as a Foreign Language to Adults
(CTEFLA), and the training they received in this program largely affected their stated beliefs and
classroom practices, basically towards using grammatical terminology as little as possible owing
to the emphasis on CLT of the program. Among the four teachers, one teacher is noteworthy. In
the initial training that he received, he was taught never to use grammatical terminology, which
had a large impact on his early practices, implying that he did not use grammatical terminology
in his class. However, pursuing a diploma five years after that, he underwent professional
training in which he learned that students have different learning styles. Then, he realized that
some students may actually appreciate the use of grammatical terminology, and together with his
own L2 learning experience of grammar translation methods, he began to incorporate traditional
teaching methods with grammatical terminology. It should be noted that this training did not
directly promote the use of terminology. This shows that the facet on which teacher training
impacts is often unpredictable.
As another study with experienced teachers, Lamie (2004) tracked the change of stated
beliefs and classroom practices of four English teachers in Japanese high school who were taking
the Japanese Secondary Teachers’ Programme at the University of Birmingham through
questionnaires, semi-structured interviews, and classroom observations. The purpose of the
program was to shift the teaching style from the traditional grammar translation to the
communicative teaching approach. The results of the questionnaires showed that following the

41

principle of the program, all of four teachers came to teach more communicatively and believe in
the communicative teaching approach more than before the program (however, such a change
did not always lead to a change in their teaching practices as written above).
Lastly, focusing on CF, in Numrich (1996), among the seven ESL teachers who were
hesitant to correct students’ errors at first owing to their negative experience of being corrected
or lack of confidence in their own knowledge about English, some of them came to realize that
their students were actually seeking CF as some of them explicitly asked for it. While some of
those student teachers still felt unsure about whether they should correct students’ errors, others
started to do so in the middle of their practicum. In addition, the researcher, who was also the
teacher of the practicum class, designed an observation sheet for the student teachers to focus
only on CF given by other teachers while observing their class. This surely also raised the
student teachers’ consciousness of their use of CF as well as their students’ expectation toward
CF, which must have facilitated the change of their classroom practices and potentially, their
beliefs about CF.
In contrast to these studies above, previous studies also show that teachers do not
change through coursework, a practicum, or a workshop, as Kagan (1992b) claims in her
influential review of literature (also see Lortie, 1975, for the same view; but see Dunkin, 1996,
for a critique of Kagan, 1992b). M. Kennedy (1990) showed a mathematical proof of it:
By the time we receive our bachelor’s degree, we have observed teachers and
participated in their work for up to 3060 days. In contrast, teacher preparation
programs usually require something in the neighborhood of 75 days of classroom
experience. What could possibly happen during 75 days to significantly alter the
practices learned during the preceding 3060 days? (p. 4)

42

Peacock (2001) described how the stated beliefs held by 146 trainee teachers who were
pursuing a BA TESL degree in Hong Kong did not change over three years, using the BALLI
(Horwitz, 1988). The results showed that the majority of items did not show any change
throughout the three years of the program; moreover, trainee teachers’ answers were different
from those of experienced ESL teachers surveyed in Peacock (1999). Unfortunately, some of
these beliefs seemed to have negatively affected their English proficiency. Precisely speaking,
those trainee teachers who believed that “learning a foreign language is mostly a matter of
learning a lot of new vocabulary/grammar rules” were significantly less proficient than those
trainee teachers who did not believe such statements. This finding was in line with Peacock
(1999) in which the BALLI (Horwitz, 1988), proficiency tests, and interviews were conducted
with 202 EFL students and 45 EFL teachers in Hong Kong. It should be noted, however, as
Peacock stated, this is the question of the chicken or the egg: that is to say, it is not clear whether
the beliefs affected proficiency or the other way around.
Lo’s (2005) case study followed one Taiwanese EFL teacher in elementary school in
depth. After teaching for three years, she was enrolled in a Master’s language teacher education
program in the US. She took the SLA class in her final semester, and an interview was conducted
after each class, which the researcher observed. Because the class was mostly theoretically
oriented, she could not find any relevance to teaching in the class; she even showed some
resistance to the content covered and homework assigned in the course. Naturally, the class did
not have an impact on her practices, and when she went back to Taiwan she resumed teaching in
the same way as she used before she had left Taiwan.
Aside from L2 teacher education, another case study with a student teacher was

43

conducted in Sendan and Roberts (1998). They investigated the change of the personal theories
regarding the effectiveness in teaching of the student teacher who was taking a four-year degree
program at a Turkish university through interviews using repertory grid data about his thoughts.
The interviews were conducted three times over the period of 15 months: (a) in the third year, (b)
in the fourth year before the teaching practicum, and (c) in the fourth year after the teaching
practicum. The data were analyzed from two perspectives: (a) content, defined as “the semantic
distinctions made by the participating student teachers when classifying and discriminating
teachers known to them, according to their pedagogic effectiveness” (p. 231), and (b) structure,
defines as “the ways in which individual constructs are hierarchically organised into a whole
system of construction” (p. 231). The data show that whereas the former remained relatively
stable except that three new constructs were added, the latter demonstrated a clear development
although its process was never linear, but “complex, evolutionary and perhaps cyclical in nature”
(p. 238). The researchers warned that the teacher education program would be easily dismissed
as ineffective wrongly if only the former dimension (i.e. content) was in focus.
Weinstein (1990) described the stated beliefs about teaching held by prospective
elementary teachers who were taking an introductory course of teacher education for a semester.
The course had two components: (a) classes on campus, and (b) field experience, including
observing class, interviewing teachers, and helping them. Questionnaires were conducted with 38
students: one at the beginning of the semester, and the other one at the end of the semester. The
researcher chose 12 students among those 38 and conducted follow-up interviews in the
following semester. The results show that, despite their “unrealistic optimism” (p. 285), their
stated beliefs were largely intact.
More broadly, a large-scale survey research was conducted by Brousseau, Book, and

44

Byers (1988). They used a revised version of the Educational Beliefs Inventory (Freeman & The
Undergraduate Program Evaluation Committee, 1982, cited in Brousseau et al., 1988), which
consisted of 60 items of five-point Likert scale categorized into five groups: student, pedagogy,
milieu, teacher, and curriculum. The participants were divided into four groups: (a) Entry, who
were currently in the undergraduate teacher education program at Michigan State University
(n=391), (b) Exit, who were finishing the same program (n=332), (c) Experienced, who were
working as full-time teachers in elementary, middle, and senior high schools (n=382), and (d)
Alumni, who were teacher education graduates from Michigan State University and also
currently full-time teachers (n=90). One of the findings showed that there were only four out of
the 60 items that were significantly different between Experienced and Alumni although they
gradated from different programs, implying that “teacher preparation loses its impact as a major
factor influencing the formation and maintenance of this set of beliefs” (p. 36).
Lamb (1995) investigated the impact of a short (two-week) in-service program at an
Indonesian university. Among the 16 teachers who participated in the program, the author
interviewed 12 and observed the class of four a year later. Although the student evaluation of the
course on the final day turned out to be very positive, what the researcher found was that “all the
participants had forgotten most of the information and ideas that they had previously been
exposed to” (p. 75). Furthermore, most of what they still remembered was either misunderstood
or rejected after a trial; only some simple teaching techniques were still in use. Thus, it seems
that the program facilitated only little change in teachers’ classroom practices while their beliefs
may have been negatively changed owing to the adverse effect of the program.
Returning to L2 teachers, other research that incorporated a workshop was conducted
by Mackey, Polio, and McDonough (2004), which consisted of two studies. In their first study,

45

the authors analyzed how the teaching experience affects the use of CF by ESL teachers. The
results showed that experienced teachers tended to use more CF than inexperienced teachers. In
order to investigate the reason for this difference, in their second study, the authors examined
how a professional development workshop affected inexperienced teachers in terms of use of CF.
In the workshop, four inexperienced teachers in an MA TESOL program were provided with
information on CF through (a) discussing FonF, (b) watching examples of FonF on video, and
(c) role-playing the provision of FonF. In addition, they reflected on their use of CF in two of
their classes, one of which was anterior and the other was posterior to the workshop, through
transcripts. The quantitative results showed that the workshop statistically had no effect on their
use of FonF although they used far more recasts after the workshop than before it. In contrast,
the qualitative results showed that the workshop raised the teachers’ awareness of the
opportunities to use FonF; however, at the same time, they expressed the difficulty of
incorporating the technique into their teaching. One of the limitations of this study is that the
workshop was only to introduce FonF, and did not have a component for the teachers to reflect
on their own classroom practices (although they may have done it on their own). This may be
why the workshop virtually had no effect on their classroom practices.
Finally, instead of focusing on teachers, the change in teachers’ practice as perceived
by their students was investigated by Kubanyiova (2006), who conducted a 5-point Likert scale
questionnaire with the students of eight EFL teachers, all of whom were non-native speakers of
English, in Slovakia. The teachers were taking a 20-hour in-service teacher development course
although one of the teachers was a pre-service teacher. The questionnaire was conducted twice,
before and after the course, with around 170 students (the numbers slightly differed between the
two questionnaires). The results showed that out of 56 t-tests, only 10 showed a significant

46

change, nine of which actually became worse. This led the author to conclude that “the [teacher
development] course failed in its goal to promote significant change in teaching practice that
would make a difference for the students” (p. 7). Through conducting interviews and classroom
observations with the teachers, the researcher ascribed the reason for this failure to three factors:
(a) the teachers’ lack of motivation to teach English (they were rather interested in English itself
or increasing their self-esteem), (b) the lack of reflective teaching culture in the research context,
and (c) an unsupportive school system.
It should be noted that in some of the studies discussed above, student teachers were
also teaching concurrently; still, these studies generally show that taking a course, a practicum,
or a workshop may hardly change students’ beliefs or classroom practices. Clark (1988) claims
that teachers’ implicit theories, which were defined as “eclectic aggregations of cause-effect
propositions from many sources, rules of thumb, generalizations drawn from personal
experience, beliefs, values, biases, and prejudices” (p. 6), will not “be quickly and thoroughly
replaced by the usual lecture, reading, discussion, practice, and evaluation methods typically
employed in teacher preparation programs” (p. 7). Weinstein (1999) claims that “[r]eviewing
video- or audio-tapes of their own teaching and engaging in stimulated recall, discussion, and
analysis may promote an awareness of the cognitive demands of teaching, as well as students’
own implicit premises” (p. 287). Such a methodology may be necessary to promote some
change. Furthermore, in his extensive review of literatures, S. Borg (2003a) argues that while the
content of teachers’ cognitions may not change, the processes and the structure of cognitive
development may change. This implies that whether a change can be observed may depend on
the focus of the study.

47

2.2.3.2 The effect of teaching experience. Once they finish a program and get a job
teaching, facing the difficulties of teaching in an institution and the real classroom, novice
teachers are likely to need to change their beliefs as well as classroom practices in order to adapt
themselves to the environment. Therefore, in contrast to the contradicting results from the
previous studies on the impact of teacher education, the previous studies on the impact of real
teaching experience seem to reach a close consensus, which is that teachers all change in some
way.
One reason may be that facing the realities of teaching and having more experience,
in-service teachers are more eager and readier to adapt and adopt new theories or ideas of
teaching than pre-service teachers even when both of these groups are exposed to the same
materials. For example, in Bullough and Baughman (1993), how a seventh grade teacher
(Baughman) had changed in the past five years was examined through interviews, classroom
observations, and think aloud sessions using videotaped classes. It was found that the main
reason for the radical change in her teaching was that she was inspired by a book introducing
new ideas of teaching that she met in her fifth year of career. Bullough expressed his doubt that
the book would have had the same impact if it had been read earlier for the reasons stated above.
This implies that even if research studies do not have a large influence on inexperienced
teachers, there is still a possibility that the story may be different with experienced teachers.
Regarding L2 teachers, after the third survey was conducted, Peacock (2001; described
above) invented an instruction package in which the trainees saw the results of the research, read
some articles, watched successful ESL lessons, and discussed communicative teaching. Although
he did not provide any evidence, he stated that these were “successful.” However, once teachers
start teaching in school, at least in Hong Kong, they may have had to revert to their original

48

beliefs as shown in Pennington and J. Richards (1997) and J. Richards and Pennington (1998),
which described how and why the five novice teachers were forced to change their classroom
practices in their first year of teaching in a way contrary to what they learned and believed in L2
teacher training. The participants were all Hong Kong Chinese, recent graduates from a BA
course in Teaching English as a Second Language at City University of Hong Kong, and teachers
of English in Hong Kong secondary schools. The data were collected through 11 classroom
observations, followed by individual interviews, monthly meeting with the researchers, and
self-tape-recording of their lessons. The results show that “much of the philosophy prompted in
the BA course was … stifled” (p. 164) for a variety of reasons. For example, although they all
expressed their belief in CLT at the beginning of the school year, sooner or later they abandoned
it for a traditional knowledge transmission mode. Also, the medium of instruction changed from
exclusively English to a mixture of English and Cantonese. Urmston and Pennington (2008) also
followed up on the graduates from the program (a different cohort from the one above) by
conducting an interview with five teachers a year after graduation and two classroom
observations and an interview a year after that with three of them. The results were identical to
those above in that teachers had to “retreat from what they considered ‘best practice’” (p. 97)
mainly due to such external factors as students’ low proficiencies in English and the
product-oriented aims of the lessons. These studies show the difficulty of applying the principles
that novice teachers learn in school to their actual classroom practices.
If the effect of teaching experience on beliefs and classroom practices is powerful, it is
natural to assume that experienced teachers and inexperienced teachers will show different
patterns (Basturkmen et al., 2004; Carter & Doyle, 1989, J. Richards, Li, & Tang, 1998).
Brousseau et al. (1988; described above) found that 48 out of 60 items yielded significant

49

differences between experienced and inexperienced teachers, showing that the former did “a)
favor a common curriculum, b) report believing that students should be given more
responsibility, c) believe that there should be common standards for all students, d) agree that
schools should act as change agents, and e) reduce their sense of efficacy” (p. 33), more than the
latter.
A study that compared experienced and inexperienced L2 teachers was conducted by
Mok (1994) described above. Among the 12 ESL teachers who participated in the study, six
were categorized as experienced and other six as inexperienced with three years of ESL teaching
experience being the cutting line. The study found that although the practicum that they took did
not entail a major change of stated beliefs in either of these two groups of teachers, more change
was observed with the inexperienced teachers than the experienced teachers in the sense that
“[the former] started to take into consideration needs analysis, social context, students’
characteristics, their limitations and input” (p. 106). If teaching experience is such a powerful
influence on stated beliefs, as this study verified empirically, it is legitimate to assume that the
impact of teaching experience is the strongest when teachers start teaching for the first time in
their life. It is unfortunate, though, that this study did not include a class observation component
to see if there were any change in classroom practices while comparing experienced and
inexperienced teachers.
However, as shown in Mackey et al. (2004), thus far studies of language teacher
cognition have paid much more attention to pre-service teachers than in-service teachers, which
S. Borg (2009b) deplored. Many studies reported the agony that pre-service or novice teachers
experienced in their teaching. For example, using dialogue journals, Brinton and Holten (1989)
and Holten and Brinton (1995) showed how ESL student teachers struggled and also how their

50

themes in their journals changed through their practicum. Johnson (1992b) investigated the
reasons for such a struggle and change by analyzing how six ESL pre-service teachers made
instructional actions and decisions in their classroom. The teachers were enrolled in a practicum
as part of a 2-year master’s degree program in TESOL, and had less than two years of teaching
experience. The data collection was conducted over the course of two semesters, using
videotaped observations of classroom, stimulated recalls, and retrospectives about the teachers’
decisions after teaching and stimulated recalls. Among the findings, it was found that their main
concern was to maintain the flow of their class while heavily relying on a small number of
instructional routines, and at the same time, unexpected student behavior had the strongest
influence of their subsequent actions. In her conclusion, Johnson ascribed the difficulty that
inexperienced teachers have to the fact that they “have not developed a schema for interpreting
and coping with what goes on during instruction, nor do they possess a repertoire of instructional
routines upon which they can rely” (p. 509).
Similar results were also reported by Nunan (1992), who investigated how ESL
teachers made decisions during lessons. The nine participants were divided into two groups: five
experienced teachers, defined as those who had more than two years of teaching experience and
four inexperienced teachers, defined as those who had less than two years of teaching
experience. The teachers prepared and taught a lesson that lasted between 60-90 minutes, which
was observed, and then followed by interviews and reflections on their teaching while reading
the transcription of the lesson. Nunan analyzed the comments made by the teachers, quantified
their contents, and compared the results between experienced teachers and inexperienced
teachers. It was found that while inexperienced teachers were mainly concerned with classroom
management, experienced teachers were mainly concerned with language issues. Also,

51

inexperienced teachers’ comments tended to be negative (e.g., T explanation too complicated –
should have demonstrated) while experienced teachers’ comments tended to be positive (e.g., T
personalized explanation rather than being abstract).
The reason why experienced teachers can focus more on language issues than
classroom processes may be explained by J. Richards (1998), who showed that experienced
teachers (9.6 years of teaching experience on average) were able to improvise their teaching
more than inexperienced teachers (1.6 years of teaching experience on average). Accordingly, S.
Borg (2003a) argues that experienced “teachers learn to automatise the routines associated with
managing the class, and can thus focus more attention on issues of content” (p. 95). However, it
should be noted that although such a dichotomy between experienced teachers and inexperienced
teachers (i.e., novice teachers or student teachers) can be seen in pervious studies (e.g., Borko &
Niles, 1982; Nunan, 1992), each teacher has a different degree of teaching experience not only
quantitatively (e.g., the number of years of teaching) but also qualitatively (e.g., the teaching
contexts). Surely, no teacher suddenly becomes experienced from inexperienced; they become
experienced gradually. Thus, the criterion used to divide between experienced and inexperienced
is arbitrary. For example, in Nunan (1992), the benchmark was two years. It is controversial
whether a teacher with two years of teaching experience can be called experienced. In addition,
in Nunan’s study, the years of teaching experience in the group of experienced teachers varied
from 3 to 15. It is quite possible that there is also a difference in their beliefs and classroom
practices between these two teachers owing to such a different amount of teaching experience.
Therefore, although it may be necessary to arbitrarily divide the teachers into two groups for the
sake of research, the reality is that the difference between experienced and inexperienced is not
dichotomous, but continuous. Furthermore, although teaching experience seems to be playing an

52

important role in changing teachers’ beliefs and classroom practices, it should be reminded that it
is still one of the various factors that influence their beliefs and classroom practices.
Thus far, only a handful of studies directly compared CF provided by experienced
teachers and inexperienced teachers. Pica and Long (1986) showed that “ESL classroom
conversation, in lessons for elementary-level students, at least, is different from informal,
noninstructional NS-NNS conversation in several ways relevant to its potential as an acquisition
environment” (p. 89). Then they compared eight experienced teachers and six inexperienced
teachers enrolled in a master’s program in ESL. Among many findings, it was found that both
groups of teachers used comprehension check much more than confirmation check and
clarification request; however, none of these showed a significant difference between these two
groups.
A similar study was conducted in the first study of Mackey et al. (2004), in which they
compared CF provided by nine experienced teachers and nine inexperienced teachers. The
experienced ones had from 4.5 to 15 years of experience in teaching EFL or ESL, and the
inexperienced ones were undergraduate students taking an introduction to TESOL methods class.
The results showed that both groups of teachers used the approximately equal amount of
negotiation; however, the experienced teachers used all other FonF techniques (e.g., preemptive,
recasts, and explicit CF) much more than the inexperienced teachers, showing a significant
difference in every case. Tarone and Allwright (2005) state that the need between novice
teachers and in-service teachers may differ in the sense that the latter prefers “more emphasis on
the declarative knowledge base or on understanding than on the acquisition of skills” (p. 15).
However, this topic still awaits further investigation due to the small number of studies
conducted thus far.

53

2.2.3.3 The effect of research studies. When students are taking coursework, they
usually read some kinds of research studies in the form of assignments (Zeuli, 1994); however,
those studies investigating the impact of coursework on stated beliefs and classroom practices of
teachers usually do not single out research studies as an independent variable. From the teachers’
point of view, among the 34 teachers of English in Turkey surveyed in S. Borg (2007a) who read
research studies either often or sometimes, 29.4% of them answered research studies had a slight
influence, 50% answered a moderate influence, and 20.6% answered a fairly strong influence. So
although the number is small (7 teachers for “a fairly strong influence”), research studies
definitely have a role to play in influencing teachers. Nevertheless, thus far, only a few studies
have actually investigated how research studies by themselves facilitate a change in teachers’
beliefs and classroom practices.
In Farrell (1999) introduced above, two of the pre-service teachers who experienced
and believed in a deductive approach to grammar decided to try out an inductive approach in
their teaching owing to the studies that they read in their class. One of them actually commented
that she was, “greatly influenced by my readings” (p. 9).
Rankin and Becker (2006) described an action research project conducted with a native
speaker of German teaching German 101 with no prior experience of teaching by using
classroom observations, the teacher’s (Becker) detailed recounting of what happened in class,
interviews, and written impressions and observations by several of his students. The teacher read
three articles on CF in a graduate-level pedagogy seminar: Long and Robinson (1998), Lyster
(1998b), and Roberts (1995), reflected on his way of teaching, and made a clear decision on his
teaching methods. Accordingly, his way of providing CF became consistent while using prompts

54

for self-correction and explicit correction tailored to student needs. However, when he actually
faced his students, he had to abandon his plan to some extent because of the diversity in their
needs. Further, he commented: “The literature does not seem to me to provide any guidance with
this dilemma” (p. 364). Still, the authors stated that the seminar meeting “prompted a deeper
level of analysis regarding feedback behaviors” (p. 365). Thus, as Kagan (1992a) claims, reading
research by itself may not change teachers’ beliefs. It may need to be corroborated by some other
intervention (in this case, discussion) in order to facilitate change in beliefs. S. Borg (1999a) also
negates the unidirectional nature (i.e., research informs practice), claiming that there is a
“reciprocal relationship where research is grounded in the realities of classroom practice but at
the same time provides teachers with insights into teaching through which they can critically
examine, and hence improve, their own practice” (p. 122). One of the factors that affected the
direction of the teacher’s change was the choice of the articles. As Rankin admitted, the
researcher’s “own bias in choosing the article [i.e., Lyster (1998b)] and the emphasis it received
in the seminar meeting” (p. 365), greatly affected the direction of his change.
Another study that showed the effect of a single study can be seen in Vásquez and
Harvey (2010). Kumaravadivelu (2001) proposed the idea that teachers should conduct teacher
research in the postmethod period of L2 teaching. Following this notion, in Vásquez and Harvey,
nine graduate students who were enrolled in Master’s level SLA course partially replicated
Lyster and Ranta (1997). In Vásquez and Harvey, following Lyster and Ranta, the effectiveness
of CF was measured by uptake, a students’ reformulation directly following CF. The study
showed that most of the results of their replications were in line with that of the original study,
Lyster and Ranta (1997), in the sense that recasts were the most common type of CF and they did
not lead to students’ uptake. The shift in the students’ awareness of CF as a result of conducting

55

this research was recorded by a variety of methods: questionnaires, videotaping a lesson,
reflection of the videotaped lesson, a research report, a reflective essay, reflective journals,
interviews, and research memos. The results showed that the affective dimension (i.e., the
face-threatening nature of CF) that many participants mentioned prior to the study disappeared,
and their focus shifted to “a more nuanced and sophisticated understanding of [CF’s] role and
function, and its interaction with student uptake” (Vásquez & Harvey, 2010, p. 430). However, it
should be noted that uptake has been criticized as its reliability has been in question in some of
later studies (Ellis, Basturkmen, & Loewen, 2001b; Loewen & Philp, 2006; Long, 2007; Mackey
& Philp, 1998) as it is possible that learners merely repeat the recasts from teachers (Gass, 2003).
It should be noted, though, that both of these studies focused on inexperienced
teachers. There is indirect evidence that the impact of research studies may be more limited for
experienced teachers, contrary to what Bullough and Baughman (1993) implied. Among the 47
teachers in Shkedi (1998), only one teacher stated that she would prioritize the research findings
over her own thoughts. About three-fifths of the remaining would reject the entire findings and
the feelings of two-fifths were rather mixed. The reasons for the former were (a) “[t]he research
does not correspond with reality”, (b) “[p]ersonal-professional experience is more meaningful”,
(c) “[d]oubts about the authenticity of the researchers”, and (d) “[a]voiding difficulty” (p. 568).
Bartels (2003) asked three L2 researchers and three L2 teachers with more than 20
years of teaching experience to read two research articles about written corrective feedback, one
of which was research-oriented and the other of which was teaching-oriented, and later the
researcher interviewed each of these six participants to elicit their reactions to each of these two
articles. It was found that the expectations toward research studies that teachers and researchers
had were incongruous to each other in the sense that, while teachers sought experiential

56

knowledge, researchers sought empirical evidence. The author concluded that teachers and
researchers had “different ways of validating ideas in journal articles and had different ways of
using and incorporating information in the articles into their professional knowledge” (p. 748).
Having established an identity as a teacher without recourse to research for extensive time,
experienced teachers may simply not find a suitable place for research to be incorporated into
their teaching repertoire any more.
Not only the experience of teaching, but also to what extent the teacher is familiar with
research may play a major role in the effect of reading research studies. In Zeuli (1994), two
groups of teachers of various subjects and levels participated in an interview study. One group of
five teachers had an experience of working with educational researchers on research projects for
at least a year. The other group of eight teachers had less experience of research. They were to
read three research studies of different kinds, and their responses were analyzed. Whereas the
former group of teachers believed that “research should have an indirect impact: that is, the value
of research is to help raise questions about their teaching and offer analytic frameworks with
which they can better understand their work” (p. 41), the latter group of teachers believed that
they should have direct impact on their teaching with pedagogical implications and were most
concerned about the products of the studies. For all the teachers in the former group, the research
studies that they read “triggered their intuitive beliefs about what constitutes good teaching” (p.
51). In contrast, in the conclusion section, referring to the latter group of teachers, Zeuli stated:
They had greater difficulty identifying authors’ main ideas and evidence in support of
these ideas…relied more on personal interpretations of the articles as opposed to
defensible interpretations based more firmly on the text…were less likely to draw on
different conceptions of learning or educational aims when responding to research

57

findings and when offering reasons for ignoring good research. (p. 53)

Zeuli called them “consumers interested in making decisions about what goods to
procure without understanding further why the decision is warranted” (p. 53).

2.3 Conclusion
Among the studies that described how teachers change, only a few have investigated
how using research studies helped teachers reflect on their own teaching, although “SLA
research has an important role to play in helping teachers evaluate their beliefs and knowledge
about SLA and to decide to change these” (Tarone & Allwright, 2005, p. 20). Due to the paucity
of studies, this should be a topic of interest in the field. Especially, to my knowledge, there has
not been a single study that investigated the effect of reading research studies about CF on
experienced ESL teachers. Therefore, in the present study, the answers to the following three
research questions will be sought.
1

What are the teaching philosophy, stated beliefs and classroom practices of CF of the four
ESL teachers?
1.1

What is the relationship between their teaching philosophy, stated beliefs on CF, and
classroom practices of CF?

2

What is the relationship between the four ESL teachers’ stated beliefs on CF and how they
read three research studies of CF?
2.1

How do their stated beliefs on CF prior to reading affect how they read research
studies?

3

What effects does the reading have on the stated beliefs and classroom practices of CF of the

58

four ESL teachers?

The present study has important implications for four stakeholders: researchers,
pre-service teacher educators, in-service teacher educators, and in-service teachers, each of
whom will be discussed briefly in the following paragraphs.
The present study will enable researchers to understand the influence of their research
studies on teachers’ pedagogical practices. While some researchers claim that L2 research does
not need to have a direct relationship with teaching, others call for such interrelatedness.
However, even the L2 researchers who argue for research studies that are applicable to
classroom teaching do not always know exactly what kind of effect their work has on teachers’
instruction. Thus, the current study is meaningful in that it will promote a better understanding of
what effects researchers can expect their studies to have on L2 pedagogy.
Also, the results of the present study will provide useful information for pre-service
teacher educators. Reading and discussing research studies is one of the most common
components in many pre-service teacher education programs. However, teacher educators are yet
to know what kind of impact these research studies have on teachers’ classroom teaching and
cognition. Therefore, the present study can inform pre-service teacher educators about how they
can effectively utilize research studies to improve the quality of teacher education.
Finally, it is possible that this study will inform in-service teacher educators as well as
in-service teachers to revalue the importance and the possibility of utilizing research studies for
professional development. L2 teachers do not normally read research studies once they finish
their pre-service training. Through the current study, in-service teacher educators and in-service
teachers may realize that, although research studies seldom seem to present substantive hands-on

59

guidance or practical solutions to the problems that they face in their everyday classrooms,
research studies may still be a valuable resource due to their indirect influence on L2 teachers
(Ellis, 1997).
In conclusion, the present study sheds light on an area that has been largely ignored in
the field thus far: the influence of research studies on teachers’ stated beliefs and classroom
practices of CF in L2 teaching. As a result, it is my sincere hope that the present study will help
to bridge the gap between research and teaching.

60

Chapter 3: Method

3.1 Setting
The present study was conducted at an intensive English program (henceforth; the
program) at a large Midwestern university in the US. The program offers various programs, but
its major function is to teach English to provisionally admitted students to the university to
which the program is attached. The newly-enrolled students are divided into five different levels,
where the level five is the highest and the level one is the lowest, based on the in-house
placement test consisting of listening, reading, and writing. However, the actual English
proficiency of the students greatly varies and even more so skill-wise even within the same level.
Except the highest level (i.e., level five)4, each level has different skills taught by discrete
teachers, but regardless of the level that they belong to, speaking and listening, reading, and
writing are the common skills that they learn. The students take four 50-minute classes every day
four days a week.
There were approximately 1,000 students at the time of data collection, roughly 70%
of whom came from China; some of the other countries of origin were the Republic of Korea and
Arabic countries such as the Kingdom of Saudi Arabic, Iraq, United Arab Emirates, and Qatar
(program staffer, personal communication, January 26 and March 12, 2012). The teaching
faculty consists of full-time teachers, semester hired teachers, and teaching assistants (TAs) from
the university’s MA TESOL program and applied linguistics PhD program.

4

The level five classes were considered as academic class, rather than language class, and
operated under a different system. Because it was not germane to the present study, the
description of those classes had not been included.
61

3.2 Participants
In the present study, the participants were four teachers, all of whom were teaching a
speaking and listening class. In Rankin and Becker (2006), from which the idea of the present
study was originated, there was only one participant. I thought that it would be more
enlightening to describe and compare several teachers because by doing so, some common
themes may be identified. Thus, it was decided to recruit four teachers.
Originally, it was intended to recruit only new TAs in order to secure that no
participants had read any of the three research studies that they were going to read for this study.
During the summer before the Fall semester of 2011 when data collection took place, I asked the
director of the program to send a mass e-mail to all of the students who had a contract to be new
TAs in the fall, asking them to send me an e-mail if they were interested in participating in the
study. Only three of them replied, and one of them turned out to become a research assistant.
Between the remaining two, after a few exchange of e-mails one declined to participate in the
present study. Thus, only one new TA consented to participate.
Having only one participant, I decided to recruit three more teachers who were already
working for the program. Because I was also working at the program as a TA, I was acquainted
with many teachers. The following criteria were used to choose the participants from among the
pool of all the teachers. The first criterion was that the teachers needed to be full-time staff.
Semester-hired teachers were avoided because it was thought to be difficult to set up a meeting
with them as they may not be likely to stay on campus all day or even come to campus every
day, and also they did not have any private office on campus. More experienced TAs were
avoided because it was likely that they had read one or more of those three research studies, as
they were concurrently pursing either an MA TESOL or applied linguistics PhD degree. The

62

second criterion was that the teachers must have worked for the program for at least a year. That
is because if they were new to the program, they may not have been familiar with the system yet,
which may play a unique role in their stated beliefs and classroom practices. The third criterion
was that they all must be teaching a speaking and listening class. This was included in order to
control the variability that a class for different skills may bring. For example, a teacher may not
give CF at all simply because it is a reading class. It was valid to assume that a speaking and
listening class would be ideal because oral CF was most likely to be provided in such classes.
Finally, the fourth criterion was that each of the teachers would have a varying degree of L2
teaching experience. This decision was made posterior to knowing that only one TA would
participate, in order to see the effects of the quantity and quality of teaching experience by
comparing among the experienced teachers. I personally asked six teachers in total by e-mail
first and then in person, and three of them consented to participate in the present study, making
the total number of the participating teachers four.
Before describing each of the four teachers, I should refer to the generalizability of
such a small-scale case study. As Duff (2006) has extensively discussed, qualitative research
focuses on internal validity and reliability rather than generalizability. Moreover, she argues that
generalizability is commonly replaced by transferability in qualitative research, in which readers
have the responsibility “to determine whether there is a congruence, fit, or connection between
one study context, in all its complexity, and their own context, rather than have the original
researchers make that assumption for them” (p. 75). In the present study, focusing on a small
number of teachers allowed me to collect a great amount of data for each of them to be analyzed
in depth. I hope that such a rich description supported by triangulation is likely to give readers
opportunities to “facilitate understanding of one’s own as well as others’ contexts and lives, both

63

through similarities and differences across settings or cases” (p. 75).
Table 2
Brief Descriptions of Each Teacher
No. Name Gender
1
2
3
4

Status

Cecile Female
TA
John
Male Full-time
Jim
Male Full-time
Tom
Male Full-time

Highest degree
MA TESOL (ongoing)
MA TESOL
MA TESOL
MA TESOL

Amount of teaching experience
US
Overseas
Total
0
2
2
11
3
14
12
18
30
10
20
30

As shown in Table 2, each teacher had different teaching experience both
quantitatively (i.e., the number of years) and qualitatively (i.e., different contexts). In order to
secure anonymity, two means were adopted: (a) all of the participants’ names are pseudonyms,
and (b) the levels of the class that they were assigned to are not described in the present study.

3.2.1 Cecile
Cecile had just started to work as a TA at the program and to study as a graduate
student pursuing an MA TESOL when she participated in the present study. Although her status
was a TA, she was assigned to teach specific classes on her own including making the syllabi,
lesson plans, teaching materials, tests, and grading although her supervisor occasionally helped
her by observing her class and having discussions with her. Before working for the program, she
had two years of teaching experience outside the US. The official L2 teacher training that she
had undergone prior to coming to the program consisted of (a) an undergraduate course titled
Methods of Second and Foreign Language Teaching, and (b) a two-week orientation session that
she took overseas before she started teaching English there. Her undergraduate major was
literature, so she was unfamiliar with empirical research studies; however, she had been assigned
to read many in her MA TESOL classes at the time of the present study. She stated that she felt

64

comfortable with teaching speaking and listening because those had been the focal skills in her
teaching overseas. Finally, according to her, she had a 10% hearing loss, which, as will be shown
in Chapter 6, had a possible effect on her participation in the study.

3.2.2 John
John had been teaching English for 14 years in total, including three years outside the
United States. He had taken a training session that lasted a few weeks right before he went
overseas, but he later majored in MA TESOL with a practicum, which was the major portion of
his L2 teaching training. He was exposed to many research studies while in the MA program
which he said that he did not enjoy reading because (a) their results were often contradictory, and
(b) they lacked practicality in teaching a real class. Accordingly, he reported that he seldom read
research studies nowadays except the practical ones related to the use of technology in the
classroom.

3.2.3 Jim
Jim had an MA degree in TESOL, which was the major portion of his L2 teacher
training. He had extensive English teaching experience in various ESL and EFL contexts for 27
years as a full-time and three years as a part-time teacher. He was also a professionally trained
editor, and had written a few books on vocabulary, which was his primary interest. Accordingly,
he reported that he read more than six research studies on vocabulary every year. He even had
favorite researchers whose names he was able to easily list during the first interview. However,
he said that he seldom read research studies on other topics.

65

3.2.4 Tom
Tom had around 30 years of English teaching experience in total in both ESL and EFL
contexts, including eight years in an intensive English program at another university and 20
years outside the United States. He held an MA TESOL degree, which was his only formal L2
teacher training, but had also attended many seminars, especially on brain-based learning.
Accordingly, he regularly read books on that topic, but stated that he had too little time to read
research studies. He confessed that teaching speaking and listening was his least favorite.

3.3 Procedures
3.3.1 Reading the Three Research Studies
Each teacher was to read three research studies: Lyster and Saito (2010), Mackey,
Gass, and McDonough (2000), and Truscott (1999). However, a concern was that the teachers
would be too busy to read the studies. Thus, along with the actual studies, the summary of each
study was also given to them. The benefits of providing teachers with the summaries of research
articles were documented by Erlam (2008): “Firstly, it attempts to bridge the different discourse
domains that are characteristic of the two types of knowledge. Secondly, a summary of the
literature addresses the problem of the sheer volume of research literature available and the little
time that is available for practitioners who often complain of overwhelming workload pressure”
(p. 255). In the present study, a PowerPoint file was used, starting with an example of each type
of CF taken from Y. Sheen (2004, pp. 278-279, with a slight adjustment), followed by the
summary of each of the three research studies (see Appendix A). The former was included in
order to help the teachers understand the research studies better.
The teachers were instructed to go through the PowerPoint file first, and if they still

66

had time they could read the actual studies. The order of reading was not prescribed. In addition,
except for Truscott (1999), they were told that they could start reading from the results section,
skipping the literature review and the methods sections. They were also told that they could
highlight some parts and/or take notes on their reflections if they wished. In such a case, the
articles were collected to be stored as another source of data. The summary of Truscott (1999)
was described in Chapter 2. Below is the brief summary of each of the other two research
studies.
Lyster and Saito (2010) conducted a meta-analysis of CF synthesizing results from 15
published studies. The studies included were restricted to quasi-experimental classroom studies.
There were four main findings: (a) CF was more effective than no CF, and the effectiveness
lasted for a few weeks; (b) the effectiveness of recasts, prompts, and explicit correction were all
significant; (c) prompts were more effective than recasts; and (d) the effectiveness of explicit
correction was not significantly different from that of recasts and prompts.
Mackey et al. (2000) investigated how accurately L2 learners could recognize the types
of CF given by their teachers using a stimulated recall. The participants were ten ESL learners
and seven learners of Italian as a foreign language (IFL). Each learner carried out a
communicative task with a native (English) or near-native (Italian) interviewer. During the task,
the interviewer provided CF, mainly recasts, when the learner made an error in morphosyntax,
phonology, semantics, and lexis. After the tasks, while the learners were watching the video of
their interaction individually, they paused the video when they wanted to describe what their
thoughts had been at that moment. The researchers also paused the video when CF was provided
and asked the learners to express what their thoughts had been at that moment. The two main
results were (a) ESL learners accurately recognized CF generally for lexis and phonology (83%

67

and 60%, respectively), but not for morphosyntax (13%), and (b) IFL learners accurately
recognized CF generally for lexis (66%), but not for phonology or morphosyntax (21% and 24%,
respectively).
These three research studies offer somewhat different findings and unique perspectives
on CF and were selected in order to stimulate the teachers’ thoughts on and responses to the
topic. For example, according to Lyster and Saito (2010), no matter what kind of CF is provided,
CF is effective to varying degrees. Mackey et al. (2000) showed that language learners often
misunderstand CF given by their teachers, especially when the target feature is morphosyntax,
but the students’ perception of CF was relatively accurate when it was lexis. Truscott’s (1999)
arguments have often been considered controversial by other researchers, but the article
nevertheless brings up many points that teachers may find persuasive and actually favor over
strictly empirical evidence. Thus, by being exposed to a range of research studies and
perspectives, the teachers in the present study had the opportunity to consider various aspects of
CF, and be selective about what they would deem most effective and relevant to their teaching
practices and beliefs while reflecting on their “fund of knowledge based on prior experience,
prior reading … prior classroom-based research … current awareness of their local pedagogical
situations” (Belcher, 2007, p. 398).

3.3.2 The Schedule of Data Collection
The main methods for data collection were classroom observations and recordings,
semi-structured interviews, and stimulated recalls. The observations and recordings were used in
order to observe the teachers’ actual classroom practices of CF, whereas the latter two methods
were used in order to understand their stated beliefs on CF. The procedures for data collection

68

are presented in Table 3.
Table 3
The Procedures for Data Collection
No.
1
2
3
4
5
6
7
8

Procedures

Time (minutes)
Inside Outside
class
class

Data to be collected

Observation
(Pre-reading)

1 The participants’ classroom use of CF before
the reading.
Background information on the participants’
L2 learning, L2 teacher training, L2 teaching,
Interview 1
the current teaching context, and their stated
beliefs on CF. (see Appendix B)
Participants
read
three articles on CF
The participants’ responses to the research
Interview 2
studies. (see Appendix C).
Observation
2 The participants’ classroom practices of CF
(Post-reading)
immediately after the reading.
Stimulated Recall 1 The participants’ stated beliefs on CF
(of Observation 2)
immediately after the reading.
Observation 3
The participants’ classroom practices of CF
(Delayed
about a month after the reading.
post-reading)
Stimulated Recall 2 The participants’ stated beliefs on CF about a
(of Observation 3)
month after the reading.
Total

Table 4
The Actual Dates of the Data Collection
Procedures
Cecile
Observation 1
September 20
Interview 1
September 22
Interview 2
October 12
Observation 2
November 3
Stimulated Recall 1
Observation 3
November 22
Stimulated Recall 2

50
60
(3-4.5
weeks)
60
50
60
50
60
150

240

John
September 23
September 26
October 17

Jim
September 19
September 20
October 21

Tom
September 22
September 23
October 19

October 28

October 27

October 25

December 2

November 22

November 29

Table 4 shows the actual dates when each piece of data was collected for each teacher.
Because teachers were often unavailable due to their busy schedules, the period between dates
differed among the teachers. For example, Interview 2 was conducted sometime between 20 and

69

31 days after Interview 1.

3.3.3 Classroom Observations
Three classroom observations, each of which lasted approximately 50 minutes, were
conducted over the course of one semester, as the data obtained from classroom observations is a
crucial component in order to investigate teachers’ cognition (S. Borg, 1999b, 2003a). The first
class was observed in the fourth week of the semester. The first three weeks were purposely
avoided for the following reasons: (a) the students might move to another level or another class
within the same level, (b) the teachers were busy with the start-up administrative work, and (c)
the rapport between the teachers and the students had not been established yet, which might have
skewed teachers’ use of CF. The second class was observed approximately a month after the first
class. The third class was observed approximately a month after the second class.
All the classes were audio-recorded as well as video-recorded. An audio-recorder was
attached to the teacher so that it could capture all the utterances by the teacher and most of what
the students said individually to the teacher. Another audio-recorder was placed on the table to
capture the utterances by the teacher and the students. The video camera was set at the back of
the classroom, operated by me as a non-participant observer. Efforts were made to capture the
teacher and as many students as possible to see their interactions. I remained silent throughout all
the classes except a case when a teacher asked me a question about a Japanese word.
One methodological drawback was that the recorders could not capture all the
utterances that the students made. Because all of the students who participated in the present
study were English learners, their speech was imperfect in many cases. However, being unable to
transcribe their speech hindered me from identifying all of the errors and CF. Even the native

70

speakers of English who proofread the transcripts while listening to the audio files could not
confirm all of the utterances. It would have been helpful to use a better-quality microphone when
the class was in the teacher-fronted style, but even this solution is of limited use when students
talk in pairs or groups. Ideally, each student should wear a microphone, but this is costly and
also, the class would probably become awkward and lose its naturalness (Mori & Zuengler,
2008). A point of compromise would be to set up an audio recorder or a wireless microphone for
each pair or group, but judging from my own experience, the larger the number of people in the
group becomes, the more difficult to identify who said what. A small class size would be helpful,
but this is uncontrollable for naturalistic classroom data. There seems to be no panacea for this
predicament for now, so researchers have no choice but to rely on whatever data they can collect
at their best.

3.3.4 Interviews
Interviews were conducted twice. The first one (Interview 1) was conducted a few days
after Observation 1 in order to collect the participants’ background information, specifically,
their experiences with L2 learning, L2 teaching, L2 teacher training, and the current teaching
context, as well as their stated beliefs on CF.
The second interview (Interview 2) was conducted 3-4.5 weeks after Interview 1, after
the teachers had read the three research studies. The purpose of Interview 2 was to investigate
the teachers’ responses to the readings. The teachers talked about what they had thought and how
they had felt about the studies. I did not state my own view on the topic. Compared to asking the
participants to write comments alone, I felt that conducting interviews with them would enrich
the data both quantitatively and qualitatively. For example, if participants filled out a

71

questionnaire, did a think aloud, or wrote their thoughts and feelings on a sheet of paper, their
remarks might tend to be short or shallow. Therefore, in order to make the data richer, I chose to
elicit answers through a series of questions. This may have made participants talk about what
they did not first intend to disclose. One more important reason to conduct Interview 2 was that
this may have facilitated their deep reflection on their practices of CF on the spot. As discussed
above, Kagan (1992a) claims that simply reading research studies may not have a large influence
on teachers; they need reflection on their teaching. Reading the three research studies alone may
not have provoked their reflection. Through answering a series of questions related to their
primary thoughts and feelings on the research articles, they may have started to reflect on their
stated beliefs and classroom practices of CF more deeply during the interview.
Throughout all of the interview sessions, an audio-recorder was placed on the table to
capture the utterances by the teacher and me. I asked questions and took notes throughout both
Interviews 1 and 2.

3.3.5 Stimulated Recalls
Almarza (1996) claims that “teaching is more than observable behavior” (p. 75). In
order to investigate the cognitive processes of teachers, some kind of introspective methods
where they verbalize their thoughts need to be implemented. There are some kinds of verbal
reporting, such as talk aloud and think aloud. However, neither one of them was applicable in the
present study because it would have interrupted their teaching. Therefore, in the present study, a
retrospective method needed to be adopted. In this sense, the stimulated recall, defined as “one
subset of a range of introspective methods that represent a means of eliciting data about thought
processes involved in carrying out a task or activity” (Gass & Mackey, 2000, p. 1), was used. In

72

the stimulated recalls, the video or the audio of the teachers’ own classroom teaching was used
as stimulus. The video was used when the teachers were standing in front of the classrooms and
led the class as a teacher-fronted style because the teachers were able to see the non-verbal
behaviors that they made in class, which may also have further facilitated their recall. However,
when the students were working in pairs, triads, or groups, the video could not capture the voice
of the teachers clearly. Therefore, in such sections the audio file recorded through the
microphone attached to the teachers was used. Regardless of the media, the use of stimulus was
important not only because this would likely to enhance the teachers’ access to their memory,
which is one advantage over interviews without any prompts, but also because researchers
“cannot deduce language pedagogies on the basis of teachers’ accounts of how they work
without reflecting with them upon actual instances of practice” (Breen, Hird, Milton, Oliver, &
Thwaite, 2001, p. 498). The following instruction was printed out, and while showing it to the
teachers, I read it aloud (adopted and modified from Egi, 2004, p. 274, and Gass & Mackey,
2000, p. 154):
What we are going to do now is watch the video or listen to the audio of today’s class,
especially when your students made errors orally regardless of whether you gave
corrective feedback or not because I am interested in what you were thinking at that
time. So you can tell me to stop the video or the audio anytime when you noticed that
one of your students made an error during the lesson. You can also tell me to stop the
video if you noticed an error now that you did not notice during the lesson. I may also
stop the video when I remember I heard an error during the lesson. I can see what you
were doing by looking at the video or listening to the audio, but I do not know what
you were thinking. So what I would like you to do is tell me what you were thinking,

73

what was in your mind at the time while you were hearing their errors, but not what
you are thinking now. Your response can be about anything and as long as or as short
as you want it to be. If you do not remember what you were thinking at the time, you
can say, “I don’t remember.” If you were not thinking anything particularly, you can
say, “I wasn’t thinking anything.” There is no right or wrong answer, so please feel
free to say anything you like.

One of the issues of the stimulated recall is the accuracy of recall. In order to avoid
memory decay, the stimulated recall should be conducted as soon as possible after the event
(Gass & Mackey, 2000). The later it is conducted, the more likely that “participants may say
what they think the researcher wants them to say or may create a plausible explanation for
themselves” (p. 54). In the present study, following the ideal, each stimulated recall was
conducted on the same day as the classroom observation, in most cases immediately after the
class, and in other cases within a few hours after the class.
Throughout all of the stimulated recall sessions, an audio-recorder was placed on the
table to capture the utterances by the teacher and me. I asked questions, but unlike Interviews 1
and 2, I did not take notes because the questions were largely discrete.

3.4 Analyses
3.4.1 The Classroom Data
For the classroom data, an AS-unit, defined as “a single speaker’s utterance consisting
of an independent clause, or sub-clausal unit, together with any subordinate clause(s) associated
with either” (Foster, Tonkyn, & Wigglesworth, 2000, p. 365) was adopted for the analyses.

74

An error was defined as an utterance by a student that contained a morphosyntactically,
phonologically, or lexically non-standard expression or a wrong choice of a word.5 Other types
of errors, such as pragmatic errors (e.g., the use of an informal expression when a formal one is
expected) or content errors (e.g., an irrelevant answer to a content question), were excluded
(although the four teachers sometimes made comments on those errors as well, some of which
will be used for analyses in Chapter 6). When an AS-unit contained at least one error, it was
coded as an error-filled AS-unit, otherwise, an error-free AS-unit was coded. It should be noted
that, although a single AS-unit may contain more than one error, each unit was coded either
error-filled or error-free. Thus, regardless of the number of errors that a unit had, as long as it
contained at least an error, it was coded as error-filled.
In the present study, CF was treated as the same as reactive FonF, defined as “attention
to form [that] can occur sporadically in response to any errors produced by the student”
(Loewen, 2011a, p. 579). Therefore, CF was identified only in the cases when an error was
identified. It should be noted that teachers often provide CF without necessarily the intention to
correct errors but merely as a natural response to students’ utterances, but their responses still
function as CF, and students may or may not notice or learn from it. In the present study, the
stimulated recalls were conducted for Observations 2 and 3, which partly solved this issue. But
even in these cases, teachers did not make comments on all of the CF that they gave. Besides,
without the stimulated recall, the intention of the teachers when they gave CF at Observation 1
was totally unknown. Therefore, in the present study, regardless of the teachers’ intention to
correct errors, all of the teachers’ utterances following their students’ errors which could possibly
5 The microphones in the classrooms captured most of the voices of the teachers clearly;
however, it turned out that those of the students were often unclear, fragmented, or overlapped,
and it was often unfeasible to detect errors in the students’ speech. Therefore, an error was coded
only when a student obviously made a hearable error.
75

function as CF were coded as CF.
In the present study, the taxonomy of CF used in Y. Sheen (2004) was adopted as
shown below:
1. Input-providing (Reformulations): The teacher gives the correct answer.
(a) Explicit correction: The teacher provides learners with a correct form with a clear
indication of what is being corrected.
(b) Recasts: The teacher reformulates the whole or part of learner’s erroneous
utterance without changing its meaning.
2. Output-prompting (Prompt): The teacher tries to elicit the correct answer from the
learner.
(a) Clarification requests: The teacher signals to learners that their utterances were
either not understood or were ill formed.
(b) Metalinguistic clues: The teacher provides technical information regarding the
learner’s erroneous utterance without explicitly providing the correct answer.
(c) Elicitation: The teacher uses one of the three techniques for eliciting the correct
form from the learners: (a) elicit completion (e.g., it is a …), (b) elicitative
question (e.g., How do you say X in French?), and (c) reformulation request (e.g.,
Can you say it another way?)
(d) Repetition: The teacher repeats learners’ ill-formed utterances without any
change.
3. Multiple feedback: The teacher combines more than one type of CF described above.

The teacher’s first utterance following the error was analyzed using the same system as

76

the one for errors, coding each AS-unit as either a CF-filled or CF-free. When an error-filled
AS-unit contained multiple errors, single (or even multiple) tokens of CF may not have targeted
all of them. In such cases, regardless of how many errors received CF, as long as the teacher’s
utterance contained at least one token of CF, it was coded as CF-filled. All of the error-filled
AS-units and the CF-filled AS-units were analyzed linguistically without considering any
non-verbal behaviors. An error may elicit more than one CF-filled AS-unit. In such a case, each
AS-unit was coded separately.
For each teacher, the frequency as well as the percentage of the total tokens was
calculated for (a) error-filled AS-units that were followed by CF-filled AS-units, (b) error-filled
AS-units that were not followed by CF-filled AS-units, and (c) each type of CF.
A graduate student who was majoring in applied linguistics coded 20% of the data
taken from either from the beginning, the middle, or the ending of each class. The interrater
reliability was 86.2% for the identification of error-filled AS-units, 80.5% for the identification
of CF-filled AS-units, and 97.6% for the types of CF. The graduate student and I had a follow-up
meeting and talked of all of the discrepancies until we came to a complete agreement.

3.4.2 The Interview and Stimulated Recall Data
For the data from the interviews and the stimulated recalls, a content analysis was
conducted, specifically using a line-by-line analysis as described in K. Richards (2003). A
content analysis can be divided into two types: deductive and inductive (Ellis & Barkhuizen,
2005). In the present study, the latter type of content analysis, inductive analysis, was conducted,
defined as “approaches that primarily use detailed readings of raw data to derive concepts,
themes, or a model through interpretations made from the raw data by an evaluator or

77

researcher” (Thomas, 2006, p. 238).
A graduate student who was majoring in applied linguistics coded 20% of the data
taken from either from the beginning, the middle, or the ending of each data, using the list of the
codes. The interrater reliability was 84.2%. The graduate student and I had a follow-up meeting
and talked of all of the discrepancies until we came to a complete agreement.

3.4.3 The Procedures of Analyses
All of the data from the audio files were transcribed by the researcher using the
software called HyperTRANSCRIBE. For the classroom observations, these transcripts were
supplemented by the video files. In addition, native speakers of English proofread all of the
transcripts while listening to the audio flies.
For the classroom observations, the predetermined coding scheme described above was
used in order to identify error-filled AS-unit as well as CF-filled AS-units. Each of the CF-filled
AS-units was further categorized into each type of CF defined by Y. Sheen (2004). Then, the
next two analyses were conducted:
1. A within-case analysis was conducted in order to analyze each teacher’s (non) use of CF for
each Observation (e.g., the amount of CF, types of CF).
2. A cross-case analysis was conducted in order to look for similarities and differences in
(non) use of CF across the four teachers (e.g., the amount of CF, types of CF).

For the interviews and the stimulated recalls, as described above, a content analysis
was conducted using the software called HyperRESEARCH. First, a line-by-line analysis was
conducted and themes were coded for each data set. Then, the next two analyses were conducted:

78

1. A within-case analysis was conducted in order to look for the themes that emerge
repetitively for each data set.
2. A cross-case analysis was conducted in order to look for similarities and differences in their
stated beliefs on CF across the four teachers for each data set.

In addition, combining the data obtained from these two types of analyses, following
analyses were conducted.
1. A within-case analysis was conducted in order to see the relationship between Observation
1 and Interview 1.
2. Using the results obtained from No. 5 above, a cross-case analysis was conducted in order
to look for similarities and differences in their relationship across the four teachers.
3. A within-case analysis was conducted in order to see the relationship between Interview 2
and Observations 2 and 3.
4. Using the results obtained from No. 7 above, a cross-case analysis was conducted in order
to look for similarities and differences in their relationship across the four teachers.

In the following chapters, the results and discussions derived from the within-case
analyses and the cross-case analyses will be presented. For the sake of clarity, each chapter will
be organized into five sections, each of which describes each teacher in depth, followed by the
final section exclusively contributing to the similarities and differences among them derived
from the cross-case analyses.
It should be noted that, regardless of whether it be within-case analyses or cross-case
analyses, comparing simply the amount of CF provided in different classes is difficult because

79

the task characteristics of the class may have an impact on the use of CF (Skehan, 1998). That is
to say, the amount as well as the kinds of CF that a single teacher provides in the classroom may
greatly vary depending upon the nature of the class. In Chapters 4 and 6 where the quantitative
data of the teachers’ use of CF are presented, although some attempts were made to compare or
contrast among the three observations (within-case analyses) or the four teachers (cross-case
analyses), no definite conclusion can be made from these analyses.

80

Chapter 4: The Teachers’ Teaching Philosophy, Stated Beliefs and Classroom Practices of CF

In this chapter, based on the data obtained mainly from Interview 1 and Observation 1,
the answer to the first research question will be discussed:
1

What are the teaching philosophy, stated beliefs and classroom practices of CF of the four
ESL teachers?
1.1

What is the relationship between their teaching philosophy, stated beliefs on CF, and
classroom practices of CF?

Previous studies show that teachers’ stated beliefs and classroom practices are
influenced by their prior L1 and L2 learning experience. In addition, although it is known that
their stated beliefs have an impact on their classroom practices, to what extent they are
congruous to each other depends on each teacher, and moreover, even within a single teacher, it
depends on each facet of teaching. In the present study, it was found that the teachers’ stated
beliefs on CF were derived from multiple sources including not only their L2 learning experience
but also their L2 teaching experience and some others. The relationship between their stated
beliefs and classroom practices of CF was largely congruous to each other.

4.1 Cecile
As a novice teacher who had only just begun her teacher training in the MA TESOL
program, Cecile had not yet developed a clear teaching philosophy, of which she was aware.
Throughout Interview 1, Cecile occasionally expressed her lack of confidence in her teaching
practices not only of CF but also in general. She specifically referred to one of the classes that

81

was required for new TAs that she was taking at the time of data collection.
[The teacher] gave us this handout of classroom situations and how would you deal
with each of these. And when we were going through them, and I felt like, I can,
someone say something I say, oh that's a good idea. Someone else would say
something different, oh, that's a good idea. So then, she would of ask me oh, it seems
like you feel like this way, [the teacher's name], And I say well, actually I don't really
know. Because I feel like, I don't know it, this is maybe part of my personality as well
in general, I'm just kind of very um adaptable, ((laugh)) which can be good, but I feel
like I don't really have a lot of concrete ideas about things, and what, what's the the
best way to make something happen and do something with teaching.6

Also, recalling her memory of learning French in high school, while she tried to
describe the teaching style that her teacher took, she mentioned, “I’m not totally familiar with
different teaching methods and approaches,” showing her lack of knowledge about teaching L2,
which was understandable as she was still studying in the MA program.
Among all the skills, she was particularly worried about teaching grammar. This was
partly derived from the fact that her undergraduate major was literature, not the English
language, linguistics, or education. She did not even minor in TESOL, but took only one course
about L2 teaching. She stated:
I'm teaching listening and speaking again. That's what I was totally focusing on in [a
foreign country], that's kind of what I'm comfortable with, a little bit, so I asked for
that my first semester, and that's what I got. So I'm really happy about that. Because
6

All of the quotes in this chapter were taken from Interview 1 for the text and Observation 1 for
the examples unless otherwise noted.
82

um I feel a little nervous about teaching explicit grammar, and those kind of things and
how to write.

At Interview 2, while we were discussing Truscott (1999), referring to one of the
points that he claimed, “teachers generally are not experts on grammar,“ after expressing her
agreement with him, she stated:
Before I started … this fall, I, I spent the summer like reading a grammar book, and
feeling like oh no. I'm gonna get hammered with hard question, I really have not yet,
explicit grammar questions … obviously in other ways.

Her lack of confidence in teaching was also observed when she was asked to express
her teaching philosophy. After she adduced the anecdote in one of the classes that she was
taking, described above, she still could not state her teaching philosophy and asked me to give
some examples:
T: …but I feel like there are a few things, I, let me think, I, I can't think right now. But,
is definitely things it that um, and what kinds of like principles like what's an
example of something?
R: Oh, like, do, do you always try to be commu, communicative, or do you always do
some activity in your class, or, you know, those kinds of things.

Then, she gave some examples of “kind of like fun stuff” that she tried in her class
including the use of “a fun video” because “those are fun to make,” and also she felt that her
students were ”gonna be more interested obviously if it's interesting subject matter” without

83

stating any teaching rationale behind them. But then, she confessed, “I don't know. I can't think
of concrete teaching, nothing I have, ((laugh)) I mean, I, I, as I'm learning all this stuff right
now.” Later she analyzed herself and stated, “I guess my teaching philosophy is really aren’t
very fluid.” Similarly, at Interview 2, she repeated her view on her teaching philosophy: “I’m
really not concrete about my teaching philosophies or what I think is the most effective … think
I, I should, but I will try to develop that.”
Throughout the present study, she often called herself as a “sponge” as can be seen in
the excerpt from Interview 2, in which she said, “I’m really feel like I’m a sponge right now, and
I’m like what should I be doing, people tell me what to do.” She expressed this concept from the
beginning of the present study. At Interview 1, she stated:
I'm just, I am very flexible person, so I feel like hearing all of these things, I should try
this, I should do this, not that I'm perfect at doing things, and implementing things but I
definitely have a desire to try things out and change things and know that I'm um far
from perfect, ((laugh)) far from adequate even. ((laugh))

In fact, though, she stated a belief about teaching that she had, which was the
importance of speaking L2.
People who sit and just listen listen listen and have all this input, and don't actually
produce the language are never really gonna get it … and examples of people who
grew up … in America, their parents just speaking Spanish, they're not speaking
Spanish, so they can understand it, but they can't speak it. And that's really kind of
shocking to me, So I feel like the more I hear that, it's not just good enough to have
students hear a lot of stuff. And then, or even I remember these students, not just for

84

my classes, none of them my class actually, but they would have these little notebooks
they would write down the most strange vocabulary words. They knew how to say,
whatever, I felt why are you memorizing that, you can't even say what you did this
weekend. Say how are you, and I say oh, I'm fine, bye and run away … that's kind of
something that's really starting to shape the way I think of things too like, if you're not
producing it, it's not worth anything for you just to hear a bunch of things.

Accordingly, she decided to give her students some homework in which they needed to
record their voices and upload the files on the Internet. She stated, “this something that they can
speak and just be confident and, even if it's not correct entirely, obviously it's not gonna be, but,
at least they're, they're producing something in speaking.” However, this stated belief came from
one of the readings that she was recently assigned in one of her classes, which can be another
proof to show that she was a “sponge.” However, this may be true for most of novice teachers
who are currently trained in a program. In this regard, it could be said that most novice teachers
are “sponges” to some extent. Being uncertain about teaching in general, it is no wonder that she
was unsure about CF as well. She stated:
So now teaching here, I’m really kind of not sure how to go about doing [error
correction], I never did in [the foreign country] … I gue, wasn't very easy for me to
understand how to do this. So now here I'm, I've um thought about it a lot but I don't
know that I really change how I’ve done things, or like, consciously realized how I'm
responding to a person's error … oh, let me tell her this, so that um, I don't really
thought that way, so, that's something I'm very, um, not sure about, I guess, honestly, I
feel like, what's the best way to go about doing that … anyways, something I'm not

85

really confident about knowing what I should be doing, that kind of thing.

Also, at the very end of Interview 1, she repeated this notion, saying, “I guess for error
correction sometimes I’m not sure about that.” Therefore, there seemed to be no firm stated
beliefs about CF in her. Naturally, she was lacking the terminology about CF, and stated, “I don’t
even know the words for some of them, like um, some of the different, the names for the
different kinds of error correction” although, in fact, she knew the word recasts.
Such a lack of confidence in her teaching as well as the use of CF came from multiple
sources. The first is that her prior classroom teaching experience was restricted to the two years
in the foreign country, and even in that teaching, she did not have to be more creative than
“simple, play the CD, and fill in the blanks.” Furthermore, she had not had much responsibility
in terms of monitoring her students’ progress with no homework or tests. She recalled:
What they had me do they give me a textbook and they said, don't try to teach anything
grammar or writing or reading. Because we'll teach that in [her students’ L1] and the
students will get it better than if you try to teach grammatical, crazy grammatical
concept in English, and they just stare at you, so like, we just want you, well this is
another thing they told me in my training. They said it’s ok if they don't learn a lot,
quote unquote, learn um, you don't feel like they're learning, as long as they are being
exposed to um, someone who's from America, and American culture, and … just
learning a little bit of that, and hearing new speak, and seeing how you do things, and
as long as they are excited and happy and having fun, then that's ok. That's all good.

She also expressed her frustration about the lack of support from the program for

86

which she had been working. She stated:
I was always kind of felt little frustrated also with [the name of the program
(henceforth: the foreign program)] organization, that they didn't really do a lot to
support teachers once they got to their schools, and again it was just kind of like, figure
it out, here's, teach this sentence, teach this structure, but not really, how to make it
interactive, or how to make it, more, than that. And again, I mean, there's, there's tons
of stuff on the Internet, but, sometimes it's hard to, ((laugh)) search through everything.
So that was kind of what um my frustration. I felt very frustrated, lesson planning. I
felt like I was always doing things at the last minute, and very much not.

With no creativity, responsibility, nor support, she felt, “I didn't do a good job, I would
think back oh, I'm, [her students] are having fun, so that's ok.” However, studying more about L2
teaching in her MA, she began to reject her teaching practices in the foreign country, even as far
as to criticize herself, which she recurrently mentioned, “now that I'm actually studying TESOL,
it's like I'm thinking back to all the things I did, I'm like oh wow. I, really screwed this up, or … I
could've done this better,” “especially, now I feel kind of guilty almost about the way I did things
in [the foreign country], definitely in retrospect.“ She was very honest about this point:
I can't lie and act like I was a very good teacher in [the foreign country], I had a lot of
frustrations, and I didn't feel like I, very adequately taught the students, I was never
really even presented with like the objectives.

At another point during the interview, she stated that she was learning to be a better
teacher:

87

I definitely think back to [the foreign country] and say oh my gosh. I didn’t do any of
that, I didn't do this, I should've done this differently, this was a terrible exercise, this
was stupid, all these things. So, I guess in that sense, in kind of reflecting on and I'm
trying not to do the things that I, um, but I definitely still learning.

Her intent to avoid the activities that she did in the foreign country was also seen in the
following quote:
I guess right now, the stuff I taught in [the foreign country], I'm not really using a lot
of it, just kind of thinking aback on, maybe how I could've improved things and, I don't
know, I mean like there were some songs, there was like at least one song I taught in
[the foreign country], that they actually like memorized a lot of it, and they all, I got
them like all to sing the song, that like, made me so happy, ((laugh)) but like, I thought
of was … using certain songs that I know were successful, but then I think I don't
know if that's really something I should be doing with not having them sing out loud
certainly.

Focusing on CF, her unfamiliarity with CF was partly derived from the fact that, as
described above, she did not need to focus on form, instead, fluency being preponderant over
accuracy, in her teaching in the foreign country, without any kinds of formal assessments. The
major reason, however, lay in her experience of being provided with contradictory information
about many issues in teaching within the orientation of the foreign program or between the
orientation and the school for which she worked. Among those topics, CF was one of them,
which led her to call CF as “the number one thing that was confusing.” She recalled vividly:

88

A lot of people said, never ever correct their errors. That's what I heard. It's not about
having them be perfect, or even right, ah, they just said, just go with it, and so I re, I
can remember specific example they said in [the foreign program] orientation, they
said if someone comes up to you and says, teacher, today birthday. Or something.
Don't even say oh today is your birthday? Just say, oh happy birthday or … I guess
even like recasting or something like that, like they didn't, they said never ever correct
errors. These were the people that were at the orientation. And then I got to the school,
and this is why it was really confusing, they were like, why don't you correct what they
are saying, correct what they are doing, they're writing sentences that are totally wrong
and, and, by the end, that was one thing I was just like, maybe I screwed this all up,
because I never really did correct any errors in [the foreign country] … I, I just felt like
I don't really know how to, and they told me not to, and then my school's telling me to,
and then, what do you want me to do, embarrass the kids in front of everyone.
Table 5
Cecile’s Use of CF at Observation 1
Error-filled AS-units
Kinds of CF-filled AS-units
Without
With
Clarification
Explicit
Total CF-filled CF-filled
Total Recasts
requests
correction
AS-units AS-units
17
8 (47.1)
9 (52.9)
9
8 (88.9)
1 (11.1)
0 (0.0)
Note. The class lasted approximately 50 minutes. The numbers in parentheses show percentages.
Because she deemed production important in L2 learning, it was expected that she
would secure a lot of opportunities for production for students in her class, which would
accordingly be likely to produce a number of errors. As can be seen in Table 5, in Observation 1,
17 error-filled AS-units were identified. Although this number is smaller than those of the other
three teachers, it is clear that she definitely provided her students with chances to speak up in line

89

with her stated beliefs.
Regarding her use of CF, because of (a) her lack of metalinguistic knowledge of
English grammar, (b) lack of confidence in providing CF, (c) lack of experience of using CF, and
(d) contradictory guidance on CF provided in her teaching experience in the foreign country, it
was expected that her use of CF would be restricted to minimal. Among the 17 error-filled
AS-units, eight of them were not followed by CF while nine of them were followed by CF. Thus,
contrast to the expectation, she gave CF over a half of error-filled AS-units.
Further examining the types of CF that she used, among the nine CF-filled AS-units,
eight of them were recasts and one of them was a clarification request. Therefore, the majority of
CF that she used was recasts. In fact, despite her lack of confidence, she said that she might be
giving recasts, but “not consciously if I'm doing anything.” This is because she felt that recasts
were “something people just kind of do naturally a lot of times.” Therefore, it was expected that
her recasts would be used as a natural part of conversation, but without necessarily the intention
of CF. Whether recasts were used in a natural flow of conversation can be examined by looking
at the interaction between her and her students. Example 1 shows a case when she used recasts.
In this case, Cecile was asking the whole class the meaning of the word, debt, after confirming
its pronunciation.
Example 1
T: Debt. Ok. Good. You said it right, This B sound is actually it's, you don't say the B.
It's not /debt/, No, It's just the t. Debt. And what's debt.
Ss: (You own money for something, isn’t there).
S: Lend, you lend so many money from others, (that is a debt).
T: Yeah. So you said borrow. Right. You, you borrow so much money from others that

90

you have no more money. You’re in debt. Not good, right? Anyway, and then ah,
adolescents, what's adolescents mean.

In this example, the student made two errors, using lend instead of borrow, and many
instead of much. After introducing borrow, Cecile reformulated the sentence into a correct one,
using borrow and much. Then she moved on to the next word in the list to check its meaning.
This example shows that she was using recasts but not by themselves, instead, rather as a part of
her long utterances in the natural flow of communication.
Example 2 shows another case where she used a recast in addition to a clarification
request (the only case in Observation 1). In this case, she was asking the whole class the reason
why people tended to get married later than before these days.
Example 2
T: Twenty-five, twenty-seven. Ok. Ok. So good, what are some of the ah, why do you
think this is happening. What are some of the reasons.
S: Economic reason [accent on co]
T: What did you say?
S: Economic reasons. [accent on co]
T: Ok. Economic reasons, So what's that mean.

In this example, after the student used the accent on the wrong syllable of the word
economic, Cecile used a clarification request. Following it, the student made the same
pronunciation error while making the following noun plural, economic reasons, but this time,
instead of giving another clarification request, Cecile repeated the words back with the correct

91

pronunciation as a recast. Then, she asked the whole class what the word meant. This example
shows that, unlike Example 1, Cecile also used a recast by itself, not as a part of a longer
sentence. However, what are common between these two examples were (a) there was no
opportunity for uptake, and (b) the interaction was done fairly quickly without any emphasis on
the corrected parts. Therefore, it can be assumed that she probably did not have any strong
intention to correct any of these errors, which is line in with her stated beliefs.
In these two examples, Cecile was asking these questions of the whole class, so it was
natural for her to echo back to the whole class the answer that a student gave, but as she was not
necessarily obliged to repeat them verbatim, her response naturally functioned as CF. Also, the
students’ utterances were rather short, so it was easy for Cecile to give recasts. In fact, in most
cases when she used recasts, the number of words that the student said was only one to three.
Nevertheless, her use of recasts was not restricted to the cases when the students’ utterances were
short. Example 3 shows such a case in which she again was asking the whole class why the
marriage age was getting later than before.
Example 3
T: What do you think, [student's name]. Why do you think this is happening.
S: Ah (do) I, I think must now, this (i) more fo, focus on the economic and the people
are (busies) on the, the job so there no time to (think about ?).
T: Ok, So they're focusing on money issues. Economic problems. Right …

In this example, although stumbling, the student expressed his opinion that economic
reasons were one of the factors while making two errors, using the adjective, economic, instead
of the noun, economy, and omitting the copula verb is. His utterance was much longer than those

92

in the previous two examples, which may have made it difficult for Cecile to repeat the whole
utterance in the correct way. However, she used a recast in the form of economic problems,
adding problems to economic. This example shows that she used a recast even when the
students’ utterances were long. But similar to Example 1, her recast was a part of her longer
utterance, which may have made it difficult for her students to notice CF. In addition, the
common features between these three examples were again that (a) she did not provide her
student with any opportunity for uptake, and (b) the conversation was fast and there was no
emphasis on the corrected parts. It should be remembered that she emphasized the importance of
production for L2 learning at Interview 1. Without securing such opportunities following her CF,
she perhaps did not regard her CF as a chance for her students to learn something new.
In conclusion, her lack of knowledge and confidence regarding CF seemed to make her
refrain from using any kinds of CF consciously, but in fact, she was consistently using recasts
and clarification requests. Judging from her recounts and the actual interactions between her and
her students, though, her use of CF in the classroom was largely unconscious without any
rationale or philosophy. Instead, she provided recasts and clarification requests without any
intention of correction but rather of a natural response in the flow of communication with her
students. In the meantime, because she knew that recasts could be used naturally in the flow of
conversation, she assumed (correctly) that she was using them in her classrooms. Lastly,
knowing little about CF, she showed her curiosity to know more about it, saying, “this is
something that actually is really interesting to me because, this was one of the things that I'd
definitely reflected on a pi, um um, upon a lately we accurately we've talked about it so much.”

4.2 John

93

In contrast to Cecile, John was able to articulate his teaching philosophy clearly, which
consisted of two points: (a) using critical thinking skills, and (b) creating a comfortable
environment for students, each of which will be discussed in the following paragraphs.
John mentioned the importance of critical thinking skills repeatedly as it was one of his
main foci of his class regardless of the skills that he taught. To the question of what he regularly
did in his class, he replied:
T: Well, I think the big one like I said before was critical thinking, Yeah. And uh,
doesn't matter speaking, listening, reading, writing, content, uh, I really focus on
their thinking of critical thinking skills because I think their education systems,
specially our Asian students, doesn't really build that into their education. They can
memorize, I don't want them to memorize. My vocabulary. I don't want them to
memorize presentation in this class, I tell them that. Uh we practice little activities
where they get up in front of the class and have to speak about the interview that
they gave for example this past weekend, But I'm very clear that I don't want it to be
memorized or stilted speech, So, that's a very common theme, critical thinking.
Presenting looking into argument two different ways, uh, you know, being able to
counter-argue and, and debate, that kind of thing.
R: Ok. So uh, Ok. So, you, you want them to have argument which counterbalance
between, counterargument between, between these two opinions. Right?
T: Right. I think that's an element that builds critical thinking skills.

Corroborating this notion, he uttered the phrase, critical thinking, to his students twice
in Observation 1. In Example 4, while groups of students were discussing questions, he

94

approached one of them and asked what they were talking about.
Example 4
T: What are you guys discussing,
S: Ah, the talking about the /(what) (?) motivation,
T:

/Ah,

T: Yeah. I wanted you to do some critical thinking because some things I think we
didn't hear on here.

As the other incident, in a whole class discussion a student gave a funny answer,
saying that if someone died in a waterfall, it would cause water pollution. John praised his
unique idea by referring to critical thinking.
It should be noted that Observation 1 was conducted prior to Interview 1. Thus, his
references to critical thinking at Observation 1 were not due to the effect of Interview 1. At
Interview 1, recalling these incidents, he explained. “in fact you heard it on Friday a couple of
times I think the discussion questions is oh, here you guys remember your critical thinking …
we've talked about that a lot in class. They're used to that, they know what the phrase means.”
His emphasis on critical thinking was also evident from his collection of books on the topic.
When he was asked a question about professional development as a teacher, he stated:
When it comes to, well as you can see I have a lot of books. And I request books is
particularly about critical thinking, That's kind of my, my shtick when it comes to in
the classroom … so when it comes to critical thinking, I do look it a lot of books, but
honestly they're mostly teaching books. Even if they are not books that we use, we
could use here there's a couple of books that I have they're inappropriate for our

95

program. That I find interesting the way that they present and work on that critical
thinking.

The second point of his teaching philosophy, creating a comfortable environment for
students, was also a recurrent topic at Interview 1. Asked a question about how his teaching
practices had changed, among others he pointed out, “making people feel comfortable. And, and
making the classroom a safe place to be, when they come in, they don't have to be afraid, they
don't have to be nervous, they don't have to worry about raising their hand.” He came back to
this point also when he was asked about his teaching philosophy. He replied, “I wanted it to be as
organic, a natural and safe process as possible, specially in a speaking and listening class.
Reading writing, it changes just a little bit.” He was further asked why speaking and listening
differed from reading and writing classes, and although the question was about his teaching in
general, he adduced written corrective feedback in his explanation:
T: Well, I, you know, lot, the feedback a lot is written, And um, I tend to be, and I tell
my students that says, tend to be fairly harsh. With my writing feedback. Uh, I don't
praise as much that good as I do in, in speaking and listening class, I tell them that
upfront say look, uh, you know, we don't have a lot of time, you don't have a lot of
time, when you get a paper back yeah, you're gonna have green all over it, but that's
how I wanna help you. I want you to look at that and examine what you think you
did how you could improve,
R: Whereas in speaking and listening, you tend to do more uh praising?
T: Yes. More praising because of that, like I said at the beginning, it's crucial I think, to
have a comfortable environment.

96

At another point during the interview, he once again mentioned the same point:
I'm very open and … I hope to create the classroom as very comfortable, that the
students are willing to take chances, are willing to ans, ask questions and answer
questions and talk in groups and, That's very important to me. And I think that, not
only is that the … theoretical approaches. That have been … kind of proven that
they're more effective them, uh, communicative approaches, but more experiential.

Such a strong stated belief was partly coming from his L2 learning background. Before
and after the previous quote, the conversation went on:
T: Well, I think that as you are a learner of a language, you become aware of pitfalls,
you become aware of the psychological aspect of learning a language, of uh, what
kind of embarrassment and fear does to one, you know, when not willing to
participate, uh, so, I learned a lot of things at the time I didn't know were important,
but looking back I think you know that really shaped what kind of a teacher I am in
a lot of regards … for me, I, that wasn't encouraged a lot, And uh, I just felt nervous
and embarrassed in the class, instead of excited to learn, And I fe, and I told myself
you know, that's not how it should be. You should make, it, it should be an
interesting experience for everybody as much as possible.
R: So which means, when you took classes, (as) of French, uh Spanish and Chinese,
you were more like uh, nervous?
T: Yes.
R: In classes?

97

T: /Yeah.
R: /So,
R: You try not, you try not to make students feel that way.
T: Right.

These two aspects of his teaching philosophy had an influence on John’s use of CF.
The first point, critical thinking skills, had an indirect impact on it. He said that, as he grew to be
a more experienced teacher, he came to be more conscious of the goals and objectives of the
lesson. At Interview 2, he expressed his concern that CF may interfere with achieving them,
including critical thinking skills. When talking of one of the claims that Truscott (1999) made,
which was “error correction interrupts communication,” the conversation went on:
T: … I wouldn't choose to interrupt a lesson that's based on listening comprehension or
critical thinking or talking about, and then suddenly, boom. All right, no, no, no.
That's wrong. Um, now let's look on the board for five minutes, and nyu nyu nyu. I
wouldn't choose to do that because that wasn't our goal for that class.
R: So maybe not communication but rather than like a class flow /or objective of the
class?
T:

/Right. Yes. Yes.
You're right.

T: Communication, but for me that will break communication too. By stopping
Something and you know, students start to focus on that, and then I pull them back
and continue, I just think that doesn't seem very efficient. Yeah.

98

Table 6
John’s Use of CF at Observation 1
Error-filled AS-units
Kinds of CF-filled AS-units
Without
With
Clarification
Explicit
Total
CF-filled
CF-filled
Total
Recasts
requests
correction
AS-units
AS-units
53
52 (98.1)
1 (1.9)
1
0 (0.0)
1 (100.0)
0 (0.0)
Note. The class lasted approximately 50 minutes. The numbers in parentheses show percentages.
With critical thinking being one of his primary objectives in his lesson as shown
above, it was expected that he would not give a lot of CF because it might compromise achieving
it or interrupt communication. As can be seen in Table 6, in Observation 1, 53 error-filled
AS-units were identified, which was far more than that of Cecile. Because this class had plenty
of time for group discussions, his students made quite a few utterances as well as errors.
However, as expected, his use of CF was strikingly limited in that he used only one token of CF,
which was a clarification request. Example 5 shows a case when he preferred not to give CF as
he prioritized critical thinking.

Example 5
S: And so, it could /also make the water pollution.
T:

/Yeah.

S: ((laugh))
T: Water pollution?
S: Yeah. /If he die.
S:

/(?) people.

Ss: ((laugh))
T: That is critical thinking. [student's name]. Good one. Yeah …

99

In this example, a student made an error in the third person singular in he die. Instead
of giving CF, John chose to praise his idea by referring to critical thinking. Example 6 is another
case in which groups of students were discussing why some people would risk their lives by
doing crazy things.
Example 6
S: I think they also have another reason that maybe that person feel really lonely, but
they won't, ah nobody care, they think,
T: Ah,
S: Ah, nobody care them.
T: Uhn-uh,=
S: =So they want to let more people to, to focus on him.
T: It's good thinking. That's a very good thinking.

In this case, the same student made three types of errors: (a) the three consecutive
errors in the third person singular, in person feel and nobody care twice, (b) omission of a
preposition after care in nobody care them, and (c) using a to-infinitive, to focus, after let instead
of a bare infinitive, focus, in let more people to focus. John kept back-channeling, and in the end,
he simply praised her way of thinking, without giving CF for any of these errors. Thus,
encouraging critical thinking was regarded more important than giving CF.
Examining the other cases when he gave no CF, what was common was that there
seemed no communication breakdown (although admittedly, it is also possible that John simply
did not try to clarify it without understanding it). Example 7 shows a case where the students

100

were discussing in groups how adventurous they were.
Example 7
S: But when I ah, have already died, I feel ah, ah very good ah personal satisfication.
T: Ok.

In this example, the student pronounced the word, satisfaction as satisfication.
Although it should have been quite easy to give CF on this as the error lay only in a single word,
John simply said, “ok,” and another student started talking following his turn. In this example,
there seems to be no communication breakdown, assuming that John understood that what his
student meant was satisfaction. As another example, in Example 8, a student told John of a
friend of his who crossed China by bicycle.
Example 8
S1: I, I knew a friend of friend these, He ride bicycle across the China.
T: /Wow.
S2: /Wow.
S1: Takes two month.
T: Two months, that's all?

In this example, the student made four errors in total: (a) failing to change the verb to
the past tense, or possibly the third person singular form in he ride, (b) failing to change the verb
tense to the past tense in takes two month, and (c) failing to change the noun to the plural form in
two month, and (d) adding the before China. However, none of these errors interfered with
understanding the message. Therefore, apparently John had no difficulty in comprehending what

101

the student meant and showed his surprise instead of giving CF. As can be seen in these two
examples, he preferred not to give CF as long as communication was successful.
In regards to the other reason of his non-use of CF that may interrupt the flow of
communication, some examples are shown below. In Example 9, students were discussing
dangerous sports, and one of them did not know the word, skydiving, so he called John and asked
for the word.
Example 9
S: How say that action, jump out the airplane with my parachute.
T: Ah, skydiving.
Ss: Skydiving.

In this example, the student made an error in omitting the auxiliary verb do and a
subject (e.g., do you) in how say. In this case, giving CF would interrupt the flow of
communication because the student was expecting to hear the answer to his question. Thus,
instead of giving CF, John responded to the content of the student’s utterance.
Example 10 is another case in which students were discussing in groups what extreme
sports that they would like to try out. Answering to John’s elicitation, one of the students said
that he chose the marathon. However, another student did not know what a marathon was. So the
first student tried to explain what it was, but he was not sure how long people run in a marathon,
and in his utterance he made a linguistic error as well as a content error. John responded only to
his error in content, but not in language.
Example 10
T: Yeah. What did you choose,

102

S1: Well ah, in my case, I will, do the marathon,
T: Marathon. /Oh, wow.
S1:

/Marathon.

S2: What is marathon.
S1: Marathon like running the thirty miles?
T: Twenty-six.

In this example, S1 made an error in omitting a copula verb in the sentence, marathon
like running the thirty miles. He used a rising intonation to show his lack of confidence in his
answer. Then, John corrected the content without approaching the error in the language. In these
two examples above, CF would be likely to hinder communication as the students expected to
hear an utterance in response to their questions, but not CF. Thus, John refrained from giving CF.
As can be seen in the examples presented above, as long as the students were able to convey
their messages to him, he seemed not to bother to correct errors and carried on the conversation.
The relationship between the second point of his stated beliefs, creating a comfortable
environment for students, and his use of CF was more direct than the first one because CF can
possibly humiliate students who receive CF in front of other classmates (Numrich, 1996;
Truscott, 1999). As shown above, in John’s L2 learning experience, he was “nervous and
embarrassed.” Although it was not clear whether he was referring to CF in the classroom or
something else, in order not to embarrass his students, it is legitimate to assume that he would
not choose to give CF that may humiliate them, as evidenced above. At Interview 2, however,
responding to one of the claims that Truscott (1999) made, which was “error correction is
humiliating for some students, so they may stop expressing themselves,” he stated that CF would

103

not humiliate students if CF was given correctly. The conversation went on:
T: Yeah. I think, from my perspective that's not, I, I think I know how to run a class
now. Maybe when I was a beginning teacher, that, this could've been a concern.
R: I see. I see. So you don't think it's humiliating.
T: I think that if you do it correctly, no.
R: Ah ok. What do you mean correctly, ((laugh))
T: Well, I mean we know, what is, you know, stopping the class say, [student's name],
that's wrong. What were you, what were you talking about. Look at their sentence
again. There is a wrong, there is an error in there. What is it. You know, of course,
that's terrible. Right? But they're all, they know that they're not a native speaker of
English, You know, as long as I do it in the right, in the correct way.
R: Um uh, So correctly means,
T: Meaning is that I'm not judging them because they made the error, I'm not belittling
them in front of their classmates, Uh, [student's name], I can't believe you don't
know that word, Wow. You should've learned that in level 1, You know, I'm not, I'm
never gonna do something like that.

His comment above sounded that he would prefer to give the implicit types of CF
rather than the explicit types of CF; however, further asked for clarification, it was found that
that was not what he meant:
R: Ok. Ok. So does, does that mean that you are trying to make it implicit so that it's
won't be humiliating? Or,
T: No.

104

R: No.
T: No. Doesn't mean that. Um, I'm, I'm just saying that even explicit correction, if, if,
you know, when you do that correctly, it doesn't have to be belittling to a student.
Yeah. I feel they are in a class, where ev, nobody speaks perfect English, there's not
an native speaker and they're all, you know, in the same boat so to speak. And they,
hopefully expect that there will be correction, that's how they were, are gonna get
that, get better.

He assumed that his students expected to be corrected by him when they made errors
because they were still learners of English and none of them spoke perfect English. Thus, to
John, “correct” CF did not involve the degree of explicitness or implicitness of CF. Rather, the
way of his giving CF “correctly” was not to “belittle” students via criticizing their lack of
knowledge or skills.
When John was asked specifically about CF at Interview 1, however, he did not have
much to say because he said that he did not think about CF as much as he used to. He did not
negate the use of CF, but expressed his hope that it was now a part of his teaching and that he
gave CF unconsciously and “naturally,” especially when there was a communication breakdown.
The conversation went on:
T: Yes. I think that, frankly, I haven't, I don't think about that so much any more
consciously. Uh, I'm hopeful that still happens, I know there, in since it that does
happen, or student will make a mistake, the, especially that hinders communication.
And then, you know, naturally I would say, oh, let's back up for a minute. What do
you mean by blah blah blah. And then, they'll re, rehash it, recast it, and then, we go

105

until we get the right thing. Um, but it is absolutely true that's not consciously in my
mind as much as it used to be.
R: You used to think about that more when you were younger?
T: Yes.
R: But how did you tend not to think about any more?
T: Well, I think, I hope that it is because it comes naturally, uh, and I think that I don't,
I think just very naturally I tend to deal with those situations when they arise
naturally.

There was only one incident of apparent communication breakdown in Observation 1,
in which, as he stated, he gave CF. In this case, groups of students were discussing what they
would do to seek a thrill.
Example 11
T: … What did you say about number three.
S: Oh, (?) just jump from the /waterfloor/.
T: One floor?
S: Oh, waterfall.
T: Oh, waterfall.
S: Yeah.

In this example, the student mispronounced the word, waterfall. Apparently being
unable to understand the word, John used a type of CF, namely a clarification request. Then the
student self-corrected, which helped John figure it out, and John pronounced the word in the

106

correct way. Because the student was able to pronounce it correctly, John did not need to give
another token of CF. In addition, this example shows that he obviously did not belittle his
students by pointing out the student’s mispronunciation of the word, waterfall, confirming the
other aspect of his stated beliefs of CF.
In conclusion, John seemed to have clear teaching philosophy, but did not articulate
many of his stated beliefs on CF. He used to think of CF a lot more when he was younger, but as
time passed, his classroom practices of CF became routine as a part of his teaching practices, and
he came not to be as conscious as he used to be, which led him to hope that he still gave CF
naturally and unconsciously. This is why he was not able to articulate much of his policy on CF
when asked at Interview 1 as the majority of his beliefs on CF had already become rather
implicit.
In reality, however, his use of CF at Observation 1 was surprisingly limited to only one
clarification request when there was a communication breakdown. It should be noted that he
seemed to have accomplished one of his teaching philosophies, which was to create a
comfortable environment for his students, even at the time of Observation 1, which took place
only four weeks into the semester, and throughout the present study. This was evident from the
fact that most of his students talked actively, extensively, and loudly in every observation.
However, the more they talked, the more errors they produced. At least at Observation 1, though,
most of these errors were not followed by CF. This is because, as long as communication was
successful, John did not interrupt interactions by giving CF, and chose to carry on the lesson. He
was consistent in this regard and CF was used only to assist communication without the function
of correction.

107

4.3 Jim
Similarly to John, Jim seemed to have clear policies on his teaching as well. When
asked about his teaching philosophy, he replied that he had “quite a few,” which were all
“practical,” but not “theoretical grounded principles.” He adduced three of them, saying that
those were the ones that he could think of at that moment, which were (a) the class should be
student-fronted, (b) teaching morphosyntax and lexis was important, and (c) the teacher should
not be too strict with their students about their misdemeanors related to class management, each
of which will be discussed in the following paragraphs.
The first point that the class should be student-centered was not evident in Observation
1. Jim acknowledged the practical limitation of applying this rule, saying, “it can’t happen in
every class, like it didn’t happen all that much in the class you saw.” However, he said that this
was an important element in order for his students to produce an L2, using an example of soccer:
At least when we coach soccer around here, when you have a practice … it's
considered bad … if the team players are not touching the ball. We, even if they are
just standing around, waiting for their turn in something, they should be kicking a ball
or on they should be juggling, they should be, they should be doing something with
that ball. And it's the same thing in the language classroom I think that they need that,
as much as possible you, you should have them manipulating.

Although the class for Observation 2 was also a review of a lecture that Jim held in the
previous class and his students did not have many opportunities to speak up, the class for
Observation 3 was almost exclusively spent for the students’ presentations, so they spoke
extensively.

108

The second point, the importance of teaching morphosyntax and lexis, was coming
from his “strong” stated belief that “there is not much of a division between lexis and syntax,”
and that “most phenomena are morphosyntactic, not morphological or syntactic.” He seemed
quite adamant in this principle and further stated:
To my mind … they haven't learned a vocabulary item until they've learned it's likely
collocations until they've learned … whether it triggers uh active clause or passive
clause … all of those syntactic features that are tied to word choice. And what makes
your choice of syntax purely the word choice if you chose a different word, the syntax
will change.

Such a strong opinion on morphosyntax and lexis was derived from his background
and experience as a L2 learner and as a MA student in graduate school. First, he was a successful
L2 learner. For example, although Latin was required only for two years in his high school, he
took it for four years because he “learned to read Latin pretty well,” and “was good at it.”
Recalling his French class that he took in high school, he stated, “it was never anything real
challenging,” “I could’ve learned a lot faster,” and the one that he took in college, “it was dead
easy,” and “I could’ve done that in my sleep.” Therefore, although he “enjoyed” these Latin and
French class because “it was fun,” in fact, he “never took it really seriously,” and “didn’t learn
all that much.” Further asked why he was so good at them, he showed his interest in grammar:
I think it was mostly because I was interested … in the grammar when we were
learning it, so I had kind of a sense of the grammar. And also I'm really sensitive to
cognates, and I was able to apply that, and so … it was not a tough course.

109

At another point during the interview, he also expressed his deep interest in
vocabulary:
I am really really interested in vocabulary … so much of the basic English words stock
is, is Latinate. And, um, I spent a lot of time just thinking about how, how words
morphed from Latin into English. And thinking about that, I am sure influences the
way I explain vocabulary to my students. Very often … I'll show them the Latin root
and its meaning, and then the English version and its meaning, and then trace how it
got from one to the other.

He expressed not only his interest, but also his confidence in grammar: “I can consider
myself pretty, pretty sensitive to grammar … I think that when a student asks me a grammar
question, I'm pretty good … at figuring out what the answer is and telling the student.” Although
he did not express his confidence in vocabulary explicitly, he constantly read research studies on
vocabulary, and had even written some books on it. Thus, it would be natural to assume that he
was confident in vocabulary as well.
Another piece of evidence of his interest in and another source of his knowledge of
grammar and vocabulary comes from his graduate study. While he was in the MA program, he
avoided taking pedagogical classes as much as possible because they were “boring and really
obvious,” and instead, he took such classes as phonology, syntax, and morphology as much as
possible. With such a strong linguistic-based background, he used the terminology for linguistics
fluently throughout Interview 1, such as synchronic language, phonetically simple, vowel,
diphthong, consonant cluster, prefix, root word, active verb, cognate, synthetic, conjugation, and
declension. Congruent with his status as a professional editor, he had abundant knowledge of

110

grammar and vocabulary.
Regarding his third principal, the teacher’s generous attitude toward students, he
explained it:
If something goes wrong, if there is a mistake or if there's even like a breach in the
rules, uh, don't come down on 'em real hard like a Hitler. Um, sort of … express
disappointment, ask them what the problem was … putting back on them, rather than
creating this, this high low conflict situation.

Later, he adduced an example of a case when a student came in late:
Normally I would stop and I would say what happened … and then get them to, get
them to explain rather than saying you guys are late again. This better not happen again
… I don't think you, they, you do yourself any favors by establishing that relationship.
But they still know that I marked them late in my, my grade book. And they, they
know that there're consequences if they do it a lot.

In regards to the use of CF, at Interview 2, he mentioned that he would not give CF
extensively owing to lack of time:
I'm not the kind of person who gives a lot of oral feedback anyway, because … when I
walk into a classroom, I've always got a lot of stuff I want to accomplish. And I'm not
likely to just kind of step back and leisurely take the conversation in another direction.

Furthermore, when asked the goal of the class that was observed, the first point that he
mentioned was to “get through at least … six, probably eight ah, long lectures.” Having such

111

clear and definite goals to cover certain topics, it was expected that he would avoid using CF
because to him, giving CF was time-consuming as it may lead the class “in another direction,”
deviating from the main focus of the lesson, or at least, slow down the pace of his teaching.
Table 7
Jim’s Use of CF at Observation 1
Error-filled AS-units
Kinds of CF-filled AS-units
Without
With
Clarification
Explicit
Total
CF-filled
CF-filled
Total
Recasts
requests
correction
AS-units
AS-units
21
9 (42.9)
12 (57.1)
12
11 (91.7)
1 (8.3)
0 (0.0)
Note. The class lasted approximately 50 minutes. The numbers in parentheses show percentages.
As can be seen in Table 7, in Observation 1, 21 error-filled AS-units were identified,
whose number was a little more than that of Cecile, but far less than that of John. Among those
error-filled AS-units, 9 of them (42.9%) were not followed by CF while 12 (57.1%) were
followed by CF. Thus, contrary to the expectation, he used CF as much as Cecile did. However,
except one case, all of his CF was recasts. His almost excusive use of recasts may be due to one
of his stated beliefs on CF. When asked his policy on the use of CF at Interview 1, without any
hesitation or pause, he was able to quickly state his belief that “I'm willing to give the correct
answer and explain why, rather than go through a long asking process.” The reason for this
straightforwardness was expressed while acknowledging the practical limitation:
I think in most cases … unless you can draw that out with one or two questions, there
is no point going through this whole thing with the students because you get random
guessing. And to me that just clouds the issue … you wanna stay focused on … where
the error lies. And, so I'm, I'm very willing to say no … if, I can see why you did that,
because this is similar to that word … I know you learned this other structure for this
meaning, but that was only one way to do it, and, no, I can see why you did it, but this

112

is the way we do it, this is why … if I can say why. Sometimes you can't say why.

This notion was derived from his experience as a L2 learner. Partly because he was
such a good L2 learner, he was frustrated by his teacher’s useless inquiry when he was not able
to even guess the answer.
I did get annoyed … as a learner early on, when, I considered myself a pretty good
language learner, and the teacher really didn't have to say very much, for me to get
what the problem was, Uh, just a tiny tiny little hint, and I could usually figure it out.
But if I couldn't figure it out, and the teacher kept asking me about it, I kept thinking
well this is a waste of time … it's not in there no matter how many questions you ask
me. It's not in there. So I'm never gonna give you the right answer. Just tell me what
the right answer is because you can't get it from me … you generally don't get, uh, you
don't make progress by asking a lot of questions from somebody who doesn't really
know it.

As mentioned above, most of his CF was recasts. Thus, although not all of his recasts
were given consciously as the stimulated recalls later showed, he gave the correct answers to his
students via recasts as he stated. His abundant knowledge and confidence in grammar and
vocabulary probably helped him feel comfortable about such direct correction in cases when he
consciously gave recasts. However, there is one more point to verify, which is whether he added
the reason of the error to the answer. His exclusive use of recasts at Observation 1 shows that,
regardless of his intentionality of correction, he did not add explanation of the error (although of
course, if he did not have an intention of correction, he would not add any explanation following

113

CF). Judging from the fact that he preferred to use recasts over other types of CF that could go
with explanation, it seems that he was not eager to add explanation to CF. Examples presented
below will prove this point. In Example 12, after listening to a CD of a lecture about the people
who jumped into the Niagara Falls, Jim asked several comprehension questions of the lecture of
the whole class.
Example 12
T: Ok. Who was trying to save their life.
Ss: The government.
T: The government. Ok. And how did the government try to save their life?
S: Make it illegal.
T: Making it illegal. So this section was about the illegality, the illegality …

In this example, the student failed to change the verb, make, into a gerund, and
possibly to add a preposition such as by. After Jim used a recast to the whole class, without any
explanation, he moved on to summarize what they heard in the listening exercise. As another
example, in Example 13, Jim was again asking a comprehension question about the content of
the lecture that his students just listened to via CD.
Example 13
T: …but what do you think that last section was about.
S: Category.
T: Categories of what.

In this example, the student failed to change the noun, category, to the plural form. Jim

114

used a recast, providing the correct form, not by itself, but as a part of his following question to
elicit more information on the answer. Again, he did not provide any further explanation. In both
examples above, CF was incorporated into Jim’s long utterances without any emphasis on the
corrected part. In addition, there was no opportunity for uptake. Thus, again, he might not have
been giving CF consciously with an intention of correction in these examples (however, it is also
possible that he was giving CF consciously with an intention of something other than giving CF,
such as eliciting further information on the topic as can be seen in this example).
In addition to giving the correct answers by recasts, Jim also used a clarification
request. It is unknown whether he used it having the intention of CF, but it was followed by a
recast as shown in Example 14. Jim was eliciting a person’s name that his students heard in the
same listening exercise.
Example 14
T: … Let’s hear from this side of room.
S: /Cook/ Jones.
T: Who?
S: /Cook/ Jones.
T: Ok. What did he do. Kirk Jones.

In this example, the student mispronounced Kirk as Cook. Jim used a clarification
request by saying, “who?” but the student’s pronunciation remained wrong. Instead of asking the
same question one more time, he shifted the focus of his question to what Kirk Jones did.
However, after finishing asking the question, he repeated the name with the correct
pronunciation as a recast.

115

Finally, considering his focus on teaching morphosyntax and lexis, it was expected that
when giving correct answers the target feature of his CF would be morphosyntax or lexis errors.
Among the 12 error-filled AS-units that received CF at Observation 1, eight errors were in
morphosyntax or lexis whereas the other three errors were in pronunciation (as described above,
both a clarification request and a recast were used for a single error-filled AS unit). Therefore,
although it is true that he was mainly concerned about errors in morphosyntax and lexis errors
and gave recasts on them, it did not mean that he disregarded errors in other categories, but gave
recasts on them as well.
In conclusion, Jim’s stated beliefs on CF were clear and straightforward. He said that
he would not use a lot of CF, but when he did, he would give the students the correct answers
possibly with some explanation. In his actual classroom practices, however, he usually provided
CF for more than half of the errors in his class, most of which were recasts. Although it is
admittedly true that not all of his recasts were necessarily given consciously, Stimulated Recalls
2 and 3 showed that many of his recasts were consciously given. Thus, his primary principle of
giving CF, which was to “give the correct answer … rather than go[ing] through a long asking
process” was conforming to his extensive use of recasts. Finally, regarding adding explanation to
CF, such tendency was not evident from the data. This may be owing to the time constraint that
he mentioned as described above because both of these moves would take extra time prior and
posterior to CF.

4.4 Tom
When Tom was asked about his teaching philosophy, he talked about it extensively and
determinedly more than Jim. His teaching philosophy can be categorized into three points: (a)

116

students must use the L2, (b) motivation is the key in L2 learning, and (c) everyone can learn an
L2, all of which are related to each other as will be shown in the following paragraphs.
Regarding the first principle of the use of the L2, he stated:
There's cert, something that are required by the thing you are learning. If you want to
learn to write, you have to write, even if you don't like to write. Even if you are not
very good at it, If you wanna learn to write, you got to write, It just a requirement. So,
you can’t not write and learn to write. Right? The same goes with speaking …You're
not gonna learn to speak if you don't actually speak.

Table 8
Tom’s Use of CF at Observation 1
Error-filled AS-units
Kinds of CF-filled AS-units
Without
With
Clarification
Explicit
Total
CF-filled
CF-filled
Total
Recasts
requests
correction
AS-units
AS-units
46
40 (87.0)
6 (13.0)
6
5 (83.3)
1 (16.7)
0 (0.0)
Note. The class lasted approximately 50 minutes. The numbers in parentheses show percentages.
Because this was a speaking and listening class, derived from his comment above, it
was expected that he would give plenty of opportunities for speaking to his students, which
would be likely to yield many errors. In line with this expectation, as can be seen in Table 8, in
Observation 1, 46 error-filled AS-units were observed, whose number was almost as many as
that of John. This stated belief of his regarding the importance of student output came from his
L2 learning experience, in which, despite his sojourn in another foreign country for a long period
of time, he could not learn the language spoken there. He adduced the example:
I'm a perfect example of this … So I use that negative example with my students, So
do you wanna be like me? You wanna spend seventeen years somewhere wh, where

117

you can't even carry on a conversation in the language that surround you? Is that what
you want? Well, I know how to do that, Don't speak. Don't try. Right? So, so I can
even use the negative things that I have in my language learning, thing to, to motivate
students to realize that they, if they do that, they're gonna end up the same way. So,
you have to speak if you wanna learn to speak.

As he used the phrase, to motivate students in the above quotation, this relates to his
second principle of motivation. That is to say, he believed that, in order for his students to use
the L2, they must be motivated to do so. This was partly coming from another L2 learning
experience in which he learned Spanish through the audiolingual method. He confessed that, in
contrast to the miserable failure described above, he progressed “very quickly” because “every
day [he] lived with a Mexican family and traveled back and forth” and “used [Spanish] all the
time.” At Interview 2, he further commented that the reason why he succeeded in learning
Spanish, despite the fact that the class was “boring,” was that he “was highly motivated to learn
it” because he “needed it to function where [he] was.” This experience led him to believe that
motivation was the key to L2 learning success. He strongly stated:
I … believe that every student is … motivated. There's no such thing as unmotivated
student, They just not motivated to do what I want them to do, and what they need to
do. So then it become my job to, to work on shifting that motivation in the best way I
can.

In the meantime, however, he acknowledged the difficulty in motivating every single
student as each of them had different learning styles and strategies:

118

It's matter of how you teach them. They can't learn the way we are doing it, but there're
the ways for them to do so, if you can motivate them. They want to be motivated they
want to do, So, the key is motivating students to do the things they need to do, There're
many ways of learning, there's not just one, and so, uh, I used to think that … there are,
oh there's a magic bullet accelerated learning this is it, what what works for one person
doesn't necessarily work for another, so you gotta do many different things, Uh, mm, if
you do many different things, and eventually you reach someone.

This is apparently why he did not believe in any single L2 teaching method or
approach:
I'm not ah, I don't believe in any one method. I used to, but I don't. I try to take the
best, doesn't mean that I don't think certain method would work better than others for
certain purposes, I do. But, there's no one best method. And anybody who says that
there is has no idea what they're talking about. None whatsoever.

Motivating every single student was his goal, and in order for him to be able to do so,
his goal as an L2 teacher was to be a “great teacher” who could reach all of the students, as
described below:
I had a goal. 'Cause I wanted to be, I wanna become a really great teacher. I don't
maybe yet. But I'm working at it. Uh, I learned in seminar at and took on teaching
stuff, The guy, the presenter, it's seventy teachers in the room, of all kinds, not just
language teachers. And he ask uh, tell us the difference between good teacher and great
teacher. So we had to discuss all the stuff, and report back … then normal kinds of

119

things, And nobody out of seventy people could come up with what the difference was
… he says a great teacher reaches seventy percent of the students …a good teacher
reaches seventy percent of the students … He says the difference is the great teacher
reaches seventy percent of the students at any one time. But over the course of a class
or week or or some period of time, he will reach a different seventy percent at different
times, so eventually reaches all hundred percentage students, at least some of the time.
Whereas … the good teacher reaches the same seventy percent of the students all the
time. Thirty percent are left out. So, ever since then, I've been trying to be a great
teacher. So I tried to do stuff that will help individual students.

Finally, his third principal that everybody is capable of learning an L2 was related to
his second principle because he admitted that everyone can learn an L2 only “if they are
motivated.” Still, he was adamant in this point:
I believe that every student can learn. All right? And when student don't learn, it's,
because there's a problem with the way it's been presented to … I think that … there is
no such thing as ah, is, so, someone who can't learn …a lot of people don't think that, I
believe that, I don't believe that there're people who can't learn languages.

This stated belief that everybody was capable of learning an L2 was derived from other
two stated beliefs that he had: (a) brain-based learning, and (b) “language learning is innate.” For
the first one, he believed that “the brain is … pattern making organization, [so] if you teach
people patterns, they learn them.” As he had “done lots and lots of studies and lots and lots of
seminars on brain-based learning,” he had abundant knowledge on brain-based learning.

120

Therefore, he seemed confident in this point. And for the second part, he strongly believed that
L2 learning was equal to L1 learning. These two points together were the basis on which he
rejected the use of the explicit types of CF, but not CF altogether, because, as shown later, he
acknowledged his use of recasts.
He did not recount why brain-based learning would negate CF, so the relationship
between these two is unclear. However, at Interview 2, responding to one of the claims that
Truscott (1999) made, which was “students may not understand error correction,” he implied that
he believed that human brains would not let people hear CF because they would hear only what
they expected to hear, rendering CF useless:
We hear what we expect to hear, we see what we expect to see, It's how we survive, s,
we fill in things that aren't there, And we, we take out things that actually are there, It's
how the brain functions. And it happens to students in class with error correction. They
don't hear it, so, yeah. They don't.

His quote above seems to indicate that he was basically negating the usefulness of any
kinds of CF. However, this does not mean that he was rejecting the use of CF altogether. As will
be shown later, he accepted the natural use of recasts. Still, he was obviously dubious of the
effectiveness of CF, and accordingly, his use of CF was fairly limited. In Observation 1, among
46 error-filled AS-units, only six of them (13.0%) were followed by CF.
His minimal use of CF was observed regardless of the difficulty in giving CF. Example
15 shows a case when it was relatively difficulty to give CF for him owing to the long utterance
by the student. In this dialog, the whole class was discussing whether they could allow their
friends to cheat in college classes. One of the students claimed that those who cheat would fail at

121

job interviews.
Example 15
S: Finally, you final (your will be found). The other do these things maybe, For
example, the job interview. He don't know this things, even he copied me, /but he,
he not actually been this knowledge.
T:

/Ah,

T: Ye maybe, sometimes, yeah. But some people are very good at job interviews.

In this example, the student’s utterance contained two error-filled AS-units in (a) using
don’t instead of doesn’t and the singular demonstrative determiner, this, instead of the plural one,
these, in this things, and (b) using an ungrammatical expression of he not actually been this
knowledge. But Tom’s response acknowledged the semantic content of the student’s utterance
without giving CF for any of these errors. This may be because it was difficult for Tom to
approach all of the errors as a single, rather long utterance had multiple errors. However, as can
be seen in Example 16, Tom did not give CF even for errors that could easily be corrected. In
this case, groups of students were discussing plagiarism. One of the students said that it was a
common practice in China.
Example 16
S1: In my country, just, just fine.
S2: Just fine? (No), first time, fine? Like there if you do it again. You will kick out?
S1: No no no. China (?) don't care (?).
S2: Don't care?=
T: =Don't care about that stuff. Cut and paste.

122

S1: Yeah. Lot of, lot of students copy in China.
T: Yeah.=
S1: =But (?) lot of very lot. So teacher don't care about it.
T: Ah,

In this example, S1 made two errors: (a) omitting a before lot of, and (b) failing to
match the subject and the verb in teacher don’t. Although these utterances were rather short, and
each of them was made in a separate turn, Tom simply replied by saying Yeah and Ah without
giving any CF. Therefore, his non-use of CF was not restricted to the cases when giving CF was
difficult; even when it was relatively easy, he did not give CF either.
Judging from the description above, although it seems that he was dubious of the
effectiveness of any kinds of CF, when he talked of the second point that language learning is
innate, he implied that he was referring only to the explicit types of CF while adducing an
example of child L1 acquisition: “you don't go around telling children, oh you said that wrong …
Well, you do, you say it right. You talk to them, and they pick it up.” In this explanation, he was
clearly negating the use of the explicit types of CF, but not recasts, which previous studies show
that are commonly used with child L1 acquisition (Loewen, 2011b). Accordingly, as will be
shown later, he stated that he would use recasts.
He was aware of the difference between the classroom setting and the language
laboratory setting and especially expressed his concern that CF may humiliate students in the
classroom, which may have an adverse effect on their motivation. He stated:
Laboratories is a great place to go to send somebody to practice … they wanna practice
how to speak, and making errors, and deh deh deh. But classroom is not a great place

123

for that, puts people on the spot, make some, make some look bad, It can demotivate
some students, Some students they look all right with, But no one really likes it, You
can see if you watch people's faces.

Therefore, although he did not negate the value of errors themselves by saying,
“mistakes are good. If you learn from it,” the explicit types of CF seemed to have no place to
play a role in his classroom as it would encroach upon one of his most crucial components in
teaching, motivation. While he rejected the use of the explicit types of CF, however, he showed
his preference to using recasts (although he did not use the exact term), possibly because he
thought that the students “wanna know how to do it right.” He said:
I don't correct errors generally speaking directly, I try not to do that, I sometimes I do
because I just, it's not, I, what I try to do is model the correct answer, so I'll repeat after
them something, but … I do it sometimes and I don't like it, but, I don't wanna say, oh
that's wrong.

When further asked why he did not like to give the explicit types of CF, he simply
said, “because I don't think it's helpful.” Interestingly, though, he acknowledged that there was
an individual difference in its effectiveness: “for most people uh, so some people that works. So
even in individual, somebody wants that, ok. But a lot, most people that, they don't like be told
they're wrong.”
Judging from his stated beliefs described above, it was expected that he would use only
recasts without any kinds of the explicit types of CF. In line with the expectation, among the six
CF-filled AS units that he used in Observation 1, except one case of a clarification request, all of

124

them were recasts (83.3%). In the case of the clarification request, whether he was simply unable
to hear or understand what the student said or he used this clarification request with an intention
of correction is unknown. In any case, he did not use the explicit types of CF and used only CF
that is commonly used in child L1 acquisition, recasts and clarification requests. This cogently
shows that his stated beliefs on CF were consistent with his classroom practices of CF.
Moreover, it was also expected that his use of recasts would be minimal as his belief in
their effectiveness was not firm; so his use of recasts would be largely confined to the cases
when they were used in the flow of natural conversation. It should be noted that, although recasts
are generally coded as an implicit type of CF, their implicitness varies and they can be fairly
explicit (Lyster & Saito, 2010). Examining his use of recasts, however, it is clear that his recasts
were quite implicit because they were used in the natural flow of conversation. In Example 17,
groups of students were discussing the differences in education systems in their home countries.
Example 17
T: All right, What about, who teaches university courses. Only professors?
S: Uhn, in China. Only professor.
T: Only professors. So it’s different from the US. Right?

In this example, although Tom provided the correct phrase by saying professors prior
to the error, the student who replied dropped the plural -s, saying professor. Tom then repeated it
in the correct form with the plural -s as a recast, and moved on to the next question. In Example
18, groups of students were talking what they would do if their friends asked them to let them
copy their essays. A student said that she did not know how to refuse such a request. Then, Tom
gave an example, but she showed her concern that such a response may break their friendship.

125

Example 18
S1: I don't know how to refuse it.
T: Refuse, you don't know how to refuse it? Let me explain it to you. It's very simple.
S1: Ok.
S2: Her mother is sick?
S1: Ah,
S2: Ah,
S1: No, /I'm (friends),
T:

/No. ((laugh))

S1: I think maybe you, maybe /I (?).
T:

/Very easy. Very small word, You just say no.

S1: But I think it's too,=
T: =I'm sorry I can't.=
S1: =It's too directly, maybe, Ok. Maybe broke the friendship.
T: If you give your paper, and she gets caught, and then you get in trouble, the
friendship will be broken. Right?

In this example, S1 made two errors, the latter of which was that the student made an
error in the verb, broke, by using the past tense. In his next turn, Tom began his utterance with
his response to the content of her concern. Then following it, he paraphrased her erroneous
utterance by reformulating the whole sentence by making the sentence passive and changing the
tense to the future by adding will.
In these two examples above, whether or not he used these recasts with the intention of

126

CF is unclear. Still, regardless of his intention, his utterances were fulfilling the function of CF.
Nevertheless, his CF was a part of his long utterance without any emphasis on the corrected
parts, rendering these recasts implicit.
In conclusion, Tom had adamant stated beliefs on CF, which were (a) that he was
dubious of its effectiveness and (b) that he would not use the explicit types of CF, because the
use of the explicit types of CF may demotivate his students, which violated one of his crucial
teaching philosophies. In addition, the rationale of his non-use of the explicit types of CF was
also reaped from multiple other sources: brain-based learning and L1 learning. His actual
classroom practices show that he used CF minimally, and when he used it, it was either recasts or
clarification requests. Such a phenomenon is said to be typical in L1 learning. Thus, his stated
beliefs and classroom practices of CF seemed to coincide with each other in this regard.

4.5 Discussion

In this section, the findings that emerged through a cross-case analysis across the four
teachers’ stated beliefs on CF, their classroom practices of CF, and the relationship between
these two are discussed.

4.5.1 Stated Beliefs on CF
The stated beliefs on CF of these four teachers can be roughly categorized into two
groups. The first group of two teachers, Cecile and John did not express any strong stated beliefs
on CF. In contrast, the second group of the other two teachers, Jim and Tom expressed an
explicit policy about their use of CF in their classrooms. However, examining their stated beliefs

127

in depth, two themes that were shared commonly by all of the four teachers emerged, which were
that (a) CF should not humiliate students, and (b) CF was not the central issue in their teaching.
John and Tom clearly stated the former point. Cecile touched upon the issue slightly.
Jim, although he did not state it explicitly, was also concerned about being too harsh on students.
Creating a comfortable environment for students to communicate and study seemed to be one of
the most important factors for a successful classroom that was understood mutually among these
four teachers. This may explain why the use of explicit correction was not observed at
Observation 1 because it could be considered as more humiliating than the implicit types of CF.
Regarding the latter point, they all seemed to deemphasize the use of CF in their
classrooms, regarding CF as something peripheral to their teaching. However, whereas for
Cecile, this was owing to her lack of any strong stated beliefs on CF, for the other three
experienced teachers, this was a firmly established stated belief on CF. Yet, each of them had
unique reasons for this decision.
John said that he used to think of CF more in the past, but as his teaching experience
accumulated, he came not to be conscious of it as much as he used to be, rendering his use of CF
unconscious and automatic. However, in contrast to Cecile who used recasts simply because they
were a natural part of conversation, John’s use of CF had been established and supported by his
knowledge and skills that he had gained through his 14 years of L2 teaching experience.
However, judging from the fact that he was mainly concerned of such other things as critical
thinking skills, communication, and creating a comfortable environment for his students, and that
he did not have much to say about CF at Interview 1, CF was not one of his main concerns in his
current teaching although it may have been in the past.
Jim clearly stated that he would prefer to give the correct answer possibly with its

128

reason rather than trying to elicit it from his students. Although he said that he was not the kind
of teacher who would give CF extensively, he had no objection to doing so. But at the same time,
he did not have any further thoughts other than this point. Therefore, it seems that CF was
something that he would use because it was a part of the teachers’ job to correct students’ errors
in the classroom, but not something that he would willingly make use of. This is also due to the
fact that one of his main concerns in his class was to cover a series of topics in the limited
amount of time. Although each token of CF would not take a lot of time, if he constantly used
CF, especially if it was followed by explanation as he stated that he may do, he would need to
spend a fair amount of time for CF. This further may have made him refrain from using CF
extensively in his class.
Tom, on the other hand, had multiple points that he adamantly believed, and he talked
of CF quite extensively. Basically, he rejected the use of the explicit types of CF, and favored
recasts over other types of CF. Similarly to John, Tom’s non-use of the explicit types of CF had
been established and supported by his past experience of L2 learning and teaching in addition to
his knowledge on brain-based learning and child L1 acquisition. He stated that he would not
mind using recasts naturally, but his comments implied that, similar to Jim, recasts were not
something that he would use actively because he was unsure of their effectiveness as CF.
Believing in the applicability of the innateness of L1 learning, where CF is not indispensable, to
L2 learning, he regarded CF as something not crucial in his teaching.
In contrast to these three experienced teachers, Cecile had not formed any concrete
stated beliefs on CF yet; therefore, she was still unsure about how to give what kinds of CF in
her classrooms. At the same time, though, she was well aware that she naturally used recasts in
her teaching because she knew that recasts were used in the natural flow of conversation.

129

However, she was uncertain about the effectiveness of recasts. Therefore, although her use of CF
was also not one of the main foci in her teaching like the other three experienced teachers, in
contrast to them, this was derived from her lack of any strong stated beliefs on CF. However,
after Interview 1, Cecile started to form stated beliefs on CF as time went with fluctuations,
which will be described in the following chapters.

4.5.2 Classroom Practices of CF
As their stated beliefs on CF were fairly similar, these teachers’ classroom practices of
CF were also largely identical in the senses that (a) their use of CF was quite limited, and/or (b)
they used recasts the most, except John who use only one clarification request. Regarding the
former point, on average, their use of CF was every 5.6 minutes (Cecile), 50 minutes (John), 4.2
minutes (Jim), and 8.3 minutes (Tom). The average rate of the use of CF per minute were 0.18
(Cecile), 0.02 (John), 0.24 (Jim), and 0.12 (Tom), which was much lower than 0.72, the average
of the 12 teachers in Loewen (2003).
Besides the fact that it could be merely the nature of the lessons observed, a few
reasons are in order. The first reason is that, as described in the previous section, the use of CF
was not one of their main concerns in their class. Related to this point, as mentioned by Jim
above, the four teachers may have been trying to go over the materials that they had planned to
cover on a daily-basis as well as the chapters throughout the semester as prescribed in their
syllabi. The speaking and listening class was only 50 minutes, held only once a day, four times a
week. This is hardly enough to cover their textbooks in most cases, so this situation may have
prevented them from noticing many errors or driven them to ignore many errors to keep the fast
pace of their teaching.

130

The second reasons is that, owing to the nature of the program for which they worked,
it was assumed that the teachers would use the communicative teaching style to some extent, and
official classroom observations (not specifically these teachers, but the teachers in the program in
general) showed that the teachers in the program followed it with a varying degree (program
staffer, personal communication, January 26, 2012). It should be noted that, in the strong version
of CLT, the focus of the lesson is exclusively meaning; thus, CF must be avoided. Therefore,
even if they tended to focus more on the content rather than the language, such a classroom
practice could be considered legitimate because it can secure enough time and opportunities for
their students’ production rather than interrupting their utterances with CF. Specifically, John
said that he would avoid CF that would hinder communication. Jim and Tom both learned how
to make a communicative syllabus in graduate school and Cecile was learning it concurrently. So
they were all familiar with this approach. With communication being the primary focus, it is
possible that, as long as their students succeeded in making themselves understood, the teachers
preferred not to correct the errors that did not cause a communication breakdown.
Because the teachers were focusing on the contents of their students’ utterances, rather
than their linguistic accuracy, the teachers might not have been alert for errors, which prevented
them from noticing them. If so, they may have been merely giving recasts and clarification
requests as a natural response to their students’ utterances without any intention of CF. Such a
hypothesis could be supported by the fact that among the 24 recasts provided at Observation 1,
only five of them were followed by opportunities for uptake.
Regarding the latter point (the extensive use of recasts), the present study also
supported previous studies (e.g., Y. Sheen, 2004) in the sense that recasts were the most common
type of CF. Among all types of CF that they used, recasts comprised 88.9% (Cecile), 91.7%

131

(Jim), and 83.3% (Tom). As written above, John did not use recasts in Observation 1; however, it
should be noted that he also used recasts the most among all kinds of CF in total throughout the
three observations as will be shown in Chapter 6. Thus, his nonuse of recasts at Observation 1
was probably due to the nature of the lesson observed. Also, there was only one incident of CF in
Observation 1, which was a clarification request, so this class was probably not congenial to
giving CF. There could be multiple reasons why these four teachers favored recasts over others.
One of the primary reasons can be that recasts are implicit, so they are not likely to humiliate
students, which is, as noted above, one of the common themes considered crucial among the four
teachers. On the individual teacher’s level, recasts were used because (a) they were a natural part
of conversation (Cecile), (b) they would give the correct answer (Jim), and (c) they are
commonly observed in children’s L1 acquisition (Tom). In addition, as described above, the fact
that the teachers wanted to go over the lesson plan can be another reason why they favored
recasts over other types of CF because recasts are one of the least time-consuming types of CF
among all. Regardless of their reasons, among all the choices of CF, recasts seem to be the one
that matched their criterion of what type of CF should be used, so it was understandable that
recasts were used the most extensively.

4.5.3 The Relationship Between Stated Beliefs and Classroom Practices of CF
Previous studies show that teachers’ stated beliefs exert a strong influence on their
classroom practices. However, the findings from previous studies show that the degree of
conformity between teachers’ stated beliefs and classroom practices can vary depending upon
what aspect of stated beliefs and classroom practices are investigated (Phipps & S. Borg, 2009;
Smith, 1996). In the present study, contrary to the previous studies that found discrepancies

132

between stated beliefs and classroom practices of CF among ESL/EFL teachers (Basturkmen et
al., 2004; Phipps & S. Borg, 2007), it seems that they conform to each other to a great extent
among these four teachers similar to what can be seen in some of the previous studies (Bailey,
1996; Brumfi et al., 1996; Hsiao-Ching, 2000; Johnson, 1992a; Mitchell et al., 1994a, 1994b).
Cecile did not have a clear mindset about CF, but was aware that recasts would happen naturally
in conversations. Accordingly, she used recasts the most. John said that CF would happen
unconsciously and automatically in the way to support communication. At Observation 1, John
provided CF only once, which was the only case when a communication breakdown occurred, in
which case, he used a clarification request. Having other things to cover in class that were more
important including critical thinking, he preferred not to give CF as long as communication was
successful. Jim’s use of recasts seemed to be consistent with his stated beliefs to “give the
correct answer.” Tom’s rejection of the explicit types of CF was evident as there was not even
one occasion of an explicit correction (however, explicit correction was seldom used by any of
these four teachers, except Jim as will be described in Chapter 6).
In summary, this chapter investigated four teachers’ teaching philosophies, stated
beliefs on CF, and classroom practices of CF. While the three experienced teachers seemed to
have established their beliefs firmly, Cecile was still unsure of them. Still, examining the
relationship between these three, it was found that they are congruous to each other to a large
extent. In the next chapter, the four teachers’ responses to the readings will be discussed. As will
be shown, it was found that their responses were deeply affected by their stated beliefs on CF
described in this chapter.

133

Chapter 5: The Teachers’ Responses to The Reading of the Research Studies

The previous chapter described the four teachers’ teaching philosophy, stated beliefs
and classroom practices of CF, and their relationship. In this chapter, their responses to the
readings of the three research studies, Lyster and Saito (2010), Mackey et al. (2000), and
Truscott (1999), will be described. Some of the findings and claims in these three studies were
congruous or incongruous with their stated beliefs. For example, both John and Jim believed in
CF, which was in line with one of the findings in Lyster and Saito (2010) and partially with
Mackey et al. (2000). Tom’s rejection of the explicit types of CF and Jim’s tendency not to give
a lot of CF could be supported by the claims made by Truscott (1999). Both Cecile and Tom said
that they would use recasts (for different reasons), but Lyster and Saito (2010) found that recasts
were less effective than prompts. In terms of their classroom practices, their frugal use of CF
would be in line with Truscott (1999), but their favor of recasts would be challenged by Lyster
and Saito (2010).
In this chapter, facing these matches and mismatches between these research studies
and the four teachers’ stated beliefs, how these four teachers read and responded to the three
research studies was investigated based on the data obtained mainly from Interview 2 in addition
to that from Interview 1, in order to answer the second research question:
2

What is the relationship between the four ESL teachers’ stated beliefs on CF and how they
read three research studies of CF?
2.1

How do their stated beliefs on CF prior to reading affect how they read research
studies?

134

It was found that not only reading research studies influenced their current stated
beliefs on CF, the latter also had an influence on the former, rendering the effect bidirectional.
The following sections will introduce the description of each teacher, which will be then
followed by the discussion for the whole chapter.
As described in Chapter 3, at the end of Interview 1 I gave each teacher a hard copy of
each of the three research studies as well as a PowerPoint file. They were instructed to go
through the PowerPoint first, and then read the research studies in any order that they liked. They
were told that they could read only the results and possibly the discussion section except Truscott
(1999), which I recommended that they read in its entirety because it was a position paper.
Because the teachers took different amounts of time to read the articles, the duration between
Interviews 1 and 2 varied in each one of them: 20 days (Cecile), 21 days (John), 31 days (Jim),
and 26 days (Tom).
Table 9
The Reading That the Teachers Did
Names
Summaries (PowerPoint)
Lyster & Saito (2010)
Mackey et al. (2000)
Truscott (1999)

Cecile
Read
Skimmed
Skimmed
Read

John
Read
Read
No
Read

Jim
No
Read
Read
Read

Tom
Read
Read
Read
Read

Table 9 shows how each of the four teachers read or did not read the research studies
and the PowerPoint file. Although they were told that they did not need to read the whole articles
except Truscott (1999), it turned out that when they “read,” they read the whole study. All of the
teachers except John either skimmed or read all of the three articles. John did not read Mackey et
al. (2000) for reasons that will be presented in his section below. Jim read all three research
studies, but stated that he simply forgot to read the PowerPoint file.

135

5.1 Cecile
Cecile read the PowerPoint file and Truscott (1999), but said that she did not have
enough time to read Lyster and Saito (2010) and Mackey et al. (2000), so she skimmed both of
them. Among the three research studies, while finding some points that she would agree with in
other studies, she most closely identified with Truscott (1999) because she said that it was
interesting, accessible, practical, and enlightening, and she began to form her stated beliefs on
CF based on this study, although her beliefs were still unstable.
First of all, she showed both her agreement with and interest in all of the articles to
some extent. She appreciated that these research studies gave her opportunities to think about
CF, saying, “I’ve never, never thought about, how I say things or how I should say things in
terms of giving like a recast,” and “I never had thought about the way students are gonna
perceive things, and thinking about the whole class as a whole.” Moreover, she said that she
wanted to read Mackey et al. (2000) again, saying “students’ perceptions … I would like to know
more about it … on a whole another layer.” However, her responses in depth to the content of the
three research studies varied, especially between Truscott (1999) and the remaining two studies.
As described in the previous chapter, her stated beliefs on CF were confused by contradictory
information that was provided from the foreign program and the school for which she worked. In
the present study, the three research studies were intentionally chosen so that their results and
claims would be contradictory, and she noticed it. However, this time, she clearly expressed her
preference for Truscott (1999) over the other two studies, saying, “I would tend to agree more
with what Truscott was saying than [the other two studies].”
One of the main reasons why she did not favor Lyster and Saito (2010) or Mackey et
al. (2000) was their lack of practicality, which was also one of the reasons why the teachers in S.

136

Borg (2009a) and Shkedi (1998) refrained from reading research studies. Cecile complained that
these two studies were presenting only what the researchers were interested in, saying, “[Lyster
and Saito (2010)] wasn’t stating what teachers should do, just what research has shown,” “maybe
less of um implication for what teachers should be doing based on [Mackey et al. (2000)],” and
“[Mackey et al. (2000)] doesn’t seem like it’s um, stating a definite view on what teachers and
students should do, it’s just kind of presenting the research.”7 Thus, she was not sure what she
could take out from these two studies, saying, “I can’t see how I can specifically relate [the two
studies], to apply, apply it to something,” and “I’m not sure how I’m gonna use [Mackey et al.
(2000)] for teaching necessarily.”
In fact, she favored Truscott (1999) for the same reason, that is to say, its practicality
to her teaching. She said, “[Truscott (1999)] laid out … things that are happening when you have
error correction.” Therefore, she was reading these research studies from the teachers’ point of
view, seeking pedagogical implications, in other words, a piece of hands-on, practical advice to
her everyday classroom teaching without caring about the validity and reliability of research.
Besides this point, she added another reason for favoring Truscott (1999), which was that it was
“easy to read,” and “pretty straightforward.”
This “easiness to read” was partly owing to the fact that Truscott (1999) was a position
paper whereas the other two studies were experimental studies. It seems that Cecile’s lack of
technical terminology had contributed to her perception of the difficulty of reading each study,
and accordingly, throughout Interview 2, she asked me to clarify such phrases as recasts,
clarification requests, prompt, morphosyntactic, interlanguage, and post hoc. This was
unsurprising, taking into consideration that she was still relatively inexperienced and a novice

7

All of the quotes in this chapter were taken from Interview 2 unless otherwise noted.
137

MA TESOL student. This shows the difficulty of understanding research studies of not only CF
but also SLA in general, even if the operationalizations of these terms were introduced in the
articles, especially for students who are still relatively unexposed to research studies, such as
those in MA programs.
Another reason for her fondness for Truscott (1999) was that it was “eye-opening” in
the sense that it gave her perspectives about which she had never thought. She stated:
I think there were more things in it that I hadn't ever taken into account … there's so
much more going on that, teachers um, well I know I don't always consciously think
about the different factors, so it was good for me to um read that, and think about that.

Additionally, she stated that one reason why she disfavored Lyster and Saito (2010)
was its lack of information about the 15 studies that were used for the meta-analysis, saying, “it’s
hard to know exactly where it is, from this study, from this to what happened here, we would
conclude this,” wishing “if I had a mini recap of some of those studies, it would have more more
of an impact on [me].”
Although she favored Truscott (1999) the most, she did not accept everything that he
claimed. She seemed to be critical when the idea did not agree with her own thoughts derived
from her teaching experience, however limited. She repeatedly said that Truscott (1999) may
have been exaggerating to some extent:
I felt like he exaggerated a little maybe he like said, oh, you won't know what's being
said and classroom noise and you can't hear properly, and certainly that does happen
but I don't feel like, perhaps that's always what's happening, I think you can
legitimately hear when a student makes a mistake.

138

Referring to one sentence, “[m]ost language teachers are not experts on grammar”
(Truscott, 1999, p. 438), she acknowledged that this was true at least in her case, but she still
refuted it somewhat by saying, “but I do think even if we are not grammar experts, any, to a
degree, I think any native speaker could point out oh, this person should’ve said this.” Another
disagreement from her was against one of his claims, “students may not take error correction
seriously.” She said, “if they’re being called out … they’re being targeted or something, they’re
gonna, probably remember the interaction … you’re gonna remember those that moment more.”
Although she did not have any firm stated beliefs on CF at the time of Interview 1,
after reading these three research studies, it seemed that she was beginning to form her stated
beliefs on CF and became able to articulate some of them. It should be remembered that
throughout the present study, she referred to herself as a “sponge.” At Interview 2, she said, “I
don’t know a lot about teaching,” and “at this point, anything I read I’m like oh that sounds
good, that sounds good.” However, examining her responses to the three research studies, her
“sponge” was “filtered” in the sense that she was selective about which study she would favor
(i.e., Truscott, 1999), and even in Truscott (1999), she did not take everything for granted and
rejected some of his claims. This shows that, although her stated beliefs on CF had not been
established, she at least had formed this “filter.” She was ready to “soak up” only the information
that permeated this “filter” while rejecting the other. Nonetheless, the data show that her stated
beliefs on CF were still unsteady, and even at the time of Interview 2, her statements fluctuated
and often contradicted each other.
Regarding CF in general, believing in Truscott (1999), she seemed determined to “give
less feedback,” saying, “I can understand why it would be negative, definitely to, especially from

139

reading [Truscott], negative to the classroom environment to students in, in so many ways,
negative to give … corrective feedback.” This notion was in line with one of the instructions that
she had received in her training in the foreign country. Thus, she was reading Truscott as
confirmation of this guidance. It also conformed to a point that she briefly mentioned at
Interview 1, in which she expressed her fear of giving CF as it may humiliate students.
Supported by the article, she explicitly expressed this point recurrently as a reason to reject CF,
saying “[I] don’t wanna humiliate students,” and “[CF] does make students feel nervous… if I
say some wrong, she's gonna tell me oh, you should say this or, so I'll just be quiet.” She further
stated:
As a whole class, when you call … you point out the student's mistake and the
students, oh other students are obviously paying attention and they are, perhaps, able to
see … what you were saying or maybe they're just lost or afraid to speak up then.

Especially, she expressed her concern about the explicit types of CF, stating that
students’ minds may be so occupied with thinking of errors that they would be unable to
concentrate in the lesson:
They might feel nervous, embarrassed, all of those things, so perhaps when you're
giving explicit correction, it seems that's the moment where um everybody or most
students are going to know that someone has made a mistake and paying attention to
oh she didn't say that right and then that student obviously is gonna be thinking about
all of those things so, perhaps less, it's less effective because there's all those other
things that they're thinking about in their mind and they're not going to maybe even,
even hear what the teacher says or pay attention or care 'cause they just worry about

140

being singled out in a sense.

Her doubts about CF also came from its ineffectiveness from the point of view of
noticing. Referring to Mackey et al. (2000), in which both ESL and IFL learners perceived CF
for morphosyntax the least accurately, she said, “I think that would be something that takes more
than just hearing the quick correction in class … it would take more repetition and study and
multiple examples to get that.” Rejecting the effectiveness of CF, she concurrently rejected that
of recasts too, which she was aware that she would use naturally in her classroom, and this was
supported by the fact that she used recasts the most in two of the three observations and also in
total. She stated:
Maybe it's not really that beneficial, it doesn't seem like … they don't recognize that
you changing the language sometimes so, so, on that note, I don't know if it's
necessarily harmful, but it doesn't seem like it's beneficial to do recasting.

To corroborate this notion, referring to a point in Truscott (1999) about how students
generally expect their teachers to provide them with the correct answers and adducing variation
in English, she stated that judging whether a certain expression in English was right or wrong
was not an easy task, which made it difficult for English teachers to give CF:
English is such … a global language … specifically thinking about pronunciation …
English is becoming less …less right and wrong. For example, when I was in [the
foreign country], they were talking of how they were gonna start allowing people from
India to come and be able to teach English in [the foreign country]. And I thought that
was a good thing because, um well, even many countries … people can speak English

141

well enough where … what is an English speaking nation … England really is the one
nation where … if you look at Malaysia, Singapore, Hong Kong … all these places
where people speak Engli[sh] … thinking of what is the right English or what is the
right accent or pronunciation or whatever … It is more fluid, It is less, there is less
right and wrong pronunciation, even grammar … sometimes I listen to the BBC and I
notice how they, have just very different structures that is not really possible in
American English so, in that sense, there really is not so much of um right or wrong, as
long as you know.

Moreover, agreeing with Truscott (1999), she emphasized the importance of
communication as another reason to avoid CF, saying, “I think [CF] does destruct
communication,” “you’re destructing the flow of communication when you give um, an error
feedback,” and “you don’t wanna disrupt the flow of communication, I think that’s definitely
more important.” It should be remembered that, at Interview 1, she emphasized the importance
of production in L2 learning. In order to secure such opportunities for students, she deemed it
critical not to interrupt communication; therefore, she found support for her prior stated belief on
CF and the importance of production.
She further compared the purpose of using languages to that of math and science in
which finding right answers is more valued:
Any teacher should … help people convey their ideas and um, in a way that people
understand what they're saying … If people don’t really get what you think they're
getting … communication is number one and there, there is less and less of um, a right
answer and a wrong answer, or maybe there always has been in language … it's, it's

142

not like math or science.

Nevertheless, in the meantime, she also expressed her dilemma about certain types of
CF from multiple perspectives, the first of which was that recasts are natural. She said, “I feel
like it's just second nature, to just recast or re-word the way someone has said something to make
it sound more natural and more what I would say as a native speaker.” She further stated:
I think, any person, no matter if they're teacher or not, communicating with um, a
language learner of the same language, has a tendency to want to recast. Tendency to
want to. 'Cause I mean, in, in a way, I feel that it kind of um, helps communication? I
mean, not not even, if you consider recasting also just to be … rephrasing and
paraphrasing … that shows that you're listening, you're engaged in the conversation.

At the same time, she reconfirmed that her of recasts was not grounded on any of her
teaching philosophies or stated beliefs.
In a sense, it's not that I'm consciously saying oh, they said it wrong, I need to correct
them. It's just more that, that's just the way, it's just like a habit or human nature, I
guess, not necessarily um teaching ideology or something.

In the meantime, knowing that she would naturally use recasts but believing that
recasts could not be beneficial, she wondered whether it would be possible for her to get rid of
her natural use of recasts:
I definitely would, um last night I went to movie … oh you went to the movies? I think
I would just naturally do [recasts], So, I, I can stop if I'm more conscious of it, and I

143

feel like [recasts are] what I definitely should do.

In addition to recasts, she said that clarification requests were also a part of a natural
conversation. She admitted that “[she is] in a habit of doing [recasts] and clarification requests.”
While again mentioning the possibility of abolishing the use of these two types of CF, she stated:
Recasting and … clarification request, certain types I think are inevitable for me, at
least … without um really working hard to not do those things, I think just in the flow
of communication, that's just how anyone, even two native speakers of any language
… that's just human nature to communicate that way, so, if I'm going to never do that, I
think it would be very hard.

In addition to the naturalness of these two types of CF, she also stated that she believed
that students expected to receive CF from their teachers. Citing Truscott (1999), she commented:
[Students] definitely believe that they should be corrected. They believe that's the role
of the teacher. I think that's definitely true … That students do feel that way, and that’s
how I’ve always felt that students expect me to do that.

Using the example of math and science that she used to explain the fluidity of
languages, she speculated that students were actually perceiving L2 learning the same as learning
those subjects.
I think students feel that, sometimes things are much more rigid, and set in stone, but
language isn't like that at all … maybe students feel that language is like learning math
or like learning something else … we need to know what all the right answers and

144

exactly what we should say and exactly what we should not say and those kinds of
things, so, maybe that's why students feel that way.

Furthermore, citing Mackey et al. (2000), in which both ESL and IFL learners
perceived CF for lexicon the most accurately, she even stated that the effectiveness of CF for
lexicon was in line with her thoughts. She said, “lexical thing … kinda latches on to your brain
and more … solidify what you can recall, vocab, easier than you can recall … grammar rules …
that seems consistent with what I understand.” At a different point, she once again stated that CF
for lexicon was something that she expected to be effective:
[Lexical feedback] … is the thing that kind of stands out more and they're able to kind
of, grasp that um, what they've said wrong, what they should've said wrong and really
go from there as opposed to a grammatical thing like that they might later be like was
it this or was it this, I can't remember now.

As such precarious, unstable stated beliefs on CF emerged, it is perhaps not a surprise
that she still expressed her lack of confidence in CF, especially recasts, saying, “I don’t know
exactly what I should be doing [with CF],” “I was a little confused about, um, how to think about
recasts,” and “in terms of recasting, I’m not sure how I feel about that.” She went so far as to say,
“maybe my perception of recasts is not completely accurate.”
Such a tendency of her to be on the fence may have come from the fact that, despite
her limited L2 teaching experience, she was aware of individual differences of each student
when it comes down to a question of whether students would take error corrections seriously,
one of the points that Truscott (1999) claimed. She said it depended upon “[their] background

145

and motivation and personality type” and “learning style and all, all kinds of things.”
In conclusion, as discussed above, her description of herself as a “sponge” was not
completely accurate. Although she seemed ready to accept any claims that she read, in reality,
she was highly selective in what she would accept. Among the three research studies, she was
particularly influenced by many claims that Truscott (1999) made in the article, which basically
led her to negate the use of CF. This became the foundation of her “fresh” stated beliefs on CF,
which began to emerge at Interview 2. Still, she seemed confused about CF; her stated beliefs
were still in the incubating stage and thus unstable. At some points, she showed her rejection of
CF. At some points, she acknowledged her use of CF, not only because she believed that some
kinds of CF were natural but also that CF for lexicon should be effective and that students may
expect teachers to utilize CF. Also at other points, she honestly expressed her mixed feelings
toward CF.
Perhaps most importantly in this regard, reading these research studies gave her an
opportunity to reflect on her own stated beliefs and classroom practices through absorbing
knowledge from and thinking critically about the three research studies. She said, “I kept
thinking about [recasts] the whole time I was going through [the study].” She also stated that her
thoughts had been not only on CF but also on her students, saying, “I’m just keep thinking back
to my own students,” and “I just kept thinking about one student I had,” and further, she
introduced several concrete examples of her students in regards to CF. Not determined firmly yet
how to deal with CF in the classroom, her conclusion at that moment was to be more conscious
of CF while admitting it was difficult:
Try to, just be more conscious of what I'm saying and how it's thinking about how it's
gonna affect the one student and all of the students, I don't know how I'm gonna be

146

conscious of that 'cause there's so much going on when you’re teaching but … what
you say is obviously very important for any teacher … you wanna um, be conscious
and … have a purpose for the things you're saying, so … try to do that more, I'm not
sure how.

At another point during the interview, she made a similar comment focusing on
students:
I will try definitely to be more conscious of what I'm saying, that's really hard, but, I
guess that everything you say has an impact on students, certain students at least, ones
that are sitting up in the front wanting to get the most out of the class, they're gonna eat
up everything you say … Students that are motivated are gonna pay more attention, try
to get the most out of things and analyze … I would think about later why did we say
this, why not this and, so, I guess being more conscious … number one thing would be
being fair and being equal about the way you say things.

5.2 John
John read the PowerPoint file, Lyster and Saito (2010) and Truscott (1999), but did not
read Mackey et al. (2000) because he thought that it would not “[add] anything new” or “draw
any new conclusions” to what he gained from other readings. However, he remembered its
summary as it was presented in the PowerPoint file. Among the three research studies, in
contrast to Cecile who favored Truscott (1999), John favored Lyster and Saito (2010) the most,
mainly because it was a meta-analysis. Before describing his responses, it is worth noting the
dissatisfaction with research studies that he expressed at Interview 1:

147

In education, there should be a right answer … what works best in the classroom,
There's a right answer. Right? There should be. We should be able to work toward
getting a right answer … My problem with research is that they tend to be
contradictory to things that should be non-contradictory. You should be able to come
to a conclusion that says this will make you a better teacher. This style will be optimal
in the classroom.

For both Lyster and Saito (2010) and Truscott (1999), he showed his interest and
agreement. But faced with these two contrasting research studies, instead of expressing his
discontent or distress as shown in the quote above, he easily chose his standpoint, and expressed
his disagreement with Truscott (1999). Throughout Interview 2, he stated a few reasons what he
favored and did not favor in each of them, which will be described in the following paragraphs.
He seemed to have found Lyster and Saito (2010) the most convincing mainly because
it was a meta-analysis, not a single experimental study. He repeatedly mentioned this point and
said, “the meta-analysis had a lot more data behind it, and the numbers made … that argument
stronger for me,” “it combined a whole lot of different studies, that have been done on feedback,
and they cited the similarities and their conclusions and I think that makes a powerful argument,”
and “it’s hard to deny when you look at the, specially the meta-analysis of all those different
studies, the evidence seems to me pretty clearly in favor of the usefulness at some level of, of
feedback.” He restated this opinion in contrast to a single experimental study:
I’m a skeptic by nature … if there is one article, that was … based on one study that
they do it by themselves, then I could find holes in it … and say well, what about this,
what about this factor but, they analyzed objectively those 15, and it’s fairly strong in

148

my mind, it’s fairly strong intellectual proof.

At the same time, like Cecile, John showed his dissatisfaction with the lack of detailed
information of those 15 studies. Although this problem is unavoidable in a meta-analysis due to
the limitation in the page numbers in academic journals, it seems to have frustrated John. He
stated:
I like to know, uh, about how much what they gave the teacher beforehand, um, where
a lot of these studies just called to they come in like you came in to my class, you can
say and they videotaped it, and they gone, they took data from that, or did they say
here's a lesson plan. We like you to do this, and we're gonna tape you on it. I guess I
like to know that 'cause it does seem to be more unnatural if you kind of massage the
material.

His curiosity to know how the data collection had been conducted can also be seen in
the quote below:
I couldn't find that information on these studies. Um, exactly I'm, what did they just
observe cold or, because I think if you give, 'cause there are some, there were hints like
I said the other quote, in this quote, there are hints that, uh, maybe they gave some,
there was a lesson, a particular lesson, maybe they didn't dictate, you're gonna teach it
this way but, ok. You are going to do a grammatical lesson of modals today … I don't
know if they get material or whatever.

This paucity of information on methodology in the 15 studies included made him

149

wonder to what extent they were conducted in the natural classroom setting, which may have
been a crucial issue when evaluating their applicability to the real classrooms. Pointing at a
sentence that he highlighted in Lyster and Saito (2010), which states, “CF in classroom settings
may be more effective when its delivery is more pedagogically orientated (i.e., prompts) than
conversationally orientated (i.e., recasts)” (p. 290), he showed his agreement with this sentence,
and added an explanation:
They say here most of the studies that they reference they set up certain situations that
would lend themselves to feedback. And I think when you do that, you, you separate
the element of the natural classroom flow. And of course to me this makes a lot of
sense. Of course when you pick as a researcher, here is the topic we want you to talk
about, Ah, you can pick one that lends itself to, to oral correction, and I think that's
something that, that caught my eyes and well, but how about in just the daily flow of a
speaking listening class when you're not talking about grammar, you're not talking
about the vocabulary, you do it in rhythm that, that recasts and rhythm … what, what is
that due to the effectiveness of it … 'Cause they say they're very honest about that,
They talked about that … a lot of these studies were set up to elicit this feedback. And
they brought up the good point, and I think it's a good point to think about, just
intellectually that what does that due, does that a, affect the effectiveness, probably
make it more effective in my mind. Than just random one off class you pull (up) the
teacher out from the … from whatever, and you watch the class and is it as effective in
that, the flow of that class, when you don't have any control over lesson or, or material.

In regards to his own classrooms, he was implying that recasts may not be as effective

150

as shown in these studies because in his class (a) communication was one of the primary foci, (b)
there was no emphasis on grammar or vocabulary, and (c) CF, particularly in this case, recasts,
was given in the flow of natural conversation.
As briefly mentioned in the quotation above, he was also wondering what the focus of
the lesson was in each of those 15 studies, which he thought would influence the use and
effectiveness of CF. Asked what questions he had while reading Lyster and Saito (2010), he
wondered whether, in the study, “there have been a study just focusing on a grammar and then
maybe only focusing on vocabulary correction … that would be interesting to see if there's a
difference in effectiveness (?) to, different categories.”
He also speculated that CF on grammar in a grammar lesson would be more effective
than CF on vocabulary in a vocabulary lesson in his speaking and listening class:
What if, in my class in here vocabulary were I say, oh, you m, you, oh, yeah. And then
I correct vocabulary word and we kinda continue on … I think the student isn't focus
directly on vocabulary at that moment, so I wonder how much they pick it up versus if
we're studying a grammar point, and then there is an opportunity and there is a
mistake, I just say, oh, you mean, blah … their awareness already peaked, we're
already studying grammar, so, boom. I think that naturally to me seems like it would
be more effective than the first situation.

As can been seen in his view on Lyster and Saito (2010) described above, John was
applying his critical thinking skills, which he deemed one of the most important components in
his teaching, in evaluating the study, and in addition reading the study from a researchers’
perspective. Such a viewpoint is in stark contrast to that of Cecile, who favored Truscott (1999)

151

from a teachers’ perspective. On the positive side, he stated that the results of Lyster and Saito
(2010) were convincing because (a) it was a meta-analysis rather than a single study, (b) it had a
lot of numbers and data, (c) it was objective, and (d) it provided “strong intellectual proof.”
Concurrently, on the negative side, he criticized this study for two points: (a) it lacked the
detailed information on the 15 studies included, and (b) the results of those 15 studies might not
be applicable to the real classroom because their classrooms were especially set up for the
experiments in order to elicit errors so that CF could be provided, and therefore were not natural
settings. This is an interesting finding because at Interview 1, he stated that he had not put a lot
of faith in research studies since he was in the MA program, similar to the teachers in S. Borg
(2009a) and Shkedi (1998), owing to their lack of practicality in teaching a real class, which
made him read only practical studies. Nevertheless, somewhat surprisingly, when it came down
to reading research studies, he evaluated them from a researchers’ perspective, partly due to his
knowledge and skills of critical thinking skills that he had gained.
For Truscott (1999), John expressed a few reasons for his interest, although the article
was not necessarily in agreement with his ideas:
What did I purely like the most the Truscott article was the most entertaining. Because
it brought up a lot of good points and it wasn't as based in the, in research, so is easier
just to read through and I found myself thinking yeah. Yeah. Yeah. Yeah. Yeah. I
understand what you're saying …I enjoyed the Truscott article the most as far as just
reading.

He also stated that the issues that Truscott (1999) brought up were good points for
argument. He stated, “he makes valid points that are worth thinking about,” “logically, I thought

152

he brought up some good points … I just think he brought up some good points that are logical
to me,” and “he brings up a lot of natural questions that I think, come up from … the research of,
of feedback and how it helps the students and not.”
On the other hand, he called Truscott (1999) “the outlier,” and claimed that many of
his points were actually what would naturally happen in the classroom and did not necessarily
negate the effectiveness of CF. He stated:
Most of 'em I found specially from the teacher perspective that I, I found that besides
the being consistent point, I felt that the other ones ([e.g., it is difficult for teachers to
understand the nature of errors correctly]) were a natural part of what a teacher does.
And not necessarily indicative of uh, in, inability to provide feedback.

Furthermore, he supported his view by the fact that the article was lacking the evidence
to corroborate Truscott’s arguments. He stated, “I don't think that he really proves anything in
this article.” Asked why he thought Truscott (1999) did not prove anything, he added:
He brings up a lot of data, but the data is, is very scattered and not real, he, he
references whole bunch of different articles, but never anything that really proved to
me that what he saying would suggest that providing feedback is useless. I guess that's
what I mean by he didn't prove anything as far as I'm concerned.

He also mentioned the difficulty of experimentally testing some of the points that
Truscott (1999) made:
I think that the, the questions that Truscott brought up … from the student perspective
a couple of them, still they weren't clearly and completely addressed by [Mackey et al.

153

(2000)]. I felt … frankly … when I read that … I don't know how we could prove
[Truscott (1999)].

John’s comment below in relation to the impracticality of proving Truscott (1999)’s
claims shows that he was well aware of the intricacies of L2 learning.
Some of his concerns. I don't think it, there, it would be very complex to try to even, I
don't even know how you would study that. And, and separate it from other issues that
were going on. I don't know how would that happen. And, so therefore … I don't hold
the against this article or that article or anything that they can't disprove discretely
every one of Truscott's points … but they're just not, for various reasons, able to …
just say Truscott is wrong, here the data of.

Despite the fact that Truscott (1999) was mainly concerned about teaching, John’s
critique of this article was again presented from a researchers’ perspective as was the case with
Lyster and Saito (2010) in the sense that he stated that (a) this study provided some ideas for
future research about CF, but (b) Truscott himself did not prove anything in his article, and (c)
Truscott’s arguments could not be empirically tested.
It should be remembered that at Interview 1, despite his firm and consistent teaching
philosophy, he did not specifically state his beliefs on CF, stating that he came not to think of it
as much as he used to, but just hoped that he gave CF “naturally,” (i.e., automatically), especially
when there was a communication breakdown in his classroom. This, however, did not mean that
he did not have any beliefs on CF like Cecile. Rather, the reality seemed to be that he actually
had had clear beliefs on CF even prior to the present study. However, because most of them were

154

already implicit due to his extensive L2 teaching experience, he was not able to articulate them
in a short interview. At Interview 2, however, he came to express his opinions on CF clearly in
the form of agreement with Lyster and Saito (2010) and more strongly in both agreement and
disagreement with Truscott (1999). Reading those research studies possibly helped him to better
articulate his stated beliefs through facilitating his reflection on his own beliefs as well as
classroom practices of CF and also from being given a chance to express them at Interview 2.
One of his stated beliefs on CF was that CF was effective, and thus should be provided.
He highlighted a sentence in Lyster and Saito (2010), which said, “it is effective to employ CF in
response to students’ nontargetlike production because it contributes to target language
development over time.” Supporting this notion, he oftentimes expressed his beliefs in CF,
saying, “I really thought that even if I am providing feedback that isn't perfectly match to a
learner's … acquisition ability or timing or whatever, it's better than not doing it,” and “when
students make … a grammatical mistake but not about the thing that we're focused on, and then
… giving some correction to that, because their mind is already there, I think, [CF] could be
useful.”
When further asked why he thought that CF should be effective, he said, “’cause it’s
logical,” and then added:
When mistakes happen, that it's a good learning moment because you don't often have
that on a test, a mistake happens because in the context of what you wrote, they chose
the wrong answer. But in the classroom where there's a natural flow … given take of
that teacher student dynamic, and then the student makes a mistake … the student
started that production. It make sense that, if I interject and say ok. Within your context
student, here is what you said wrong, here and here what is correct. That to me seems a

155

lot more natural than like, the one off assessment measure that I create for them. So, I
guess just logically to me, it seems like it makes sense to me intellectually, and in here
… that it would be helpful.

Although he did not explicitly state at Interview 1 that CF would be effective, judging
from his comment, “I’m hopeful that still happens,” it can be assumed that he at least believed in
its effectiveness to some extent. Thus, he was corroborating his beliefs prior to reading with a
claim by Lyster and Saito (2010). In regards to his classroom practices of CF, since Interview 1
he had emphasized that his use of CF was “natural,” as also seen in the quote above. As
discussed in Chapter 4, for him “naturalness” meant “automatic and unconscious;” thus, it was
natural for an experienced teacher to give CF. In contrast, Cecile’s “naturalness” meant “a part of
a natural flow of conversation;” thus, it was natural for a person to use recasts and clarification
requests. He mentioned this “naturalness” in other parts of Interview 2 as well, saying, “it's not in
the forefront of my mind … I'm hopeful that this is something that happens naturally during the
class at this point,” and “my hope is that some of [CF] is natural now.” Nonetheless, it seems that
his consciousness of CF had been raised by reading those three research studies and expressing
his thoughts on CF at Interview 2. In fact, the quote above seems to indicate that he believed in
the use of CF by “interjecting” interactions. It should be remembered that at Interview 1, he
stated his belief that even the explicit types of CF would be beneficial if it was done in a
“correct” way; therefore, he found pieces of information that could support his stated beliefs as
these three research studies gave him frameworks for expressing them.
As another reason why he gave CF in his classroom, he stated that CF for a learner
may also be beneficial for other learners in the view of (a) L2 acquisition, and (b)

156

communication breakdown. For the former point, citing one of the points that Truscott (1999)
made, which was “an error correction for a student may not be appropriate for other students,” he
stated, “I think that that's where that point is jumping and I think that's a pretty large jump.” He
further said:
If you have a level, you have some general … the general weaknesses and strengths of
those students. And of course there's variable. But, I would be shocked if one student
made an error, and if I found that it was worth to stop and correct that the majority the
students in that class wouldn't often have that same error at some point. That's why I, I
think that arguments (like) doesn't hold lot of water.

For the latter point, after he said that he thought that he used the implicit types of CF
unconsciously especially in his speaking and listening class, he stated that he gave recasts for
other learners to understand the erroneous utterance correctly:
I think sometimes that feedback is also for the rest of the class, oftentimes at that
moment, because someone says something. And … If I don't clearly understand what
they said, the class didn't and, to continue this discussion, everybody needs to know.

This is in line with one of his stated beliefs on CF that he gave CF when the error
hindered understanding and impeded the flow of communication. This implies that when
communication breakdown happened, his focus was not so much on correction, but rather on
clarification. In other words, CF was used in a way to support communication; therefore, he did
not give CF as long as the meaning was clearly conveyed:
During a normal class, we talk about comprehension, getting me ideas of we talk about

157

critical thinking, I tend to let grammatical things go by the side, go by the wayside …
in the flow of a lesson like that, I tend to weight the meaning more in the flow of that
lesson.

Using an example of the article system in English, he said that he would not correct the
errors in articles, and its reason was that it would not cause a communication breakdown:
I wouldn't choose to correct implicitly or explicitly a lot of those areas that, unless they
impede meaning … But articles, they rarely impede meaning and I think … I probably
do let most of those mistakes go.

As described in Chapter 4, he had no hesitation to give CF as long as it was given in a
“correct” way, which meant that it would not humiliate students so that it would not violate one
of his teaching philosophies to create a comfortable environment for his students. Nevertheless,
he did not believe that CF was always effective, and stated that the effectiveness of CF depended
upon various variables, being aware of individual factors, contextual factors, and practical
difficulty in using CF in the classroom.
For example, he thought that the type of error was one of the factors. Whereas he
believed in CF for lexicon, as seen in his comments such as, “students just innately … they
always want more vocabulary, that's always a desire they want more academically orientated
vocabulary, I just think that's something they would immediately latch on to,” he expressed his
doubts about the effectiveness of CF for pronunciation:
Pronunciation … that's a whole another field of research that is … there's lots of
different opinions. Can we explicitly teach them, the t-h sound. Um, we can tell them

158

mechanically how do you do it, We can put and tell them where the tongue is held and
stuff, but … are they able to actually get better because of that … I haven't found a
good answer to that one.

As another factor, he cited the relationship between the difficulty of form and the
proficiency level of the student. He stated that, although he would not correct errors in articles as
described above, he would correct errors in tenses in the current class because “[the students]
should have a, uh, hold on and their productive ability and most of the tenses.”
Some of his comments that CF depended upon the lesson were already quoted above.
In addition, as he said when evaluating the 15 studies in Lyster and Saito (2010), he adduced
such other factors as “how [students] present ideas,” “the lesson material that you’re doing,” and
“how [the material] lends itself to correction.” He also mentioned the focus of the lesson played
a role in it, He stated:
When I'm doing a speaking listening lesson and I'm trying to get comprehension, uh,
and they make a slight grammar error where they leave out an article or something … I
go right by that … I don't care. Because the goal of my lesson isn't about that.
Necessarily. And not that it has to be, but uh error like that, I will let go because that's
the, I don't wanna their draw off a focus of what we're talking about at the moment.

He even stated that teachers may not notice their students’ errors under certain
circumstances, adducing an example of himself with critical thinking:
If I'm concerned about the content, we're having a discussion, I'm really looking at
gaining critical thinking and pulling ideas out … I might not stop and correct some a

159

grammar flaw that wasn't anything to do with what we are talking about … I might not
even notice it because we're focused on something else.

Several claims that Truscott (1999) made were derived from the practical difficulty of
CF. While admitting that they were representing a true picture of the classroom, John still
believed that they would not negate the effectiveness of CF. As can be seen in the quote above,
he admitted that he may not be able to notice some of the errors in the classroom. But he further
added, “but still I don’t think … that impacts whether [CF] is effective or not.” While he agreed
with a point in Truscott that “it is difficult for teachers to make error correction consistent,” John
added, “my question was … does it have to be consistent.” Finally, regarding another point,
which was “error correction facilitates explicit knowledge, but does not facilitate implicit
knowledge,” he broadened the issue to language teaching in general, and stated:
I think that could be said with a lot of language teaching. Uh, and how do we make the
explicit into the implicit, That's right … that's the struggle with any kind of instruction.
In my opinion, that's just one of the challenges of teaching. And it's practice, it's
repetition, it is correction, and I think you have to have that step in that process or else.
You don't have a shot at the implicit part.

He expressed the same opinion using an example of tests.
I agree with [Truscott (1999)] … [not] Just … ESL te, as an educator in general, I
think that of course, It's what we always struggle for. Is this test going to actually help
them, or is it just, 'cause I need to give points. I hope that it's … mostly that it's going
to help them, and of course, I need to have points too to assess, but I want it to be

160

valid.

In this sense, his critique of Truscott (1999) was also conducted from a teachers’
perspective. Therefore, he was reading these three research studies as both a researcher and a
teacher in contrast to Cecile, who only had the latter. This is possibly derived from the difference
in the amount of training in reading research studies in the MA program that each of them had
received; Cecile was still in an incipient state whereas John had finished it. Also, it is a
legitimate choice of him to use both of these perspectives considering the main target of readers
in each of these two studies: researchers for Lyster and Saito (2010) and teachers for Truscott
(1999).
In summary, among the three research studies, although the lack of information was an
issue, John was in favor of Lyster and Saito (2010), mainly because this was a meta-analysis.
This may explain why Mackey et al. (2000) was simply disregarded because it was a single
experimental study. Although he found Truscott (1999) amusing and interesting, he found both
agreement and disagreement in this article. However, this was actually the one that he had the
most to say about. This may be simply due to the fact that Truscott listed many discrete points
for argument. But also, it may be due to the fact that it was easy to read for John and Truscott
brought up many issues that were thought provoking for L2 teachers. Regardless of whether he
reacted positively or negatively to the article, Truscott was the article that facilitated his
reflection on his use of CF the most. In this sense, Truscott was the most influential study among
the three.
As mentioned above, John did not explicitly state his beliefs on CF at Interview 1
except a few points, but talked of them extensively at Interview 2. Taking into consideration his

161

14 years of English teaching experience and also his claim that he used to think of CF a lot more
before indicates a stronger possibility that he had already established his beliefs on CF at the
time of Interview 1, but most of them had been unconscious, which prevented him from
articulating them. In this sense, reading these articles and expressing his thoughts at Interview 2
succeeded in raising his awareness of thoughts on CF as well as eliciting them, both of which
would have been difficult without these opportunities, which Ellis (1997) claimed as one of the
benefits of reading research studies for L2 teachers. For example, at Interview 2, he said, “as I
read through this, I was thinking about [my use of CF] … I brought a grammar [textbook] today
because we started on a grammar chapter … I did find myself doing this much more in the
function of grammar.” Also, he stated, “I've been thinking about what my natural way is. And
I'm not sure if I know the answer to that, like it's hard for me to pull myself out objectively and
think about what I just did an hour ago.” Using these references made it easy for him to articulate
his stated beliefs more clearly than at the time of Interview 1 in which his stated beliefs were
elicited without any stimulus.
Finally, he mentioned two possible effects that the reading had on him in regards to his
classroom practices of CF. The first point was that he would be more conscious of CF, a point
mentioned also by Cecile. He stated, “I would like to think that [reading these studies] will make
me more aware of those opportunities [of CF], uh specially like I said going to class today … a
grammatical lesson, really there's opportunity here, for that … lot of it … when you speak about
grammar, it becomes explicit correction, uh, but certainly there are opportunities for … more
implicit [CF].”
The second point was that he would “like to give and be able to use a variety of ways
just to make sure that their exposure isn't only one style of feedback, it's not always explicit, for

162

example.“ This may be derived from his awareness that the effectiveness of CF depended upon
many variables as described above in addition to his hope that his students “expect that there will
be correction” in his class.

5.3 Jim
Jim forgot to read the PowerPoint file, but because he read all of the three research
studies thoroughly, this did not impede Interview 2 in any way. Overall, he showed his general
interest in and agreement with each of the three research studies to a certain degree, saying, “I
thought that they were all very interesting;” however, whereas Jim seemed to appreciate Lyster
and Saito (2010) and Mackey et al. (2000), he rejected the usefulness of Truscott (1999). This is
somewhat similar to John’s standpoint; it is, however, different in the sense that (a) although Jim
did not talk about Mackey et al. extensively, he showed his favor of the article, and (b) his
evaluation of Truscott (1999) was more severe than that of John.
For Lyster and Saito (2010), he stated that he did not have any points to disagree with.
He said:
There wasn't all that much to disagree with them there, I think they make a point …that
the effects were stronger in laboratory situations than classroom situations and yes, that
doesn't surprise me … their categorizations of … explicit, implicit of um, recasting,
prompts, etc., I could see how interactions would fall into those categories.

While stating the same opinion, he called this study a literature overview.
Since this was essentially a literature overview, um, I just have to say that I don't find
any of this really … all that controversial … what they said about… the effects of age

163

and … instructional setting and all that kind of stuff, once again, very easy for me to
believe.

The mere fact that the study included 15 different studies was enlightening to him,
which interested him. He said, “I'd say I probably learned the most from the meta-analysis. A
because that … introduced me to literature that I hadn't read before and that … was nice to, to get
that perspective.”
However, he confessed that he skipped the statistics section. Later, it was found that he
was confused between the correlation coefficient and effect size, and asked me to clarify their
difference. Lyster and Saito (2010) explained the formula in the article, but it was probably not
enough for him to understand what the effect size was, with which anyone who is unfamiliar
with statistical analysis, including L2 teachers, are generally less conversant than with
correlation coefficients.
Although Mackey et al. (2000) was a single experimental study, Jim also showed his
interest in it from a different perspective, saying, “I thought that [Mackey et al.] was the one that
was actually, oh covering new territory in research, it was actually moving the research
boundaries a little bit further, and so I appreciate that.” Still, he pointed out its small sample size,
which was 27 in total, as its weak point. Referring to the table showing the percentages of CF
perceived correctly by the L2 learners, he stated, “none of these percentages really … impressive
to me, because the sample size was so small … I don't think that the percentages are significant
to me and maybe statistically significant.” Therefore, like John, Jim was reading Mackey et al.
(2000) from a researchers’ perspective. This might have been derived from the fact that he had
been trained to do so in the MA program and that he regularly read research studies on

164

vocabulary.
Finally, for Truscott (1999), Jim “was able to react most strongly” “because … it
wasn't a matter of any research that he was … more capable of talking about them … you and I
can both uh, look at that stuff and give our opinions.” Jim also said that what Truscott was
claiming was the practical difficulty in giving CF in the classroom situation, of which Jim was
well aware, and such complexity was not anything that could negate the effectiveness of CF.
Therefore, to him, the article was not useful in the sense that all of those issues were obvious to
L2 teachers. He stated:
[Truscott] was complaining about all of these problems that are not, to my mind, really
problems. They're just characteristics of human interaction and … they don't really
stand in the way of the validity of error correction. Of oral error correction.

At another point during the interview, he further elaborated on this issue, stating that
what Truscott (1999) was saying was simply the difficulty of communication, but not the
ineffectiveness of CF:
I actually found um, the Truscott one to be, kind of obvious … I was reading it, and I
was thinking, well, ok. But all you’re really doing is saying stuff that, anybody already
knows about interpersonal communication … to my mind … he was as essentially
saying that … oral feedback is not effective because it introduces a communication
problem. Well of course, it introduces as communication problem but that doesn't
mean it's not effective … I thought his arguments against it … simply pointed out that
communication is difficult, it didn't really offer much of an argument … to show
ineffectiveness.

165

While elaborating the notion presented above, he made specific comments on each of
the claims that Truscott (1999) made such as “that's just stating the obvious,” “doesn't mean that
[CF]'s ineffective,” “that's a trivial complaint,” “it's just natural,” “that doesn't invalidate the
effort … doesn't make the effort worthless,” and “just a natural aspect of communication, it
doesn't have anything to do with the effectiveness or ineffectiveness [of CF].” He also made
similar comments from the teachers’ viewpoint: “that doesn't mean you shouldn't do [CF],” “that
doesn't remove your obligation to provide [CF],” “that's part of what, not only ESL teachers or
not only language teachers, but every teacher has, it's part of our skills,” and “that's part of what
goes into becoming a professional teacher … that's part of our job is to get good at that.”
It should be remembered that, to Cecile, most of these claims made by Truscott (1999)
were new to her, things that she had never thought of before. However, to Jim, they were
obvious points, and he mentioned that every L2 teacher should know them. This is probably due
to the differing amount of L2 teaching experience between Cecile and Jim. Cecile was still
relatively inexperienced; therefore, she was not aware of these points. In contrast, Jim had
already become aware of them through his lengthy L2 teaching career.
The fact that he favored Lyser and Saito (2010), but not Truscott (1999) implies that
like Cecile and John, he was selective about which article to identify with. In addition, his choice
was far from random; he chose the study that most closely coincided with his stated beliefs on
CF prior to reading. This is one reason why he chose Lyser and Saito (2010) as this study stated
that CF should be provided, which he stated that he had been doing by giving the correct answer
directly.
Toward the end of discussion on Truscott (1999), he speculated that such discrepancies

166

in perspectives on CF between him and Truscott may be ascribed to the differences in the
relationship between teachers and students derived from the unique L2 classroom contexts where
CF occurred. He stated:
Maybe in my head, I've got a different picture of teacher student communication than
Truscott had when he was thinking of it. Know I'm thinking about a class in which,
well, like our level [X] section. We know their names, we've got relatively good
relations with 'em … we're with them for 13 weeks, that there's kind of a personal,
there's a personal interactional um, situation set up that maybe doesn't occur in a lot of
other foreign language classes, or that doesn't occur under research circumstances
because there's such a fleeting relationship between ah, the teacher and the student, I
keep referring to what seems natural … but that's being colored by the fact that there is
this personal relationship between me and the student. It wouldn't seem so unnatural if
it were a student, or if it were a non-native speaker who I was just talking to … maybe
I see 'em for one day or for one week or something like that. There it would seem …
more natural for me to sometimes maybe not give them as much as I give them in my
class.

Asked to further explain, Jim speculated that the reason why Truscott (1999) would not
prefer to give CF in contrast to Jim could be that the relationship between the teachers and their
students that Truscott (1999) imagined may have been different from that of Jim:
I think it's because of the enduring nature of the relationship that um, all those things
that Truscott brought up about um, the problems of introducing error correction, um, I
think that they, they’re more of a problem if, if your communication is short-term,

167

directed at very well-defined individual targets that you have to reach in a very very
short time. And it's not so much of matter of, I know you, [student's name], you know
me, we've talked many times before, when you say something to me, I pay attention to
it, I have some idea of your personality, I know how much you can understand and
how much you can't understand … all of those factors that negate Truscott's
complaints, exist in my class, but they may not exist in other classes.

Regarding this issue, he concluded, “all of my statements that well it would just seem
unnatural to not correct somebody. All of those statements are based on a classroom where
there's a lot of teacher student understanding.” This clearly represents his stated beliefs on CF. At
Interview 1, he stated that, when an error occurred, he would prefer to provide his students with
the correct answer possibly with some explanation, but did not elaborate on this issue. At
Interview 2, however, he further explained his reasons why he would do so. One reason is that
giving CF was a “natural” part of instruction in the classroom, especially when students expected
CF from teachers. He stated, “it would seem … weird to have a teacher and a student, conversing
and the teacher unnaturally avoiding correction.” Also, he said:
To me, it's less natural to not to say anything. When, when you've got the time …
when you're in instructional situation where the student clearly looks at you as the
arbiter of correct language … there's no time reason, there's no attention reason, there's
no interpersonal reason not to say something. It seems, only natural that you would say
something. That's characteristic of your relationship that you are not the student’s
period or not the students teacher. And the student expects certain things from you that
it's unnatural of you not to deliver … granted some students want you to correct

168

everything … if the student wants me to correct major things, and we've got the time,
it's my job.

He restated the same point while adducing an example of the UK.
There are times, of course, where you'll focus on a morphosyntactic feature and you'll,
you'll say, no, no … I would say THE United Kingdom. Not just United Kingdom.
Because, that's one of the countries that has THE in its name. To me, it seems totally,
totally natural.

This notion was partly based on another stated belief of his that CF was a natural part
of normal conversation even between native speakers of English, stating, “even when I’m talking
with native speakers of English … we do what [Truscott (1999)] would call an ineffective
correction process, but it’s just really clarification. And it’s a natural part of communication.”
Then, he adduced an example of Ukraine:
You often hear native speakers refer to THE Ukraine. Well, they shouldn't say THE
Ukraine any more because it's an independent country now. But it's not just a region,
and it's very, would be very natural for me to hear somebody say oh, I hear she is from
THE Ukraine. I would say, no, not THE Ukraine. Ukraine. They don't wanna be called
THE Ukraine any more. There's nothing unnatural about that.

He later used another example of an upcoming election.
Let's say I'm talking to my wife. She says, something like … I don’t think um, Mitt
Romney has, has any chance … in the New Hampshire caucus. And I say, the New

169

Hampshire caucus? She says yeah. You know where they, where they try to sort out
the, the nominee for the Republican Party. The New Hampshire, Oh, you mean the
New Hampshire primary. And … it's like a natural, people make mistakes, and when
other people respond to 'em, you have what sounds like a correction situation, which is
really just a clarification situation.

Like Cecile and John, Jim also used the word “natural” as can be seen in the few
quotes above. It should be remembered that, although they used the word commonly, Cecile and
John used it with a different connotation: “a natural part of conversation” for Cecile and
“automatic and unconscious” for John. Jim’s “naturalness” seemed to represent twofold; one is
the same as that of Cecile, which was that CF was a natural part of conversation, and the other
one is unique to him, which was that CF was “a part of job as an ESL teacher” as it was “natural”
in classroom discourse. It should be noticed that he said, “if the student wants me to correct
major things, and we've got the time, it's my job” in one of the quotes above.
Focusing on recasts, he thought that they were “very common,” and he acknowledged
their efficiency. He stated, “people appreciate recasts in some ways especially when you're
talking about … pure morphosyntactic or in some cases … very discrete lexical issues. Recast is
very efficient and it gets things done.” However, when the topic shifted to which type of CF he
thought would be more effective, he expressed his strong stated belief on prompts, saying, “the
prompt, to me is, more likely to be remembered because it's, it's really communicative. It's, it's a
deeper phenomenon. It's not so much of surface phenomenon.” Accordingly, he stated that he
would prefer prompts to recasts in his classroom while saying that he used the former more than
the latter:

170

If you're having a real conversation um, that, prompts are actually … more like
conversation and I think that, I am more likely to prompt than to recast, just the way I,
I talk in class. To me, recasting … doesn't sound like conversation, it sounds kind of
patronizing in a way like, I feel like there's a greater distance between me and the
person I'm talking to when I recast … that distance narrows when I prompt. It's more
like, um, an even communication situation.

At the same time, he thought that explicit correction may not be the best type of CF in
terms of effectiveness, saying, “I think that explicit, explicit correction, even though it might
seem to be more directed and more … incisive, it … doesn't necessarily uh stay with the learner
more than a prompt or even a recast.”
All of these stated beliefs on specific types of CF were in line with the results of Lyster
and Saito (2010), which found that (a) CF was effective to some extent regardless of the type of
CF provided, (b) prompts were more effective than recasts, and (c) the effectiveness of explicit
correction was not significantly different from that of prompts and recasts. This shows that Jim
had no disagreement with this article.
In the meantime, however, he was aware that the actual implementation of CF
depended upon many variables, such as the level of learners:
If I hear my student make a mistake in um, a future hypothetical conditional, and that
student is a level 2 student, I'm not gonna correct him about it … It's part of our skills
that … you learn what's valuable to say and what's not valuable to say.

Also, as described in the previous chapter, although he was “not the kind of person

171

who gives a lot of oral feedback,” he said that it may vary depending upon the type of lesson,
saying, “I give feedback more in formal speech um, situations.” Also he stated that he would
tend to give more CF in his reading class than in his speaking and listening class:
I'm more inclined … to do error correction with reading because you can go back and
you can find evidence in the reading … you can't really go back, over the spoken
interaction all that much and so, I guess I'm more tempted to do it in a reading class.

Furthermore, he stated that he would expect to use CF extensively in his grammar
class:
I will do a lot of error correction in grammar because the point at least as we've taught
it and as our books teach it … it's a very form focused, kind of an approach … I think
that that's what we're supposed to do if we’re focusing on form we're supposed to teach
form … I do error correction in grammar.

However, he mentioned that giving too much CF would break communication. Thus,
the amount of CF was another factor to take into consideration, saying, “if you keep stopping and
stopping and stopping and stopping to the point where the conversation never gets anywhere, I
can see that a, a classroom correction situation might become more frustrating.”
In summary, among the three research studies, he found Lyster and Saito (2010) the
most convincing, mainly owing to the fact that it was a meta-analysis. The fact that it consisted
of the results extracted from 15 different studies gave him the impression that the study was
cogent. In addition, the findings in Lyster and Saito were congruous to his stated beliefs, which
is probably another reason why he particularly favored this study. In this sense, like Cecile and

172

John, Jim’s prior stated beliefs on CF influenced his way of reading these three research studies.
He did not mention much about Mackey et al. (2000). This may be because Jim knew one of the
authors of this study in person. He stated, “I’m not in any position to criticize … [the name of the
author]’s research techniques.” This may have created a situation where he refrained from saying
anything negative about this study. Finally, for Truscott (1999), he cast doubt on its usefulness as
all the claims that Truscott made were obvious and natural to L2 teachers. Jim was well versed in
the practical difficulty in teaching, but expressed that it had nothing to do with the effectiveness
of CF, which was in contrast to Cecile’s impression of Truscott that the study was enlightening.
When he was asked how much these research studies would influence his use of CF in
his classroom, he basically said that there was only a slight likelihood, and showed his
confidence in its current CF practices, saying, “I'm still pretty confident that the level of
feedback I give is an appropriate level.” Also he stated:
I don't think that it will increase, increase or decrease the amount of oral correction I
do … I don't really feel differently about it than I did before I read the articles … I still
feel like, my level of oral correction is a suitable level and, I probably will continue to
do about that much.

The only impact that he stated was that he became probably more aware of CF than
before through reading, a point mentioned also by Cecile and John. For the same question, he
replied, “maybe it will heighten my awareness,” and “[reading] systematize[d] my thought …
maybe I'm a little bit more, aware of kind of a taxonomy of um, of oral feedback.” Also he
added:
I think that, primarily it's gonna make me more conscious of what I do in the oral

173

correction. I'm sure I'll catch myself doing oral correction at some point and then
thinking about these articles and thinking how it fits into the, the scheme, that the
schema they've laid out.

He further adduced an example in which he stated that he may reflect on his use of CF
posterior to teaching a class:
I probably won't think about it right as I'm doing it, but I think that maybe after class,
when I think back on what I did, I’ll realize that, oh yeah. That was that kind of oral
correction. That was a recast and that was prompt. Um, and maybe I'll think a little
more specifically about how effective it seemed to be.

Finally, in line with his preference for prompts over recasts, he stated, “if I catch
myself recasting too much, I might … say don't do that. Don't do that. Move to prompts.”
In conclusion, it did not seem that reading these research studies had an influence on
his stated beliefs on CF. Rather, he was able to find some evidence to support his stated beliefs
that he had had prior to reading, and his confidence in his implementation of CF was stable.
Nevertheless, it could be said that the reading did have an impact on him in two points: (a) his
heightened awareness of CF, and (b) his increased knowledge about the taxonomy of CF.

5.4 Tom
Tom read the PowerPoint file as well as the three research studies. Overall, his
responses can be summarized as general agreement with Truscott (1999) and disagreement with
Lyster and Saito (2010) and Mackey et al. (2000). This is somewhat similar to Cecile; however,

174

his argument was much stronger than hers.
To begin with, at Interview 1, he expressed his willingness to learn new things, even
from his students and novice teachers. He stated:
I believe that I can learn from anybody … because I have thirty years of experience …
why am I gonna listen to these people, they just got out of graduate school, That's
ridiculous. They know stuff that I don't know. They learn things in graduate school …
and from their own life experience … They’ve just gotten through with all those latest
stuff, I should sit listen 'cause I might learn something. So I learn from my students. I
just wanna my beliefs that I can learn from anybody.

In accordance to his motivation to learn, he said that he was grateful for being given an
opportunity to read the three research studies, saying that if he did not participate in the present
study, “I probably wouldn’t read them otherwise, so, see this is one way I motivate myself to do
these kind of things.” He regarded it as a chance to learn new things. He reflected:
[Reading] makes me think again about why I do things and where, why I don't do
things, every time I reread something, even though I've known about it, doesn't mean I
use it all the time. So now, ok. Let me try this out again. Let me see how this works,
let's see what this, so, actually doing that, it's helping me relearn things, and reevaluate,
and I think that's the most useful thing about reading this ok, When I was younger,
reading … this would've been enlightening because I didn't know anything. Thirty
years of teaching I know many things. But I'm still interested in continuing to know.

Tom himself had published two articles in a journal, so he was familiar with research

175

and also, was aware of difficulty in conducting research. In this regard, he commended Lyster
and Saito (2010), calling it a “credible job.” He said, “they did uh commendable job in a very
difficult topic. I could’ve never done that. So I’m not gonna criticize them for … that.”
Although he said that the study was “interesting,” his responses to its content were
mostly negative. He stated that one reason was its overall “really complex” style of writing,
which was, according to him, a standard for publications in academic journals. He complained
that this made the study “difficulty to read” and understand. He said:
My biggest complaint is [it’s] badly written. I understand why. They need to get it
published … their audience were the, people reviewing it and if you wanna be taken
seriously … if you wanna get published in that journal, you better write badly. In this
style … But it's very confusing in that sense … they give all this, age factors and
present study, then they talk about different thing … I'm trying to figure out which one
is going where and what, It was very confusing in that sense.

He further stated that such a “bad” style of writing was the norm in academic journals:
They say a whole a lot, they use a whole lot of words to say very little. It’s bad writing.
Which tends to be true for almost all academic articles. Because, that’s the style. You
must write badly in order to get published.

He also showed his discontent with its long reference list:
If you wanna get published in these kind of serious journals, you have to have a
bibliography this like almost as long as your paper. And that means you have to insert
that everywhere, which means that's a quarter of your paper is all insertions and they're

176

very annoying. They serve useful purpose, but they're very annoying.

In addition, referring to the concept of a “good teacher” who can reach the 70% of the
students, Tom criticized the nature of meta-analyses, which was to deal only with numbers, but
not human beings directly. This is in stark contrast to John’s claim that it was these numbers that
made Lyster and Saito (2010) convincing to him:
If you reach 70% of the students, and they do well, you can show that your class got
better. But 30% of the students are lost and they're gone. That's a lousy class. As far as
I'm concerned. But these studies don't show any of that because they're only looking at
numbers, and not looking at people. And that's the problem with quantitative studies.
They look at numbers and not people … not that … you shouldn't do them, they, they
need to be done and they have some uses and so forth, but you have to take that all
with a grain of salt. People get wrapped up in numbers and they think that numbers are
like magical, because I've done all these things and I do all this math and forth better
than something that isn't done that way and, just isn't true.

This continued to the criticism on the use of statistics. First, he admitted that he was
not familiar with statistics, saying, “I have a very basic understanding of statistics, so, I have no
idea whether the statistics they used were the right ones to use for the article or not, So I had to
take that on faith,” and ”I have no idea what any of this means, data computation, I have to take
that as, I have to go to a statistician and ask them … ‘cause I don't know.” Specifically, he
honestly expressed his unfamiliarity with k samples by saying, “I don’t know what k samples
are.” However, at the same time, he understood that the results of statistics could be manipulated

177

in order to fit the researchers’ expected outcome. He said, “I do know that it makes a difference,
what you use and if you use something wro, wrong, it can skew the things.” Furthermore,
adducing a famous quote, he emphasized the trick that statistics could play:
There're three kinds of lies. They're lies, damned lies, and statistics … how you choose
and what you choose and what, what formulas you use and so forth … even though it's
math also in art in the sense you have to choose the right one to use in order, so you
choose the one who makes your data look the best. Human nature. No one wants to
choose it, one that's gonna make your data not look good … And both of these … are
legitimate, it's not like, you're choosing one that's completely off, they're both
legitimate, but they're gonna skew it one way or the other.

Accordingly, he cast doubts on the results in Lyster and Saito (2010) partly due to the
results of the statistics that they used. In the study, the confidence intervals of the effect size
were quite large, and oftentimes overlapped between groups, upon which he pounced. He stated,
“I don't think that they really proved their points strongly, strong enough to actually say that it's,
it's worth the effort,” and “even though it's supposedly positive on the side of it, it’s marginally
positive, even they say that, right? It's slightly? Well, if there’re so many ‘slightly,’ it's not worth
the effort.” Pointing out the two phrases, “the difference between these two measures [for the
effectiveness of CF] almost reach significance” (p. 285) and “the differential effects of [two
different measures] remain unclear” (p. 285), he said, “all these kinds of things, makes me think
that, that's not necessarily true.” Moreover, he said:
So if [the effect is] not a really strong … I tend to discount it, and this isn't really
strong, They say there is a slight, well, slight to me means meaningless. I'm not

178

interested in slight … what about the people that don't make it.

Related to this issue, the conversation went on:
T: … but, it doesn't really add a whole lot. I don't think to the, to what we know,
R: Umn, so basically, but what they say was uh, error correction was more effective
than not giving error correction.
T: Yeah. I don't think they proved that. That's, they kind of say that, and they don't say
it. Their data doesn't support that. I don't think, I'm not convinced of [the data].

Using this rationale, he went so far as to cast doubt on all of the studies that were cited
in Lyster and Saito (2010) that claimed that CF was effective:
T: … there's no real research that shows that [CF] actually helps.
R: Why, why do you think so,
T: Well, because it doesn't show. Their conclusions don't, don't match.
R: Match with,
T: With, they doesn't, it doesn't show that it is actually helpful. They … talk about …
all these statistics … first of all, they got a whole bunch of, of studies. Which
they mention and they say ok, because there are so many studies and all these
problems with them, that's the problem. There's all these studies and there's all these
problems with them. You can't really draw any real conclusions from these.

Through interviewing, it was found that his rejection of statistics was partly due to his
doubts on the validity and reliability of a pretest and posttest design. He said, “there are many

179

reasons why people can do well on pre and post test. If I gave 'em … a difficult pretest and easy
posttest, I don't have to do anything in the middle.”
It should be remembered that, at Interview 1, based on his knowledge on brain-based
learning, he was dubious of the effectiveness of CF. In addition, although he did not refer to their
effectiveness, he said that he would use recasts. Lyster and Saito (2010) found a result somewhat
contrary to these stated beliefs of his in the sense that the study showed that (a) CF was more
effective than no CF, and (b) the effectiveness of recasts was less than that of other types of CF.
Therefore, like the three other teachers, it is possible that he was questioning this study because
its claims did not fit his stated beliefs prior to reading.
Tom marked Mackey et al. (2000) as “interesting,” because of “the groups, groupings
of people that they used.” However, he criticized the study from two standpoints. The first one
was, similarly to Lyster and Saito (2010), its long list of references, and Tom said, “it's, again,
lots of, tons of references, that's how you get it published … And if you don't get published, then
you, you don't get tenure, That's the main function of journals.” The second one was its small
sample size, a point which Jim also mentioned, and Tom said, “when it’s a small group, that’s
always a problem,” and “this is a small study. See, [Lyster and Saito (2010)] was too much,
[Mackey et al. (2000) was] too small.” However, his focus was on Lyster and Saito (2010) and
Truscott (1999), so he did not make further comments on Mackey et al. (2000).
Like John and Jim, Tom was reading Lyster and Saito (2010) and Mackey et al. (2000)
from a researchers’ perspective, judging from the fact that he was criticizing them from such
viewpoints as (a) the use of statistics, (b) the procedure of experimental studies, and (c) the
sample size. This was probably due to the training in reading research studies that he received
during the MA program and his experience of writing two research studies. At the same time, in

180

contrast to this researchers’ standpoint, the way of writing in these two articles, which was
typical in research studies, bothered him. This is similar to Cecile to some extent; although she
did not show her discontent, definitely, they were not reader-friendly for teachers. Thus, Tom
was evaluating the contents as a researcher but reading the articles as a teacher. For Truscott
(1999), Tom generally expressed his agreement with the study. One reason that he cited was the
same as the one just described above, its accessibility, saying, “its simplicity. This one, much is,
in a basic style, but it's much easier to understand,” “simpler to understand but because it's well
written,” and not “badly written.” The other reason, which seems more crucial than the first one
was that the main topic of this article, rejection of CF, was somewhat in line with his stated
belief on CF prior to reading based on brain-based learning (It should be remembered that he
rejected Lyster and Saito (2010) for the same reason as described above), saying, “[Truscott]
resonated with me because I’ve, this kind of the experience that I’ve had.” He further stated that
these stated beliefs of his were coming from many sources:
Partly is, is, partly on my own bias, and partly my own experience, and partly what I've
learned about psychology … and the human brain and how it operates, not just in
language, but in general, so, this rings more true to me.

He specifically expounded on his “own experience” from two different perspectives:
L2 acquisition and teaching. For the former point, he showed his doubts on the effectiveness of
CF by saying, “It doesn't seem to help … whether someone corrects you or doesn't correct, it
doesn't seem to have any varying of whether you learn it or don't learn it.” For the latter point, he
stated:
It … does have an effect on how you feel about what's going on depending on, as a

181

student, and as, as a teacher watching students … when I was doing that, I could
sometimes tell … that was a mistake because it wasn't well received or other times it
was oh that work pretty well.

At Interview 1, he showed his reluctance to give the explicit types of CF, and said that
he preferred to use recasts. In contrast, Truscott (1999) actually claimed that all kinds of CF
needed to be completely abandoned in L2 classrooms. This seems contradictory; however,
apparently, Tom’s use of recasts was not grounded on his belief that recasts were effective.
Instead, he was using recasts simply because his students “wanna know how to do it right”
(Interview 1). Thus, he did not believe in or at least was dubious of their effectiveness, which can
be seen in the quote below where he emphasized the unhelpfulness of CF based on his
knowledge of brain-based learning:
Human brain has lots of holes in 'em, And in a sense, we hear what we expect to hear,
we see what we expect to see, It's how we survive, s, we fill in things that aren't there,
And we, we take out things that actually are there, It's how the brain functions. And it
happens to students in class with error correction. They don't hear it.

In the quote below, he expressed the same opinion, but this time, he directly referred to
CF:
If you know anything at all about neuroscience and how complex a brain is, a little bit
about psychology and … how their, the brain functions and how people see and don't
see things, and, and hear and don't hear things, then it stands a reason that error
correction is gonna be ineffective.

182

In order to justify his claim that CF was ineffective, he used the same reasoning as
what he had used at Interview 1, which was that L2 learning was identical to child L1 learning:
When I was raising my kids, you don't go around correcting kids all the time, you just
give them input, input, input, input, and uh, they learn it. They correct themselves over
time. And I think that uh all humans have innate ability to learn languages … if we
have that enough intelligence, we can speak, even people with a relatively low
intelligence still speak, uh, some of them speak more than one language if they were
exposed to it, So, we have a natural ability to learn languages … even for second
languages.

He made a similar comment referring to Krashen (1981)’s input hypothesis, and
adduced the accelerated learning that he used as a successful example of L2 teaching:
I always think how do kids learn. They don't know any grammar at all … They just get
input … And that's the input hypothesis … it goes … not just Krashen and those guys,
but it's also other people, doing the same thing. So, the most success I found was really
hard, students have a hard time to learn is accelerated learning, They learn it without
realizing that they're learning it.

Apparently, in the accelerated learning that he implemented, learners were exposed to
abundant input and CF did not play a major role in it, as when people learn their L1. He
explained:
I've taught using accelerated learning, and there, there I just give them a whole piles of

183

input, it's pretty much native-like, yeah. And with the input, with the practice, they
learn. And you don't have to worry about explicit grammar teaching … worry about
the order that it comes, it comes the way it comes naturally. So, you, you have a
natural conversation, and whatever grammar comes up in that conversation, that's what
it is.

He further claimed that CF was not only ineffective but also potentially harmful,
saying, “you gotta be careful. It's dangerous,” and “it's real dangerous to do that kind of oral
correction with people. Not that I say inappropriate … but it can … backfire real easily.”
Precisely speaking, he adduced that CF “embarrasses” students and said that such incidents
happened to him when he was learning an L2:
I just wouldn't try any more, I wouldn't try to say anything. Because I couldn't, and I,
then I would tell myself that I can't do it. I keep making these stupid mistakes, so I'm
just not gonna do it any more. And then I would just use fossilized stuff that I already
knew, and I wouldn't try anything new, yeah. So I see students doing that sometimes.

All of these factors led him to say, “I personally prefer to focus on what people do
well.” However, despite his rejection of CF based on several reasons, in contrast to Truscott
(1999)’s claims, he still stated that he would use recasts, but concurrently, he admitted that their
effectiveness was not guaranteed. It should be noted that the use of recasts would support these
two points above that he believed in because (a) recasts are one of the input-providing types of
CF, and (b) recasts are usually implicit, and not likely to humiliate learners. Despite his virtually
overall rejection of CF, owing to these two points, he may not have stopped giving recasts

184

without necessarily believing in their effectiveness. It is also possible that, when he mentioned
CF in those quotes above, he was referring only to the explicit types of CF, excluding recasts.
Asked whether CF would interrupt communication, he acknowledged that it generally would, but
not always, and adduced recasts as a way not to interrupt communication:
You can sometimes slide it in, so the specially in the kinds of where you're restating
things, That's the ones I tend to use the most, and of course, sometimes people catch it,
sometimes they don't, but, restate what they said.

Asked whether or not he cared that some students would not catch his recasts, he
added:
I can't care about it, because if they catch it, they do, if they don't, they don't … that's
true of any of my teaching … I can do the best I can, but if the people aren’t receptive
for whatever reason, it's not gonna stick …[teachers] can only present it and as an
interesting, and even if you are really interesting, there is always somebody who isn't.

Here, he showed that he was well aware of individual differences in L2 learners that
had an influence on the effectiveness of recasts, which was a source of practical difficulty in
giving effective CF. Regarding the “receptiveness” that he mentioned above, he also said:
As a teacher watching students … when I was doing that, I could sometimes tell …that
was a mistake because it wasn't well received or other times it was oh that work pretty
well … So it depends on the, the student themselves whether they're receptive or that
they want it, uh, their mood for the day, there's like so many many psychological
environmental factors. Cultural factors that come into it … it's hard to determine

185

whether it's a right thing to do or not.

To one of the claims that Truscott (1999) made, which was “students may not notice
error correction,” he responded, “That's true. They may, may not. Sometimes they do, sometimes
they don't.” Further asked why, he added:
There are thousands of reasons for it. We don't really know …they could be tired, they
could be, nowadays, looking on their cell phone, they could be thinking about what
they're gonna do after class, they could be thinking about what they did before class,
they could be worried about. Uh the test, they could be worried about the test in the
next class, they could be, they could be worrying about what they're gonna do with that
guy or that girl over there in the other area, and across the room … they are hungry.

To another claim by Truscott (1999), which was “students may not take error
correction seriously,” he responded:
That happens. You can see it … Some students, not, it's not important … They don't
care … that can be for many reasons too. They are more interested in communicating,
than worrying about the minute, whether I've got that present perfect correctly or
whether I, if the person understands what I said, that's what's important. And, that's
normal human.

However, like Cecile, Tom did not believe in every claim that Truscott (1999) made.
For example, he showed his uncertainty about the validity of another claim made by Truscott
(1999), which was “error correction facilitates explicit knowledge, but does not facilitate implicit

186

knowledge.”
I can't say that that's true, because we don't know, Uh, it's for some students, yeah.
They, they explicitly, they can learn it and see that but then they, they can't use it, But
for other students … they can't explicitly do well but they can use it correctly, so, I
don't know if that's necessarily true, I think it varies.

He also commented that whether or not CF would be harmful depended on the
relationship between teachers and students. Such an opinion of his was in contrast to the claims
made by Truscott (1999), who believed that CF could not be effective. Drawing examples from
his own L2 learning experience, Tom said:
I've had some great teachers that I remember. And I know that they, they did
corrections in class and so forth, But because, I had a great relationship with the
teacher, there was no adverse impact any of those things. And if there were, it didn't
last very long. But I've had other, other classes where, I've had teachers that, there
wasn't a, a bond or like uh a relationship or it was a negative one. And even if they did
a really good job, it came across as bad. Not because of them, but because of me. My
attitude toward them. So, one never knows that as a teacher. You can sometimes guess,
you can sometimes read people's face sometimes, but you can't always be even certain
of that, They may be frowning because they're tired, not because they're angry …
there's all kinds of things going on.

He generalized this notion to “correct in a correct way.” This is similar to John, but in
contrast to John who used it clearly referring to “not belittling his students,” Tom’s definition

187

was intentionally vague because he believed that “a correct way” depended upon each of the
students:
If you really like your teacher, and they correct you, and they do it … In a correct way
and whatever that means, that individually, that's for individual students to determine,
it's not for the, the teacher or somebody else to determine what's the correct way.

Being aware of these individual differences may have been another source for him to
claim that the effectiveness of CF could not be guaranteed.
In summary, among the three research studies, while admitting the valuable work done
by Lyster and Saito (2010), he basically rejected its claims altogether mainly because of its rather
inconclusive statistical results. He did not criticize Mackey et al. (2000) as much as he did Lyster
and Saito (2010), but still he did not take it seriously partly because of its small sample size. In
addition, from the point of the style of writing, he rejected both of these studies based on the
norm of writing in academia that they adopted. In contrast, he generally showed his favor for
Truscott (1999), whose claims seemed to partly coincide with his own stated beliefs, which was
to abandon CF completely in L2 classrooms.
However, examining his comments closely, the picture is not so clear and
straightforward. When he talked of CF directly, he rejected its use, based on several reasons. One
reason was that it was ineffective. However, referring to the conclusion regarding the
effectiveness of CF in Lyster and Saito (2010), he said:
They kind of say that at the end. We think maybe … But we don't know. Well, if you
don't know, then you don't know … they did a lot of studies to find out that they didn't
know, Yeah. So, we still don't know. Whether does or not. We go by our beliefs. If it

188

works or doesn't work.

In short, he stated that (a) he was still unsure whether or not CF was effective, and (b)
we had no choice but to depend upon our own beliefs on CF when using it. This quote is
interesting in the sense that, although he often expressed his beliefs that CF was not effective so
adamantly and clearly that they would not seem to fluctuate in any way, the reality was that he
was still unsure and felt that it had not been proven one way or the other. Thus, to him, CF was
something that did not have to be utilized necessarily as it was not indispensable for L1
acquisition and its effectiveness was not guaranteed.
Moreover, when the topic was about individual differences, he admitted that there may
be some occasions where CF would not be harmful and he actually adduced examples from his
own L2 learning experience. Also in that anecdote, he qualified his other reason for rejecting CF,
which was that CF may humiliate students, by saying that it would be acceptable on the
condition that a rapport was established between teachers and students and students were willing
to learn. Thus, although Truscott (1999) was the study that Tom favored the most, he was not in
complete agreement with every claim that Truscott made.
It seems that he was well aware of the potential danger that CF had. It may work if
certain conditions are met, but it may not work or even work negatively in some situations.
Fearing the consequences that CF may bring, in addition to the fact that it seemed that no
research had shown CF was definitely effective thus far, he may have retreated from giving CF
instead of trying to solve the unsolvable puzzle of judging when and how to give CF “correctly,”
and he supported his theory by such other sources as brain-based learning.
Neither Lyster and Saito (2010) nor Mackey et al. (2000) helped him to solve the

189

puzzle in any way, and Truscott (1999) generally supported his stated beliefs on CF, which was
to abandon it. Thus, none of these studies added anything particular to his original stated beliefs
on CF, which ended up by not having much influence on his use of CF. Asked how reading these
studies impacted his teaching, he responded, “not a whole lot in the sense that 'cause … I've
already known about, these things and these particular arguments.”

5.5 Discussion
In this section, the findings that emerged through a cross-case analysis across the four
teachers’ responses to the three research studies in relation to their stated beliefs on CF is
discussed. First of all, it seems that reading the three research studies provided the teachers with
the chance to reflect on their stated beliefs and classroom practices of CF; however, this issue
should be discussed from a broader perspective because participating in the present study itself
seemed to be a whole large opportunity to do so. Thus, this will be taken up in the final chapter.

5.5.1 The Responses to the Three Research Studies
In this section, some of the common points as well as differences in the four teachers’
responses to the three research studies will be discussed. A behavior observed commonly among
the four teachers was that they all read Truscott (1999) from a teachers’ perspective. This was
quite understandable because (a) they were all L2 teachers, and (b) the prospective readers of the
study were L2 teachers. However, except Cecile, all of the three other teachers also read the
other two studies, Lyster and Saito (2010) and Mackey et al. (2000), and to a small extent,
Truscott (1999) as well, from a researchers’ perspective. Again, this was understandable because
these two research studies were written mainly for researchers. It should be noted that Cecile was

190

the only one who was lacking a researchers’ perspective. This is probably because of the fact that
she just began to study in the MA TESOL program, and had not been exposed to various kinds
of research studies nor learned how to read them critically. On the other hand, the other three
teachers had already been trained to do so in their MA TESOL programs. However, because they
were not recent graduates from those programs, some other factors might have been playing a
role, such as the amount of teaching experience and their motivation for professional
development.
Another common point observed among the four teachers was that all of them were
basically agreeing with the studies whose findings or claims concurred with their stated beliefs
on CF prior to reading, and disagreeing with the other studies. Concisely speaking, John and Jim
believed in the effectiveness of CF, and both of them favored Lyster and Saito (2010), which
found that CF was effective, as the most convincing study and disfavored Truscott (1999), which
claimed that CF should not be used. On the contrary, Tom, who did not feel the necessity of
using CF, chose Truscott (1999) and rejected both Lyster and Saito (2010) and Mackey et al.
(2000). What was particularly interesting was Cecile’s case. Although she did not have any firm
stated beliefs and expressed her willingness to absorb anything from any studies as a “sponge,”
she was highly selective in the sense that she preferred Truscott (1999), and furthermore she did
not accept every claim that Truscott made. Thus, the relationship between all of teachers’ stated
beliefs on CF and their responses to the three research studies was bidirectional, rather than
unidirectional (S. Borg, 1999a). Not only did their reading have an effect on their current stated
beliefs on CF in the sense that it raised their awareness of CF (as was also seen in Mackey et al.
(2004) using a workshop), but their stated beliefs on CF prior to reading also had an effect on
how they read the studies. This connects to another important point that each teacher’s

191

expectation for research studies differs, and cannot be predicted easily. Thus, the effect of
reading research studies will be highly individualized for each teacher even if they read the same
list of research studies. Judging from Cecile’s responses, how teachers would read research
studies is highly unpredictable even if the reader is a novice L2 teacher.
Another noteworthy point commonly observed among the four teachers was that there
were only scarce responses to Mackey et al. (2000) from the four teachers. This study was
included with the intention of providing them with the information on the effectiveness of CF
from the viewpoint of learners’ perception of CF. However, none of the four teachers showed a
deep interest in it. One reason is that it was not clear what this research study was trying to
convey as the main idea, and therefore, the four teachers were unsure of what they were
supposed to extract from it. In contrast, Lyster and Saito (2010), on one hand, clearly stated that
CF was effective and some other minor points, such as prompts were more effective than recasts.
Truscott (1999), on the other hand, claimed that CF should be abandoned. Thus, these two
studies’ claims were at the complete opposite poles, so it was easy to contrast them. However,
the position that Mackey et al. (2000) took was somewhat vague, somewhere in the middle
between these two extremes. Its findings were not as straightforward as those in other two
research studies in the sense that its main conclusion was that CF may be effective in some cases
depending upon the type of error. So it may have been somewhat difficult for these four teachers
to take their stance on its claim. Moreover, the notion that CF depends on many variables was in
fact in line with the stated beliefs of all of these four teachers as will be described below;
therefore, they might have perceived the results in this study common sense, and they did not
have any strong opinion either for or against the study.
Regarding the four teachers’ responses, overall, the evaluation of Lyster and Saito

192

(2010) and Truscott (1999) was completely split among the four teachers. While John and Jim
favored the former, both Cecile and Tom generally favored the latter. The reason for favoring
Lyster and Saito (2010) by John and Jim was mainly that it was a meta-analysis. As John stated,
a single experimental study may be inadequate to be cogent, but the results derived from 15
different studies seemed to them convincing enough to consider the claims in Lyster and Saito
valid and reliable. Although 15 is a smaller number than most of other meta-analyses, it seemed
not to bother John and Jim at all. At the same time, Cecile and John expressed their
dissatisfaction with the lack of detailed information about those 15 studies. This is partly
unavoidable due to the nature of meta-analyses and the strict limitation in the number of words
or pages in academic journals. Having said that, this shortcoming did not hinder John’s belief in
this article.
Nevertheless, even Jim, who favored Lyster and Saito (2010), confessed that he did not
understand its contents completely. Several points that made reading research studies difficult for
these four teachers were found through interviewing. The first one was the use of statistics, as
discussed in Chapter 1 (e.g., Zeuli, 1994). The statistics in Lyster and Saito seemed too
complicated for some teachers to understand. Tom explicitly expressed his insufficient
knowledge on statistics and did not know what k sample was. Jim avoided reading the statistics
section, and was confused between the correlation coefficient and the effect size and asked me to
clarify it. Although Cecile did not express any concern (she skimmed the article), as a first-year
MA student with a background in literature, it would be legitimate to suppose her knowledge of
statistics was far from enough to understand a meta-analysis. So these teachers had to trust
whatever was presented in the research studies at its face value. Lyster and Saito explained the
formula in the article, but this was quite difficult to understand for anyone without a background

193

in statistics or meta-analyses. In contrast, Truscott (1999) did not include any statistics, mainly
because it was a position paper. This is one reason why these teachers commended Truscott on
the grounds of its easiness to read.
Another point that would make research studies incomprehensible for teachers is
teachers’ unfamiliarity with research studies, specifically their terminology, as discussed in
Chapter 1 (e.g., Bolitho, 1991), and methodology. Regarding the former, Cecile asked quite a
few questions regarding the terminology of CF. Some questions were basic while others were
advanced. Regarding the types of CF, although it was explained in the PowerPoint file, she
showed her confusion at Interview 2. In addition, some of the expression used in the research
studies were written on the assumption that the readers already knew them, so there was no
explanation or operationalization for such words as interlanguage and post hoc. Cecile’s lack of
such terminology was understandable because she was still novice in the field of TESOL, but
this shows that these research studies are not written in a way so that MA TESOL students can
understand the contents easily, although they are definitely one group of potential readers.
Regarding the latter, Tom thought that a pretest and posttest each used different test items, and
therefore the use of gain scores was questionable. This issue, as well as the importance of
securing a control group in order to see the test effect while using the same test items in pretest
and posttest, is somewhat common sense for experimental researchers, and Lyster and Saito
(2010) compared the effectiveness of each type of CF against the test effect observed in the
control group in each of the 15 studies included, but such methodology may not be obvious for
L2 teachers. These points may prevent L2 teachers from understanding research studies
thoroughly.

194

5.5.2 Stated Beliefs on CF
Analyzing the four teachers’ stated beliefs on CF, five issues that were common among
all or most of the four teachers emerged. The first point was, except Tom, all of them claimed
that the use of CF (at least recasts and/or clarification requests) was “natural.” However, as
described above, the meaning of “natural” was different among those teachers. The use of recasts
and clarification requests was “natural” in the sense that (a) “it was a natural part of
conversation” for Cecile and Jim, (b) it had become “unconscious and automatic” in his teaching
for John, and (c) it was his job as a L2 teacher for Jim. Regarding (a), the use of CF was a part of
interactions in not only the classrooms but also daily conversations, and not only with non-native
speakers but also even between native speakers; therefore they thought that they neither could
nor should avoid using recasts and clarification requests. Even Cecile, who was reluctant to give
CF, and Jim, who believed prompts were more effective than recasts, confessed that it would be
unnatural not to use recasts. Finally, even Tom, who disfavored CF, said that he would use
recasts although he did not say it was “natural” (however, he would say so at Stimulated Recall 1
as will be shown in the next chapter). Therefore, although their reasons for using CF may differ
from each other, the data from the three observations supported this notion in the sense that these
four teachers used recasts the most and clarification requests the second most (except Jim, who
used explicit correction more than clarification requests) among all kinds of CF.
The second point was that the teachers believed that their students expected CF from
them in their classrooms. As described above, in both John’s and Jim’s cases, their use of CF
was partially in order to fulfill the expectation of their students. Tom also stated that his use of
recasts, although he was dubious of their effectiveness and necessity, was for the sake of his
students. In Cecile’s case, however, she was determined to give less CF although she believed

195

that her students expected CF from her. This may be because of the following three points.
The third point was that CF should not disrupt communication. As discussed in the
previous chapter, communication was important and all of these four teachers adopted CLT to
some extent. Both John and Jim, who believed in CF, stated this point. Thus, in these teachers’
minds, communication was more prioritized than correcting students’ errors.
The fourth point was that CF should not humiliate students, and this seemed to be a
crucial criterion in deciding when and how to give CF. Both Cecile and Tom listed this point as
one reason to refrain from giving CF. Both John and Jim, in contrast, while stating that CF
should not embarrass their students, claimed that CF would not do so if it was given in the
“correct” way. Actually, even Tom argued for this idea on the assumption that certain conditions
were met, such as a good rapport between teachers and students. This is probably coming from
his abundant experience of L2 learning and teaching. However, Tom also said that “the correct
way” was so individualized that he was incapable of knowing it in each case, which may explain
why he was reluctant to give CF. In contrast, as a novice teacher, Cecile may not have been
aware that there was “the correct way,” or even if she did, she could not have found what “the
correct way” was yet. Either way, it was natural for her to decide to diminish the use of CF in
order to exercise the greatest caution not to humiliate her students.
The final point was, as briefly mentioned above, all of these teachers were well aware
of the fact that the effectiveness of CF depended upon various variables. One of them was
individual differences, such as students’ motivations and proficiency levels. Even Cecile, who
was relatively inexperienced, expressed this concern. Another one was contextual factors of the
class, such as the goal and the objective of the lesson and the skills to teach. The teachers were
also aware or came to believe that the error type (e.g., morphosyntax, lexicon, pronunciation)

196

played a role in the effectiveness of CF, which was one of the findings in Mackey et al. (2000).
Thus, they knew that there was no single definite answer to CF, and “the correct way” to give CF
varied in each moment, in each class, for each student, and they all knew the practical difficulty
of giving CF. But each teacher decided to deal with this variability differently. While Tom
preferred not to tackle it, knowing that it may sometimes succeed, both John and Jim preferred to
still try their best, knowing that it may sometimes fail, believing in the usefulness and
effectiveness of CF. Thus, although their stated beliefs on CF showed a number of differences, it
seems that these three experienced teachers all shared the same perception on CF to some extent,
which was the success of CF was not guaranteed and it could be harmful.
In summary, this chapter explored the four teachers’ responses to the reading of the
three research studies. It was found that the relationship between their stated beliefs on CF and
their responses was bidirectional. That is to say, while their awareness of CF was raised by the
reading, their choices of the articles to identify themselves with was influenced by their stated
beliefs on CF prior to the reading. Moreover, while the stated beliefs on CF of the three
experienced teachers did not show any fluctuation, in contrast, Cecile began to form stated
beliefs on CF through the reading although they were still unstable.

197

Chapter 6: The Effect of the Reading on Teachers’ Stated Beliefs and Classroom Practices of CF

The previous chapter described how the four teachers responded to the reading of the
three research studies. It was found that the relationship between the reading and their stated
beliefs on CF was not unidirectional, but bidirectional. In this chapter, the effect of the reading
on their stated beliefs and classroom practices of CF will be examined in depth while describing
them posterior to reading using mainly from the data of the two classroom observations and
following stimulated recalls in addition to those from Interviews 1 and 2 and Observation 1.
Because of the busy schedule of the four teachers, as can be seen in Table 4, the duration
between Interview 2 and Observation 2 varied for each teacher: 22 days (Cecile), 11 days (John),
and 6 days (Jim and Tom). Likewise, the duration between Observations 2 and 3 also varied for
each teacher: 19 days (Cecile), 35 days (John and Tom), and 25 days (Jim). All the stimulated
recalls were conducted either directly following or a few hours after the observations on the
same day. While answering the third research question, as shown below, the following sections
will introduce the description of each teacher starting with the quantitative results about their
classroom practices, which will be then followed by the discussion for the whole chapter.
3

What effects does the reading have on the stated beliefs and classroom practices of CF of the
four ESL teachers?

6.1 Cecile
Table 10 shows the number and the percentage of error-filled AS-units, those followed
by CF-filled AS-units, those not followed by CF-filled AS-units, and each kind of CF provided
by Cecile. On average, an error-filled AS unit occurred approximately every 2.9 minutes, 1.9

198

minutes, and 3.1 minutes (Observations 1, 2, and 3, respectively), which was not as frequent as
those of the other three teachers. One reason for such relatively low incidents of errors can be
ascribed to the fact that the utterances of her students were quite short and in many cases the
audio recorders could not capture their voices clearly.
The data show that her use of CF was every 5.6 minutes (0.18 CF per minute), 2.3
minutes (0.44 CF per minute), and 4.5 minutes (0.22 CF per minute), respectively. Although
these rates of giving CF was smaller than all of the 12 teachers observed in Loewen (2003),
because the number of error-filled AS-units was also small, she actually provided CF for around
half of all the errors-filled AS units (44.4-52.9%).
Table 10
Cecile’s Use of CF at Three Observations
Error-filled AS-units
Kinds of CF-filled AS-units
Without
With
Obser
Clarification
Explicit
Total CF-filled CF-filled
Total Recasts
vations
requests
correction
AS-units AS-units
1
17
8 (47.1)
9 (52.9)
9
8 (88.9)
1 (11.1)
0 (0.0)
2
27
15 (55.6) 12 (44.4)
22
8 (36.4)
13 (59.1)
1 (4.5)
3
16
8 (50.0)
8 (50.0)
11
6 (54.5)
5 (45.5)
0 (0.0)
Total
60
31 (51.2) 29 (48.3)
42
22 (52.4)
19 (45.2)
1 (2.4)
Note. All the classes lasted approximately 50 minutes. The numbers in parentheses show
percentages. The number of the error-filled AS-units with CF-filled AS-units may not
correspond with the total number of CF-filled AS-units because a single error-filled AS-unit may
elicit multiple CF-filled AS-units.
Cecile did not have firm stated beliefs on CF at the time of Interview 1, but said that
she used recasts because they were a part of a natural conversation. The three observations
showed that she gave CF for almost half of all the error-filled AS-units, and the majority of the
CF that she used was recasts, although she may have been using them unconsciously. However,
influenced by Truscott (1999), at Interview 2 she stated that she came to believe that CF should
not be given. At the same time, however, due to the naturalness of the use of recasts and

199

clarification requests, instead of abolishing their use wholly, she stated that she would try to give
less CF than before.
In the three weeks from Interview 2 and Observation 2, which was the longest duration
among the four teachers, her stated beliefs on CF had started to sway and she expressed her lack
of confidence in the claims that Truscott (1999) made. She said, “I feel like maybe less
conclusive than the last time … whatever I said so, about agreeing with Truscott” (Stimulated
Recall 1). Her recount revealed that this was because of the classes that she was taking
concurrently. She said, “since the last time we met, we talked about error correction in a couple
of my classes and perhaps … I don't know exactly how to deal with this. So I, I would like to
learn more about it” (Stimulated Recall 1). This was in line with one of the points that Nisbett
and Ross (1980) and Pajares (1992) argued in the sense that, because her stated belief that CF
should not be given was a newly formed one, it was still vulnerable. Therefore, when another
source came in, it was easily shaken.
At Stimulated Recall 2, which was conducted approximately three weeks after
Stimulated Recall 1, it was found that her stated beliefs on CF switched to the opposite side. She
expressed the same opinion repeatedly, saying, “I think I’ve definitely changed my opinion from
when I first talked with you,” “I think my, my opinion changed as I’ve studied more and learn
more about it,” “I would definitely say my mind has changed,” and “I really feel like, ((laugh))
my mind has changed” (Stimulated Recall 2). Accordingly, she openly expressed her beliefs in
CF, saying, “I feel it is important to, correct students … it’s what they’re spending money to be
there to learn, improve,” and “you're the authority, you have to kind of tell … this is right, this is
not right is the is the only American in the classroom as the expert … I am the expert”
(Stimulated Recall 2).

200

As described in the previous chapter, she was comparing L2 learning with math and
science, and claiming languages were much more flexible than those subjects; thus it was
difficult to judge whether certain linguistic expressions were right or wrong. At Stimulated
Recall 2 after finishing watching the video of Observation 3, she rephrased this point, but this
time, used it to justify the use of CF:
In a math class, you don't just let any answer fly, you have to say that, that’s wrong. So
like saying that’s true true even though language is much more, adjustable or more …
fluid … than numbers and math … but still … there needs to be … some kind of
measuring stick of what's right and wrong, and, they're gonna be judged by the way
they speak so I should help them to understand that.

Although believing that CF should be used, she had not formulated any concrete ideas
as to how she would give CF in her classroom, which she perceived as a problem, stating, “I
should just develop a, a system for that … I think it would be helpful, um, I don’t know how,
how to do it though, exactly that’s the problem” (Stimulated Recall 2).
Multiple sources of this fluctuation in her stated beliefs between Stimulated Recalls 1
and 2 were identified through the stimulated recalls. One of them was one of the classes that she
was taking, and especially reading another article in one class. Recalling Interview 1, she said,
“at that time, I hadn’t really read a lot about those kinds of things, so I didn’t have a clear
formulated opinion” (Stimulated Recall 2). The study that she read was Chandler (2003), which
showed that students’ writing improved through written corrective feedback, and among the four
choices (a) correction, (b) underlining with description, (c) description of type only, and (d)
underlining, correction was the most effective. In addition, the study investigated the students’

201

perception of these four types of correction, and although their responses were rather mixed, they
generally preferred correction. Despite the fact that this study was on written corrective
feedback, not on CF (i.e., oral corrective feedback), this also influenced her stated beliefs on CF
in general. Again, toward the end of Stimulated Recall 2, She stated:
I saw different research … and the more I thought about it and, and read things … I
think this is valid … there's studies for written … he kind of had an experimental
design to see if oral feedback had an effect but we talked about lots of … research
setups for written feedback and there seems to have … positive results in giving
feedback.

Although she said that it was the reading itself that impacted her stated beliefs on CF,
she also felt that how the professor dealt with the article in the class had an impact on her
thoughts. Also at the end of Stimulated Recall 2, asked whether the reading would have had the
influence on her stated beliefs to the same extent assuming that she had read it by herself, she
answered:
I think so, The article was well written and I think it made, probably not as strongly,
'Cause I think [the professor] … drew out the main points and really helped to clarify
… Maybe her, her really saying what she thought about it had an impact.

Another source of her fluctuation was the feedback from her advisor who frequently
observed her class and had sessions with her. The advisor criticized her for not being precise
about answers that students produced in her class. She recalled, “[the advisor] observed my
teaching too and sometimes, you don't have concisive or … decisively, right or wrong answers

202

for the students, so, that can be confusing” (Stimulated Recall 2). In another class that she was
taking, one of her classmates also criticized her for this point, saying, “you can't do that. You
can't let it slide” (Stimulated Recall 2). Watching the video of her own teaching, she reflected on
this point, and she confirmed this point as her shortcoming, saying, “watching today on, on this,
perspective of students, they might feel like well, she never really tells us what the answer really
is. Well, not never. But sometimes I leave things kind of open-ended, which, isn't the best”
(Stimulated Recall 2).
At the same time, throughout the two stimulated recalls, she acknowledged her
“spongeness,” which was exactly why her stated beliefs had swayed in such a short period of
time, but concurrently implied that her stated beliefs on CF were still unstable, and may further
fluctuate in the future. She said, “I really feel, like everyday, I have a new thought about things,”
“kind of changing and constantly thinking oh I should be doing this instead of this and that”
(Stimulated Recall 1). Such an attitude was evident through the end of the present study. She
said, “I'm new and I'm not exactly sure what I think about everything so I am easily influenced
probably” (Stimulated Recall 2).
Throughout the two stimulated recalls, Cecile made comments on 22 incidents of
errors. Among them, she stated that she had not been aware while teaching that a student made
an error in seven cases (four and three cases in Stimulated Recalls 1 and 2, respectively), and did
not remember whether or not she was aware of an error in four cases (three and one cases,
respectively). Among the 11 cases when she was aware of an error while teaching (seven and
four cases, respectively), she could not remember what she was thinking about at that time in two
cases (one and one cases, respectively). Therefore, this leads to the fact that she was aware of an
error and consciously gave CF in only nine cases while teaching (six and three cases,

203

respectively). This proportion is much less than those of other teachers. Through interviewing, it
was found that she was heavily focusing on content, such as the answers that her students made
in the class while she was hearing errors. Together with the fact that she was still relatively
inexperienced, this may have prevented her from noticing the errors (This issue will be further
discussed at the end of this section). Two examples are shown below. In Example 19 at
Observation 2, the students were working on a listening activity. After listening to news, they
answered three questions about its content. The answer to the second question was he had a
hunch, which a student answered in response to Cecile’s elicitation. Cecile picked up the word,
hunch, and asked the student the definition of the word, which they had already checked in
advance in the same class.
Example 19
T: He had a hunch? Ok, Then what's a hunch again?
S: Ahn, I think (less punch can let pink to) pick up the, their reminiscents and take just,
just take a, just to pick up and remember want to, (have seme), just go through the
old story, just, like that.
T: Ok, so he had an idea to go through old stories, so you're right …

In this case, the student made at least two linguistic errors: (a) making the adjective,
reminiscent, plural, treating it as a noun or using the singular noun, reminiscence (which would
sound like reminiscents) instead of the plural form, and (b) failing to change want to either a
to-infinitive, to want, or a gerund, wanting after remember, while patching several portions of the
sentence. However, Cecile only picked up the last error in content, which was failing to change
story into its plural form, stories, and gave a recast on that part. At Stimulated Recall 1, she

204

stated:
I think … I didn't notice, I don't think all those errors that she made, but now I can hear
them … I probably noticed she wasn't saying it, totally correct or whatever but um, I
wasn't thinking about uh, her exact the words she used as so much as I was thinking
about the content, and uh I think she said um, something about having a hunch, so I
noticed that, she was using the word and so I thought that was good, that's what I was
thinking.

Example 20 was taken also from Observation 2. In this incident, the students were
presenting the metaphors that they had created at home. One of them said, “he is edible like an
elephant” (It should be noted that the terms metaphors and similes were used interchangeably in
this class). Being unsure what she meant by this phase, Cecile further asked a clarification
request.
Example 20
T: … but what do you mean by the elephant. That he is edible like an elephant.
S: Like, elephant usually eats so much. Like, he has really big stomach.
T: Ah, ok, So that elephant can eat a lot. Ok,

The student made an error in omitting a determiner before elephant. Cecile gave a
recast adding the demonstrative determiner that before elephant. At Stimulated Recall 1, she
stated:
I was just starting to get a better understanding of what she was saying or what she
didn't mean to say the word edible, she meant hungry or he is like an elephant because

205

elephants eat a lot or, I was understanding her logic, so I wasn't paying attention to
maybe that error, maybe I didn't hear it or notice it.

Watching herself teach and recalling her thoughts while teaching at the stimulated
recalls made her realize that she was focusing on the content much more than the language,
which she used for the reflection on her own teaching. She said, “I'm realizing I am worried a lot
about the content, so maybe I need to spend more time, focusing on the language and not, oh
they got the main idea, so that's good.”
It should be noted that, in both of these two examples above, she said that she was
focusing on content and did not notice all of the errors. Nevertheless, she used recasts following
the utterance. Throughout the present study, she had been aware of her use of recasts in a natural
way, which she mentioned again at Stimulated Recall 2. In Example 21, the class was creating a
dialog for a cartoon in which turkeys were demonstrating against Thanksgiving.
Example 21
T: … what did you write,
S: I’m not taste good.
T: I don’t taste good? Ok, …

The sentence that the student said was grammatically incorrect, using am not instead of
do not. Immediately after that utterance, Cecile used a recast. She recalled:
It was just natural, yeah. I just, oh I know what you were trying to say so here I’ll say it
for you. Without being super conscious of which type of feedback or something I just,
repeated back what I figured what she wanted to say.

206

As can be seen in Table 10, she used both clarification requests and recasts quite
frequently, which was in line with her comment that they were naturally given in her
interactions. She used only one case of explicit correction for the phrase presented above, he is
edible like an elephant, by saying, “edible means something that is able to be eaten.” However,
this was provided after an extensive interactions between her and her students, so it was
impractical to use recasts any more; in addition, she probably thought that the student’s
misunderstanding of the word, edible, would not be removed without explicitly explaining it
because, replying to Cecile’s elicitation, the student gave the definition of the word, which
turned out to be wrong. Thus, it was clear that she was using clarification requests and recasts
most of the time.
Finally, there was a moment when she stated that she was thinking of the research
studies that she read in the present study. According to her recall, in Example 22, Cecile was
unable to hear a metaphor that one of her students created, which was embarrassed like a
downed rat. After several clarification requests, she finally succeeded in recognizing the phrase,
but still could not figure out the meaning, so she asked a further question.
Example 22
T: Downed rat. Ok, So he's embarrassed like a downed rat, So what does that mean to
you. He feels, how does he feel.
S: Embarrassed.
T: Ok? Like the rats?
S: (Like the rats).
T: Ok, Ok. He's embarrassed like a downed rat, Ok, that, could be one …

207

In this example, the student did not make any linguistic error, but the expression,
downed rat, was incomprehensible to Cecile. At Stimulated Recall 1, she confessed that she was
still incapable of understanding the phrase although she acknowledged it by saying, ”ok, that,
could be one.” She added:
Actually maybe at this point … maybe I was thinking this back about the things that
you asked me to read, I s, I don't remember maybe it was Truscott or somebody maybe
meant multiple times they mentioned … I didn't really exactly correct her error in the
sense of what language she used wrong, but I was probably … this moment thinking,
oh no. What are other students thinking when she said downed rat, like that's not, any
kind of natural metaphor that any one would write in English, so I wasn't meaning to
make it seem like that's a good answer, but I was meaning to say let's just move on, but
it was kind of an awkward situation because her metaphor was so off, and I, I didn't
know how to deal with it, so maybe I was wondering if other students were wondering
what she was saying and, how I dealt with her response.

Her concern that other students may have been wondering about the phrase was not
directly germane to any of the claims that Truscott (1999) made, but rather, it was a moment
when CF could be given in order for other students to understand the student’s utterance
correctly. Cecile actually used a few clarification requests, but as they did not help her
understand the true meaning of the metaphor, she was incapable of clarifying it. Still, her
concern for other students’ comprehension seemed to have been derived from the reading, which
may have provided her with another perspective as to when she would deal with errors in her

208

classroom.
In conclusion, Cecile’s stated beliefs on CF kept fluctuating from the outset of the
present study over the course of two months. She had no concrete stated beliefs on CF at first.
Then, influenced by the claims in Truscott (1999), she came to believe that CF should not be
given because it was ineffective and possibly harmful, but this stated belief was still far from
fixed. Finally, her thoughts on CF were reversed, and she came to believe that CF should be
given because it was effective. However, this latter process was not caused by the reading in the
present study, but by her experiences of teacher development. But as she said, it is interesting
that the strongest influence came from another reading about written corrective feedback. It was
on a slightly different topic, but still reading. This again shows that the vulnerability of her stated
beliefs and her “spongeness” in the sense that she was readily influenced by reading a single
study. Yet, this also proves the potential power that reading research studies has. It was Truscott
(1999) first time, and later it was Chandler (2003). Nevertheless, as shown in the previous
chapter, she was selective in terms of which articles to identify with. She must have read
multiple research studies concurrently with Chandler in the classes that she was taking. It is
unclear how many of them were related to CF, but it is true that she chose Chandler to identify
with for some reason among all, and deemed it more convincing than Truscott (1999). It is
unclear why it was Chandler (2003), but not Lyster and Saito (2010), that had an effect on her,
although both of these studies claimed that corrective feedback was effective and moreover, the
latter had more direct relationship with CF (i.e., oral corrective feedback) than the former (i.e.,
written corrective feedback). One possible reason is that while she was on her own when reading
Lyster and Saito (2010), the teacher in her class explained Chandler (2003), and this may have
helped her better understand the latter than the former. In any case, it is clear that the power of

209

meta-analyses on John and Jim, as discussed in the previous studies, were not exerted on Cecile.
Before moving on to John, it is worth while to discuss the issue that her noticing of
errors was much less frequent than the other three teachers. Also, there were some cases when
she even could not recall whether or not she had been aware of the error at the time when it was
made, which was unique to her. Additionally, she was the only one who could not recall what
she was thinking about regardless of whether or not there was an error, which happened four
times. Thus, it seems that she was the only one who was having difficulty in noticing errors as
well as recalling her thoughts as the other three teachers were successful in both of these tasks.
One reason is her relative lack of experience. Her recall showed that she was focusing
on the content of her students’ utterances at least when they were speaking. But then, as her
thoughts were occupied by comprehending her students’ talk, she may not have had any more
capacity left to notice errors or think of other things. Being relatively inexperienced, she may
have been mainly concerned with classroom management rather than language issues (Johnson,
1992b; Nunan, 1992). Although being an inexperienced teacher may have been enough to create
this situation, the three factors below may have further contributed to making her case more
difficult: (a) her students usually made quite a few errors as she was teaching a lower level; (b)
many of her students did not articulate their speech and spoke in a small voice (which was why
the audio recorders could not capture many of those errors); and (c) she confessed that she
sometimes had difficulty understanding what her students said in the classroom owing to her
10% hearing loss.

6.2 John
Table 11 shows the number and the percentage of error-filled AS-units, those followed

210

by CF-filled AS-units, those not followed by CF-filled AS-units, and each kind of CF provided
by John. On average, an error-filled AS unit occurred approximately every 0.9 minutes, 0.6
minutes, and 2.9 minutes (Observations 1, 2, and 3, respectively). Also, his use of CF was only
once (0.02 CF per minute), every 1.6 minutes (0.64 CF per minute), and 8.3 minutes (0.12 CF
per minute), respectively. The amount of CF that he used greatly varied across the three
observations. Except one case, none of the error-filled AS-units were followed by CF at
Observation 1, but around 25% of the error-filled AS-units were followed by CF at Observations
2 and 3. Especially, Observation 2 produced a large amount of CF partly because the class
mainly consisted of group discussions and his students talked to each other and John extensively,
yielding a large number of errors. In contrast, Observation 3 had the fewest number of errors
because this class mainly consisted of practicing reading aloud written manuscripts. In contrast
to Cecile, who used CF for almost half of error-filled AS-units, on average, he did not use CF for
81.5% of the error-filled AS-units.
Table 11
John’s Use of CF at Three Observations
Error-filled AS-units
Kinds of CF-filled AS-units
Without
With
Obser
Clarification
Explicit
Total
CF-filled
CF-filled
Total
Recasts
vations
requests
correction
AS-units
AS-units
1
53
52 (98.1)
1 (1.9)
1
0 (0.0)
1 (100.0)
0 (0.0)
2
81
58 (71.6)
23 (28.4)
32
18 (56.3)
14 (43.8)
0 (0.0)
3
17
13 (76.5)
4 (23.5)
6
3 (50.0)
0 (0.0)
3 (50.0)
Total
151
123 (81.5) 28 (18.5)
39
21 (53.8)
15 (38.5)
3 (8.1)
Note. All the classes lasted approximately 50 minutes. The numbers in parentheses show
percentages. The number of the error-filled AS-units with CF-filled AS-units may not
correspond with the total number of CF-filled AS-units because a single error-filled AS-unit may
elicit multiple CF-filled AS-units.
John seemed to have firm beliefs prior to the present study; however, because they had
become implicit through his long L2 teaching career, he did not articulate them at Interview 1.

211

Through reading the three research studies, he reflected upon his beliefs and classroom practices
of CF, and, being given an opportunity, elaborated on his stated beliefs regarding the
effectiveness of CF at Interview 2. Although he hoped that he was giving CF “naturally” (i.e.,
automatically and unconsciously) in his classroom, he said that he became more conscious of CF
through reading, and in practice, he wanted to give various kinds of CF. He mentioned neither of
these points directly at the stimulated recalls. Still, one indirect piece of evidence of his stated
belief in the effectiveness of CF can be observed in his comments on specific instances of CF
described in the following paragraphs.
Among 32 recalled instances of errors at the two stimulated recalls, he stated that while
teaching he noticed an error in 25 of those instances and did not notice an error only in seven of
them. In addition, out of those 32 instances, there were only two cases when he was not able to
recall what he was thinking about while teaching, and interestingly, he claimed that he had
noticed both of these errors while teaching. Thus, he noticed quite a few errors in his classroom
while teaching, and moreover, regardless of his awareness of an error, he remembered his
thoughts in most cases. Nonetheless, although he was aware of many errors, as can be seen in
Table 11, throughout the three observations, on average, 81.5% of error-filled AS-units, varying
between 71.6% and 98.1%, were not followed by CF.
As mentioned above, he explained several reasons why he did not choose to give CF in
many instances of errors. One of them was that he prioritized communication over CF, and this
has twofold meanings: (a) he did not give CF when it would hinder communication, and (b)
when communication was successful, even when he noticed an error, he did not give CF. Two
examples for each of these accounts are extracted from Observation 2. In Example 23, the
students were discussing the perception of beauty, and one of them mentioned that it had

212

changed in the course of time.
Example 23
T: Uhn-uh, [student's name], Anything to add?
S: Ah just, I, I said to him like the stand of beauty's changed, [pronounced as
“change-ed”]
T: Uhn-uh,=
S: =So at, at that time, that fat woman was beautiful, but now it is the thin woman is
beautiful,=
T: =Uhn-uh,=
S:=So, I just, I just told like the stand of beauty's changed, [pronounced as
“change-ed”]
T: Yeah.

In this case, the student made several errors, such as using stand instead of standard,
and omitting a relative pronoun, who/that, in it is the thin woman is beautiful. But at Stimulated
Recall 1, John focused on only one of them, the mispronouncing of the past tense –ed, because
there were two incidents of error of the same kind. Instead of giving CF, John had only
acknowledged the students’ contribution simply by saying, “yeah.” He recalled:
He often makes that, it's a very common Korean mistake that, the ed ending is in, is
incorrect. But in that case … I did notice that because that happen if you heard it two
times in succession, changed, changed. But that didn't impede communication with his
partner … I felt like it would've broke the flow of what he was saying by correcting a
word that wasn't even in his content word … just the ed part was … a little bit

213

incorrect … now that I hear that again, I thought at that time … that doesn't impede
communication.

His recall contained both of the two points above: (a) the error did not impede
communication, and (b) CF would have impeded communication. Thus, without a valid reason to
give CF together with a possible detrimental effect, he avoided giving CF. Example 24 is another
example taken from Observation 2 that shows his reluctance to give CF when it was likely to
hinder communication. In this case, a student talked of his thoughts on the change in the
perception of beauty.
Example 24
T: Ok. What i, if, [student's name], you have an idea?
S: Yeah. I think an (met) ah sixteens, ah people is, probably lack of food.
T: Ah,=
S: =And girl who is fat, it means they, ah they're ah rich, /And,
T:

/Uhn uh,

S: There's, make the money is attracting, so the society think that the fat girl, I mean
the big girl is ah, ah, attractive. Beautiful. But in 2000, ah, everybody have a enough
food to eat. /Ah, so ahn, probably they, they won't think this again. And, maybe the
opposite. Yeah.
T:

/Uhn,

T: Ok …

The student made several errors in his utterance. At Stimulated Recall 1, John

214

mentioned one of them, the subject verb agreement, probably because it was recurrent in the
student’s speech. Instead of giving CF, he simply said, “ok.” He recalled:
I remember thinking at that time, that, he has a lot of uh, agreement problems … but, at
the time … I didn't want to stop and correct it, that was gonna break the flow and that
… In that setting with the big group, I wasn't inclined to, I, I felt that I would just break
the flow, p, it would be confusing rather than helpful.

Thus, he decided not to give CF because he thought that it would hinder the flow of
communication. This tendency was common especially when the student’s utterance was lengthy
as can be seen in both examples above, or when the error was made at the beginning of the
sentence even when the utterance was relatively short. This is because giving CF in such cases is
more likely to break the flow of the student’s talk than when the student’s utterance was indeed
short or the error was made at the end of the sentence, as in such cases the teacher can give CF
immediately after the error without interrupting students’ utterances.
Another example where he did not give CF because communication succeeded can be
seen in Example 25 taken from Observation 2. In this example, the whole class was discussing
using such software as Photoshop to alter faces in any way that they wanted. Then, a student
mentioned another way to change faces, not virtually but actually, namely plastic surgery, which
he said was becoming popular among women.
Example 25
S: Even nowadays, (China do shut at enter. Eches).
T: Uhn uh,=
S: =They do, they do a surgery.

215

T: Yeah.
S: To change the face, change the nose,
T: Uhn uh,

In this case, he made an error in do a surgery, instead of have surgery. John simply
said, “yeah” without any CF. At Stimulated Recall 1, he stated:
I think that that wouldn't impede, to do surgery rather than have surgery in the context
of what he was saying that would prevent no, miscommunication to his group … I did
hear that, but I just, he just kept going.

As can be seen in these examples above, he did not give CF when communication was
successful. This in turn means that when communication was unsuccessful he would give CF,
which in fact accords with his explanation at Interview 1 about when he would use CF. A
communication breakdown occurred when John and his students were conversing and he could
not understand what they were saying. In such a case, his CF had threefold functions: (a) in order
for John himself to understand the utterance; (b) to provide the correct form to the speaker of the
error, using CF as an input; and (c) to secure that other students could understand the utterance as
well.
Example 26 was taken from Observation 2 where the pairs or groups of students were
discussing a cartoon in which two women at different eras were imagining their ideal body
images. One of the students described one of the women.
Example 26
S1: This one, like what's, what's she, she get about her future? To be polite, and to get

216

married and to be like be, ah to be in /front, about, about (?),
S2:

/Always? So you think is the her pitch, I, her

futu?
T: Her future.
S1: Yeah. Maybe, I mean, hope she hopes to be.
T: Uhn-uh, Right. Her dream.
S1: Her dream.
T: Yes.

In this case, S2 could not pronounce the word, future, so John gave a recast saying,
future. Then, S1 rephrased what he meant by future, and John found out that the word that S1
was looking for was presumably dream, not future, so he gave a recast saying dream. He
recalled:
I just felt like it had impeded communication … I was clarifying for because he was
working with his partner, so I didn't think that it, his partner would understand what
exactly he was trying to get at by just saying future … the cartoon they were …
thinking about or wishing what they could be … So I guess I wanted to make sure that
his partner understood what he meant by future.

Thus, his CF was not only for himself and the speaker, but also for the partner to
understand the speaker clearly. An intention to give CF for other students can be seen even in
instances in which he was able to understand the utterance; that is, while no real communication
breakdown occurred between him and the speaker, he perceived a communication breakdown

217

between the speaker and other students. Such an example can be seen in Example 27 taken from
Observation 2, in which John elicited responses from some students who could present what they
talked about in pairs or groups to a whole class. One woman mentioned about differentiating
inner beauty and outer beauty of human beings.
Example 27
T: … Ah, [student's name], your group, did you guys have anything different from
that?
S: Yeah.
S: We have different ah, different ah, definition, but they (?), I think beauty of (?)
under beauty for universal.
T: The inner beauty?
S: Yeah.
T: Like inside of you?=
S: =Yeah.
T: Ok,

In this case, the student said, under beauty, which was not a correct expression. John
used two recasts in the form of questions to clarify what she meant. However, through
Stimulated Recall 1, it was revealed that he actually comprehended her remark even before the
first question, but used them for the speaker as well as other students’ benefit. He recalled:
She was explaining the inside, the in, and she kept saying something like that. So, I just
gave her basically the vocabulary word. Oh, the inner beauty. She said yes … in that
situation, she was trying to explain something in the class was listening to her … I

218

think it would've been hard for them to understand. So I clarified, I gave, the correct
vocabulary so everybody could s, hear what she was trying to say.

Through watching his own teaching and recalling his thoughts, as can be seen in the
recall below, John became able to verbalize another reason why he avoided giving CF during
whole class discussion. This shows a case where the stimulated recalls worked as a source for
self-reflection, which possibly facilitated his better awareness of his own teaching practices. In
the middle of Stimulated Recall 1, he mentioned:
I think that the pattern is that I'm seeing is that … a large group, I'm asking for people
to give feedback … communication is the important thing. I want everybody to
understand. And if communication is impeded, then I think what's happening is that's
when I step in and I give either vocab or I give some pronunciation change. At least
that's what I'm noticing so far. And I see a lot more, I noticed a lot more and in that,
even as I think back like in small groups, I gave a lot more discrete little … things that
only [student's name] would struggle with or only [student's name] has a problem with.
I notice that I, I corrected more things like that in the, smaller setting as opposed to the
larger setting when we were debriefing, we wanna move fairly quickly, because we
wanted to go to s, the next part of the lesson.

The data from Observation 2 that he was referring to when making the comment above
show that, among his 32 CF-filled AS-units, 23 were given while students were working in
groups and nine were given during whole class discussion. Thus, his self-analysis was actually
quite accurate.

219

Besides communication, another reason why he tended not to give CF was that he was
focusing on the objective of the lesson rather than the language that the students produced. At
Interview 1, he adduced, “being aware of what your goals are all the time” as one of his teaching
philosophies. Example 28, taken from Observation 3 shows such a case. In this class, his
students were learning how to speak in presentations while especially paying attention to pauses,
stresses, and intonation in and between sentences. Each student wrote an excerpt from their own
presentation script, and in groups, they were reading it aloud to prepare for the upcoming in-class
presentations. John approached one group where a man was practicing.
Example 28
S: The (airs) of game is a very good (activity) thinking for people, especially, for
those who want to try to (manipulize), but it cannot be (realized). It can train your
tactical senses, and do (strong body). Also in the video of brotherhood with your
teammates, and (folk) ah, (falling) the spirit of cooperation. So try to become brave
and strong, by playing this game.
T: Oh, that was really good.

Judging from John’s recall, the student manuscript had a few grammatical errors and
he was aware of them. However, instead of giving CF, he simply commended the student’s
speech. The recall session went on:
T: He clearly had a couple … minor grammatical slipups, but because he really nailed
the ah, he the rising and falling in the intonation and pauses the groups of words … I
gave the him some positive reinforcement there because it was a really good
example and all of his, the people around 'im were listening, were like wow, good.

220

R: So you notice the errors in uh, /his speech.
T:

/Now I did … But I was really focused on today …
the rising and the falling and, some of the things that we've talked about this
semester and he did such a great job of improving …

R: But how how about in the class, did you notice the errors in the way he was reading
our, out, when /(?),
T:

/Uh,

R: Or you didn't,
T: Well no. I did. I did notice that it wasn't perfect, I guess that I just noticed that but
there was never a thought in my mind that I would correct it because it would take
the focus off … I just didn't … think twice about correcting it in that circumstance.
Specially 'cause it's a couple there're article mistakes and this and that mistake …
That would just pull their attention away from what I … was really trying to get at
there.

Thus, the reason why he did not give CF in this case was that although the student’s
speech had flaws, the objective that John was trying to achieve in this particular lesson was
fulfilled by the student, so John’ positive feedback was on that component rather than the
language which was not the focus of the lesson. In turn, in this lesson he gave CF on a number of
errors, usually explicitly, on the three features that were the focus of the lesson: pauses, stresses,
and intonation. Example 29 taken from Observation 3 shows such a case. In this situation,
another student was reading his manuscript, but his intonation for each sentence was always
rising. As soon as he finished reading, John gave feedback on his speech:

221

Example 29
T: [student's name] not bad, Good groups of words but remember, you're going up all
the time at the end.
S: I (do go), just just I realized that.
T: Ok. All right, all right, so, like in this situation that “my answer’s really simple”.
That would be a natural falling. “My answer’s really simple.” Blah, blah, blah, blah,
blah. Tu, tu, tu. All right? Think about, this weekend when you practice. All right,
[student's name], [student's name],
T: Read the last line,
S: This is the aim of freedom.
T: Ok. This is the aim, of freedom. Some dramatic, right? This is the aim, of freedom.
S: Oh, yeah.
T: Yeah. 'Causes that's pow, that's a good powerful line, Use, use your pauses to, let
the reader know and give them some time known, hey, this is, I'm gonna hit you
with this.

In this dialog, John gave lengthy feedback on the student’s rising intonation quite
explicitly. He even had the student repeat a sentence to show an example. At Stimulated Recall
2, he recalled, “because we're focused on, the speaking techniques and the presentation
techniques, I wanted to focus on his incorrect intonation patterns.”
In conclusion, throughout the present study, John seemed to have prioritized
communication in his class, which can be seen in the three points described above: (a) he would
not give CF that would likely to hinder communication, (b) he would not give CF when there

222

was no communication breakdown, and (c) he would give CF when there was a communication
breakdown. Moreover, when he gave CF, he often thought of three goals: (a) for himself to
understand the utterance, (b) for the student who made an error to learn the correct form, and (c)
for the other students to understand the utterance. It should be noted that whereas the purpose of
(a) is mainly communication, that of (b) is purely for L2 learning, and that of (c) could be both,
although John mentioned only communication. Thus, (b) was the only one in which he was
purposefully fulfilling the task of providing correct input. Finally, he focused on whether the
objective of the lesson was accomplished by each student more than on the linguistic errors that
his students made and made feedback on the former point preponderant over CF for the latter.
Examining all of his data, it seems that, unlike Cecile, John’s stated beliefs on CF did
not fluctuate throughout the present study. He was unable to articulate his beliefs on CF at
Interview 1. But as he recounted, this was because as his teaching experience accumulated giving
CF had already grown to be automatic; thus, the truth is likely that he had already established his
own way of using CF in his classroom. However, it was so automatic that it had become rather
implicit, which made it difficult for him to express it explicitly without being provided any
external framework. However, through reading research studies on CF and reflecting on his
stated beliefs and classroom practices of CF, his level of consciousness of CF was raised, and he
became able to articulate his beliefs on CF at Interview 2. Thus, the reading made him aware of
his beliefs on CF, but did not have any major effect on them, possibly because he did not
perceive the necessity to reconsider his use of CF owing to his abundant L2 teaching experience,
except that he may give more various kinds of CF than before. This is because he was very clear
about the points that he regarded as important in his class, such as creating a comfortable
environment for students, successful communication, and critical thinking skills, and these were

223

prioritized over the use of CF. Because CF was not one of his foci in his class, CF was often
avoided for the benefit of these. And when CF was given, care was taken so that it would not
interfere with any of them. Such a policy, which he had established over the course of time, was
consistent throughout all of his data, leaving no leeway for any fluctuation.

6.3 Jim
Table 12 shows the number and the percentage of error-filled AS-units, those followed
by CF-filled AS-units, those not followed by CF-filled AS-units, and each kind of CF provided
by Jim. On average, an error-filled AS unit occurred approximately every 2.4 minutes, 3.6
minutes, and 0.6 minutes (Observations 1, 2, and 3, respectively). Also, his use of CF was every
4.2 minutes (0.24 CF per minute), 4.5 minutes (0.22 CF per minute), and 2.4 minutes (0.42 CF
per minute), respectively. The reason why Observation 3 yielded a large amount of CF was that
this class mainly consisted of students’ short presentations in the front of the class. Their
utterances contained a large number of errors, but as it would have been awkward to stop them
and give CF each time when they made an error, Jim preferred not to interrupt their talk. The
amount of CF that he used remained stable at Observations 1 and 2 and he gave CF 12 and 11
times, respectively. Although he secured opportunities of output for his students, which yielded
some errors, the rate of giving CF was smaller than most of the 12 teachers observed in Loewen
(2003). However, the percentages of the error-filled AS-units followed by CF were 57.1% and
64.3%, respectively. Therefore, although he stated that he would not give CF extensively, he
used CF as much as Cecile did. Having said that, on average, over half of his CF was recasts, and
he also used clarification requests, both of which can be used as a part of conversation.
Therefore, it is possible that he was using these two types of CF unconsciously without the

224

intention of giving CF.
Table 12
Jim’s Use of CF at Three Observations
Error-filled AS-units
Kinds of CF-filled AS-units
Without
With
Obser
Clarification
Explicit
Total
CF-filled
CF-filled
Total
Recasts
vations
requests
correction
AS-units
AS-units
1
21
9 (42.9)
12 (57.1)
12
11 (91.7)
1 (8.3)
0 (0.0)
2
14
5 (35.7)
9 (64.3)
11
5 (45.5)
2 (18.2)
4 (36.4)
3
86
74 (86.0)
12 (14.0)
21
10 (47.6)
4 (19.0)
7 (33.3)
Total
121
88 (72.7)
33 (27.3)
44
26 (59.1)
7 (15.9)
11 (25.0)
Note. All the classes lasted approximately 50 minutes. The numbers in parentheses show
percentages. The number of the error-filled AS-units with CF-filled AS-units may not
correspond with the total number of CF-filled AS-units because a single error-filled AS-unit may
elicit multiple CF-filled AS-units.
Jim’s stated beliefs at the time of Interview 1 was that, although he would not give a
lot of CF, when he did he would give the correct answer possibly with an explanation. As can be
seen in Table 12, as he stated, his use of CF was not extensive, and he used recasts the most
without any explanation. Owing to Jim’s confidence in his way of giving CF that he had
accumulated through his extensive L2 teaching experience, the reading did not seem to have an
impact on his stated beliefs or classroom practices of CF. However, he stated that he might try to
use prompts more than recasts if he found himself using the latter too much, although his
determination was not strong as can be seen in his own remark that he “might” do so, and he did
not mention anything related to his switching from recasts to prompts at the two stimulated
recalls.
One notable difference between Jim and the other three teachers was that Jim was the
only teacher who used explicit correction quite often, comprising 25.0% of his overall CF. For
example, at Observation 3, seven explicit corrections were used for three errors during
presentations when the students were unsure of the vocabulary to use. Example 30 shows a case

225

when the student was presenting the procedure of baking a cake.
Example 30
S: Just make ah, it, and the cake to come uh, when we bake it, can come more, yeah.
T: Oh, ok. Very simple word. Rise.
S: Rise?=
T: =You want to make the cake rise.
S: Ok. To make cake rise …

The student was lacking the word, rise, and instead used another word, come, wrongly.
But Jim figured out what he meant by it and provided the correct word, rise, directly followed by
an example sentence. This may be because speaking up in the middle of students’ presentations
was already intrusive, so he decided to provide more information to clarify how the word should
actually be used in the sentence, although he could have used a recast by simply saying, “Rise,”
without any further utterance. At Stimulated Recall 2, however, he simply said, “[I was] filling in
a vocabulary gap,” so his true intention of choosing explicit correction over the other implicit
types of CF is unknown. At Interview 2, agreeing with one of the claims that Lyster and Saito
(2010) made, he expressed his belief that explicit correction may not be more effective than other
types of CF, but he did not state that explicit correction would be ineffectiveness. Thus, he must
have had no hesitation to use explicit correction in any way, and he used explicit correction four
times at Observation 2 as well. So it is legitimate to assume that he simply did not have a chance
to use explicit correction at Observation 1. However, it should be remembered that he
emphasized the importance of relationship between the teacher and the students when giving CF
in Chapter 5. Thus, it is also possible that he refrained from using explicit correction in

226

Observation 1 as he thought that he had not established a rapport with his students yet. In any
case, the class at Observation 3 was rather exceptional in the sense that it was devoted to the
students’ presentations. Thus, it would be worthwhile to examine an incident of explicit
correction at Observation 2 because the classes at Observations 1 and 2 were conducted in the
same teacher-fronted style. In Example 31, Jim was eliciting the four senses that people had in
addition to the five fundamental ones.
Example 31
T: And the last one,
Ss: Proprioception. [pronounced wrongly by multiple students]
T: Yeah, ((the teacher writes the word on the board)) Yeah. That one is pronounced
proprioception. This is a long o. Proprioception. Ok. And I believe you all know the
meanings of these words. Right?

In this example, after writing the word on the board, Jim explained how to pronounce
it correctly. But instead of using a recast, he used explicit correction with an explanation, “this is
a long o.” Although he could have used a recast, the reason why he chose explicit correction was
possibly that he thought that this word was difficult to pronounce for his students just by hearing
the correct pronunciation. In fact, among the four cases of explicit corrections at Observation 2,
three of them were for pronunciation errors as Jim found his students were having difficulty in
pronouncing proprioception, analogies, and antonym.
At Stimulated Recalls 1 and 2, he was able to comment on his thoughts on every one of
17 incidents of errors. In addition, he stated that he was not aware of the error only twice. This
shows that he noticed many errors that occurred in his class and remembered his thoughts at

227

those moments. Two recurrent themes emerged from his rationales for using CF that were
obtained from the stimulated recalls: (a) the use of direct CF, and (b) giving CF for other
students, each of which will be described in the following paragraphs.
As he expressed his thoughts at Interview 1, his stated beliefs on CF was to provide the
students with the correct answer if it could not be elicited through one or two questions.
Throughout the three observations, he used clarification requests to some extent, and the use of
clarification requests consisted of 15.9% of his CF-filled AS-Units. However, as his recalls at
Stimulated Recalls 2 and 3 show, it was obvious that during the three observations he preferred
to directly correct errors in accordance with this stated belief.
Example 32 was taken from Observation 2 where the class was discussing the senses
that human beings had. Jim asked the whole class what the fundamental five senses were, and the
students were answering the question together. While hearing the answers, Jim wrote down each
of the answers on the board.
Example 32
T: … What are the nine senses? Ok. The five are easy. Right?
S: /Yeah.
T: /Sight, ((the teacher writes on the board))
S: Hearing,
T: Hearing, ((the teacher writes the on the board))
S: Touch,
T: Touch, ((the teacher writes the on the board))
S: Smell,
T: Smell, ((the teacher writes the on the board))

228

Ss: /Taste,
Ss: /Test,
T: And taste, ((the teacher writes the on the board)) Those are the classic five, What
are the four others.

When they reached the last sense, taste, although some of the students correctly
pronounced it, others pronounced the word as test. During the stimulated recall, Jim said that
when he repeated the word as he did with all of the other senses he had “emphasized the word
taste a little bit louder than the other ones that [he] said.” So although his utterance seemed to
have been given only to reiterate the answer, the repetition of taste had a different function. He
used it with the intention of CF, namely a recast. He recalled, “all I was thinking is just about
pronunciation … I knew that the content was right, I was just thinking about … let's get the oral
form of vocabulary item right.”
Example 33 was also taken from Observation 2. In this example, Jim was asking the
whole class a question, which was to name a sense that people could not use despite the fact that
they had the organ for it. One of the students replied.
Example 33
T: … Can you think of any case where there is a real stimulus, but we don't have an
organ that can work to pick it up.
S: /Mag-ne-to-ception./
T: Magnetoception. Like we talked about before with that little piece of iron in our
body …

229

In this instance, the student’s pronunciation was choppy, and was not natural, which
elicited a recast from Jim. This one also seems to be simply a repetition of the students’
response, but he had an intention of CF. He recalled, “he said, /mag-net-o-ception./ And, I didn't
want that to be let go … obviously that's not … how you say that word. So … I felt like its real
pronunciation should come out.” An example from Observation 3 is shown in Example 34 in
which a student was making a short presentation. She was explaining the mechanism of rowing a
boat using the concept and terminology of physics.
Example 34
S: Work on these, and the, the water work also on this boat and so made the boat move
forward and the so, I guess it's the, /kanegic/, /kanetic/,
T: Kinetic, yeah.
S: Yeah.

While she was talking, she found herself having difficulty in pronouncing the word,
kinetic. She mispronounced the word twice and then looked at Jim, asking for assistance.
Prompted by the student, Jim gave a recast with the correct pronunciation. He recalled, “I was
just thinking that um, her pronunciation was too far … they wouldn't understand kinetic from
/kinesic/ … /ka, kinesic/, and so I thought I had to say.”
As can be seen in these recalls above, his explanation was short and straightforward.
He basically said that he gave the correct answer because there was an error that would interfere
with comprehension. Judging from the fact that he used more explicit correction than
clarification requests, instead of using output-prompting types of CF he preferred to give CF
directly following the error.

230

Another point that he mentioned regarding CF was that he used CF for the benefit of
other students, a point also mentioned by John. Example 35 was taken from Observation 2, in
which a woman asked a question to clarify the difference between nociception and
proprioception.
Example 35
S: Ah, what's the difference between noceptio, noception and properception.
T: Nociception and proprioception?
S: Yes.
T: Ok. Proprioception doesn't involve pain …

In her utterance, both of the words were wrongly pronounced. He immediately used a
recast in the form of a question to clarify what she meant, but his recall showed that he
understood that she meant these two words and used his utterance as a recast for other students to
hear the correct pronunciations rather than wrong ones. He recalled:
Because when she asked the question, she got the words wrong. Which is why I
repeated the words out loud to the class. It wasn't just that her voice was soft although
it was soft … she said something like um, noception and ah properception or
something like that. And so … I wanted to repeat the correct pronunciations for the
class so that … their input wasn't screwed up.

Another example taken from Observation 2 shows the same pattern. In Example 36,
Jim was asking the whole class a question to elicit the name of the stimulus for the sense of
smell. A student answered the question by saying, fume.

231

Example 36
T: How about ah, smell. What's the stimulus.
S1: Fume.
T: Fumes? Very good. Good word. ((the teacher writes the word on the board)) We can
say fumes, what i, you know this word, [student's name], and I'm not sure that
everybody knows this word. Can you explain this word?
S1: No.
T: ((laugh)) Ok. Anybody else know this word who can explain it?
S2: The odors?
T: It, ok. Fumes are odors. Fumes actually are currents of an odor …

In this case, although S1 said the word, fume, he was not able to respond to Jim’s
request for explanation. After S1 indicated that he could not explain what the word meant, Jim
asked the whole class again for a volunteer. Then, another student (S2) said, “the odors.”
Although the meanings of these two words, fumes and odors, were overlapping to some extent,
Jim wanted to clarify the difference between them. Thus, he used explicit correction after
acknowledging that these meanings were similar by saying, “fumes are odors.” He recalled, ”it
was getting toward that definition, but I didn't want them to walk out of the room, thinking that
every odor was a fume. And so, that's why I proceeded with the definition.”
In these cases, CF was given for the benefit of other students. But when giving CF and
the benefit for other students seemed to conflict, he prioritized the latter over the former. Toward
the end of Observation 2, a student was called upon to read a paragraph in the textbook that the
students were going to discuss in the next class. In the reading, he often stuttered and

232

mispronounced many words. However, Jim did not interrupt his reading. After the reading was
finished, he gave explicit feedback on only one of the words that he thought would be crucial for
the discussion in the next class. At Stimulated Recall 1, he recalled:
[The other words] weren't absolutely essential terms for our future discussion. And, I
really don't like to break up fluency when I ask somebody to read something, My, my
main point is, we get the students reading it at, at a slow enough pace because it's been
paced by a student reader, And if I interrupt him, everybody’s comprehension gets,
gets chopped up, and it doesn't, doesn't proceed well enough.

In conclusion, like John, Jim’s stated beliefs on CF did not show any fluctuation at all,
and his classroom practices of CF accorded with his stated belief on CF that he preferred to give
CF in a direct way. It should be noted that his use of CF was also for the sake of its potential
benefit for other students. Thus, for him, CF was not only for the student who made the error, but
also for the other students who were hearing the CF. However, for some reason, he did not
mention this stated belief until the stimulated recalls.
Providing CF for the benefit of other students is identical to one of John’s reasons to
give CF (Cecile also mentioned this point, but she did not make such a comment in relation to
any concrete incident of CF); however, whereas John’s giving CF for the other students was
mainly for comprehending the utterance that originally included an error for the sake of
communication, Jim’s giving CF for the other students was mainly for providing the whole class
the correct input instead of incorrect input. This tendency was evident also in the fact that
whereas John gave CF mainly when there was a communication breakdown, Jim sometimes gave
CF even when was no apparent communication breakdown, which made his recasts often sound

233

like he was merely repeating his students’ utterances. This was probably derived from the
difference in their main purpose of giving CF: communication for John and linguistic accuracy
for Jim.
Having said that, Jim did not seem to have any deep thoughts on CF. At his recalls, his
explanations for giving CF were mostly short, and his logic was simply to give the correct
answer because there was an error. Additionally, although he believed in the effectiveness of CF,
as described in the previous chapter, he called himself “not the kind of person who gives a lot of
oral feedback anyway” (Interview 2). He had other objectives and goals in his lessons which
were more important than giving CF. Thus, despite his being aware of its effectiveness, CF was
not something that he would ponder. It was rather something that he would just “naturally” give
as a response to an error in his role as a teacher.
Finally, it needs to be explained why the reading did not seem to have an impact on his
stated beliefs or classroom practices of CF. The most plausible reason is that he did not see any
necessity to reconsider any of his stated beliefs or classroom practices of CF because he had
already established them throughout his rich L2 teaching career, and he was already feeling
confident in and comfortable with his use of CF. As time went by, his stated beliefs on CF
became robust, regardless of the depth of his thoughts on CF (Nisbett & Ross, 1980; Pajares,
1992). Therefore, he agreed with Lyster and Saito (2010) whose arguments were congruous with
his stated beliefs that had been solid even before reading. Since this research supported his
existing stated beliefs, there was not reason for him to change as a result of the reading.

6.4 Tom
Table 13 shows the number and the percentage of error-filled AS-units, those followed

234

by CF-filled AS-units, those not followed by CF-filled AS-units, and each kind of CF provided
by Tom. On average, an error-filled AS unit occurred approximately every 1.1 minutes, 2.0
minutes, and 1.7 minutes (Observations 1, 2, and 3, respectively), which was almost as frequent
as in John’s class. Like John, Tom’s class also had plenty of time for group discussions, and his
students made quite a few utterances as well as errors. Also, his use of CF was every 8.3 minutes
(0.12 CF per minute), 4.5 minutes (0.22 CF per minute), and 4.5 minutes (0.22 CF per minute),
respectively. On average, his rate of giving CF was less than 0.2 per minute, which was much
smaller than all of the 12 teachers observed in Loewen (2003).
Table 13
Tom’s Use of CF at Three Observations
Error-filled AS-units
Kinds of CF-filled AS-units
Without
With
Obser
Clarification
Explicit
Total
CF-filled
CF-filled
Total
Recasts
vations
requests
correction
AS-units
AS-units
1
46
40 (87.0)
6 (13.0)
6
5 (83.3)
1 (16.7)
0 (0.0)
2
25
15 (60.0)
10 (40.0)
11
6 (54.5)
5 (45.5)
0 (0.0)
3
29
20 (69.0)
9 (31.0)
11
8 (72.7)
3 (27.3)
0 (0.0)
Total
100
75 (75.0)
25 (25.0)
28
19 (67.9)
9 (32.1)
0 (0.0)
Note. All the classes lasted approximately 50 minutes. The numbers in parentheses show
percentages. The number of the error-filled AS-units with CF-filled AS-units may not
correspond with the total number of CF-filled AS-units because a single error-filled AS-unit may
elicit multiple CF-filled AS-units.
Tom had firm stated beliefs on CF from the outset of the present study, which was not
to use the explicit types of CF. He preferred to use recasts because he thought that his students
wanted to know the right answers. Table 13 supports these stated beliefs in the sense that, on
average, (a) he used CF only for 25.0% of the error-filled AS-units, and (b) 67.9% of his CF
were recasts and the remaining 32.1% were clarification requests, which could also be used in
the flow of natural conversation.
Nevertheless, it should be noted that Tom did not necessarily believe in the

235

effectiveness of recasts. Such stated beliefs coincided with Truscott (1999), with which he
agreed the most. Even through the reading, his reluctance to give the explicit types of CF was
stable and did not seem to fluctuate at all. He expressed his confidence and stated at Interview 2
that the reading did not have any impact on his stated beliefs or classroom practices of CF.
What is noteworthy is that in line with his stated beliefs, Tom never used explicit
correction during the observations, and at every Observation over half of the error-filled
AS-units were not followed by CF. However, he still used recasts and clarification requests.
Clarification requests were used when there was a communication breakdown in which he could
not understand what the student said. At Stimulated Recall 1, he stated that he tended not to give
CF “except in cases where I can’t understand what they’re saying … it makes difference.” Thus,
he was aware of his use of clarification requests. Example 37, taken from Observation 2, shows
such a case in which students in groups were discussing who were cultural heroes in the US. One
of the students adduced Warren Buffet.
Example 37
S1: /Baffeti./
T: What?
S1: /Baffet./
T: Who, who?
S2: /Baffet. Baffet./
S1: /Baffet./
T: Buffet. Warren Buffet. Warren Buffet. Yeah. Yeah. Yeah.

In this case, two students mispronounced the last name, Buffet, a few times. Being

236

unable to identify whom they were mentioning, Tom used clarification requests twice. Finally,
when he understood the name, he used recasts with the correct pronunciation. At Stimulated
Recall 1, he stated that he “didn't understand what [the student] was trying to say.” Such a case
was also observed in Observation 3. Example 38 shows a case where the whole class was
discussing some of the differences between American English and British English. One student
adduced the spelling of the word, traveled and later, favorite.
Example 38
S1: I think traveled.
T: Tra,
Ss: Traveled.
T: Traveling?
Ss: /Virorate/.
S1: My /favorites.
S2:

//Varorate/.

T: Favorite. Favorite. Favorite. All right? …

A student adduced traveled as an example. Tom apparently did not understand what
word the student meant, and used two clarification requests, “Tra,” and “Traveling?.” Following
them, other students spoke up, but during the interaction, another word, favorite, replaced the
original word, traveled. Again, Tom seemed unable to recognize the word. Presumably (although
not clear from the recording), the original student who pronounced traveled said favorite along
with the word, my, and this seemed to help Tom understand the word, favorite, correctly (but he
missed traveled). Then, he gave recasts three times. At Stimulated Recall 2, he stated that he was

237

“trying to understand what they are saying.” As can be seen in these examples, although the
number is not large, he used clarification requests when there was a communication breakdown.
Regarding recasts, as can be seen in Table 13, Tom’s classroom practices data show
that he used them to some extent at every Observation. At Stimulated Recall 1, he said that
recasts were a part of “natural” communication for the first time in the present study while
casting doubt on their effectiveness. This “naturalness” of recasts was identical to that of Cecile,
but not to that of John who meant “automatic and unconscious” or Jim who meant “a part of
teacher’s job.” Tom stated:
Sometimes I might say the correct thing, but, it's just … as part of a conversation, it fits
into the conversation. Yeah. We often repeat what somebody else said, so, I’ll repeat it
in a correct manner as supposed to non-correct … it's very indirect … but do they
catch it or not, I don't know, because I don't think it really has a whole lot of
effective[ness].

Although he said that his use of recasts happened “sometimes,” comparing the
percentages of the four teachers at three Observations, his use of recasts was approximately as
frequent as the other three teachers. Also, despite his dubiousness of their effectiveness, he was
aware that he used recasts in his classroom “naturally,” and those examples above show such
cases. However, these recasts can arguably be nothing but his reaction to his sudden
comprehension. In fact, he sporadically used recasts on their own without prior clarification
requests as well, as can be seen in Example 39 taken from Observation 2. In this case, Tom and a
student were talking about American heroes, and Tom broached football players who had led
their team to win by a Hail Mary pass at the very end of the game as an example.

238

Example 39
T: So, but entrepreneurs are considered to be American heroes. What about the
quarterback who threw the pass, Hail Mary pass, and the guy who caught it, were
those guys heroes at [the university’s name]? They were. Right? Yeah. Last week,
they, they're heroes this week.
S: Yeah.
T: Yeah. Because they help win the game, right? Yeah. One second. ((laugh))
S: Or less second.
T: Less than a second.

The student’s utterance was lacking than a, and Tom added them in his recast. Thus,
his use of recasts was not confined to the cases when he came to a sudden comprehension after
clarification requests, but also even when there was no apparent communication breakdown.
However, it should be noted that in most of such cases whether or not he used recasts with the
intention of CF is unknown because of the reason described in the next paragraph.
Before Stimulated Recall 1, I showed the printed instruction and read it aloud while
Tom was looking at it. However, at the recall session, he stopped the video only three times
although there were many instances of errors. At Stimulated Recall 2, I emphasized this point
and asked him to stop the video even if he did not remember anything or he was not thinking of
anything at those moments. This time, although he stopped the video more frequently than
Stimulated Recall 1, most of his comments were not related to errors, and the ones related to
errors were counted only twice (although in both stimulated recalls, I sometimes stopped the
video when an error occurred and asked him to recall his thoughts). In total, he commented on

239

nine incidents of errors and among them, he remembered his thoughts at eight of them.
Even in such a limited set of data, though, his attitude toward errors and CF
conspicuously emerged. Two examples are shown below. The first one, Example 40, was taken
from Observation 2. In this case, led by Tom, the whole class was checking a list of words and
phrases in the textbook. The phrase in focus was have something in common.
Example 40
T: … Have something in common. Do you have something in common.
Ss: /Share (?)
S: /Sharing,
T: Share the same thing. Right? So, what do we all have in common in this room.
S: Speak English,
S: No.
T: Yeah, we do. We speak English. We only speak English in this room.

In this case, a student said, “Speak English” as an incomplete sentence. Despite the
fact that Tom used recasts of this sentence twice, his recall at Stimulated Recall 1 showed that he
was not concerned about the error, and comprehension was his primary concern as shown in the
interaction below. In this case, he was the one who stopped the video:
T: Uh, there was a mistake there, speak English,
R: So did you, did you notice in the class or right now?
T: Yeah. No. Uh, I notice but I don't pay any attention to it,
R: So you noticed in the class?
T: Yeah.=

240

R: =Ok. /Then,
T:

/But … it's understandable. So I just keep going … what I'm trying to do is not
correct for errors and [I’m] trying to get them to see whether [they] understand
what's going on,

R: Ok. So you're thinking about like the /answer (?),
T:

/Yeah. The, the answers and not the
pronunciations and so forth. As long as it's understandable, it doesn't matter.

A similar instance was observed also in Observation 3. In Example 41, Tom asked the
students in a group to say their birthdays. Then a student asked another student the same
question, and as Tom heard her answer, he made a joke by saying that was the best birthday in
the world because he found out that she had the same birthday as his.
Example 41
T: This is the best birthday in the world.
Ss: Why.
T: Take a guess.
S1: /Ah,
S2: /September fourth.
S1: September /first is /Septem,
S3:

/Septem,

T:

/Why would it be the best birthday in the world?

S4: Same you?
T: Yeah. See?

241

S4 made an error by saying, “same you,” lacking the before same and as between the
two words. Tom did not give CF, and the conversation carried on. This time, it was I who
stopped the video, but his recall at Stimulated Recall 2 vividly shows his thoughts on errors:
R: Did you, did you notice the /error?
T:

/Yeah. But it’s uh, it’s uh, not a, I mean, doesn't hinder
communication. So I don't worry about it.

R: So you were thinking about oh he got it? /((laugh))
T:

/Yeah. He understood. Yeah. He is the one
who'[s] got it. Right?

As he mentioned above, his primary concern in his class was communication. After
Simulated Recall 1, when asked whether or not he noticed more errors than the ones on which he
made comments, he replied that he did, but then added, “I tend to ignore them” twice.
Furthermore, he stated:
The only thing … I'm worried about is, is in these kinds of situations is
communication, can get their point across. So, that's all I care about. So I don't really
think about the errors themselves … communication is what I'm interested in … what
we were trying to get at, what are they, they've listened to a lecture, so now, it doesn't
matter to me that they don't say it, as as, absolutely correctly or … they don't use the
exact grammar.

At Stimulated Recall 1, he also stated, “if I can understand what they’re saying, there’s

242

no need for me to, to give anything.” Then, there is no wonder why he would not give CF
because CF “disrupts the flow of the communication” (Stimulated Recall 1). Obviously, his
focus in his class was communication, and as CF would hinder it, it was natural for him not to
give CF.
His stated belief regarding non-use of CF except recasts and clarification requests as
described above was quite adamant and also consistent. Nonetheless, he admitted that he would
sometimes use CF unconsciously partly because he used to use CF to a large extent. However, as
he learned to apply accelerated learning and brain-based learning, both of which were adduced as
reasons for his reluctance to give CF, he said that he came not to give CF as much as he used to,
and especially the explicit types of CF. Still, he stated that when he found himself using CF, he
would regret doing it, which was apparently the reason why he tended not to give CF
“purposefully” (Stimulated Recall 1) as CF is “just not necessary” (Stimulated Recall 1). At
Stimulated Recall 1, asked whether or not he would ever give CF, he said:
I do sometimes … there're occasions. I don't like to do it, but sometimes I do it because
I've done it in the past … used to do it a lot … But once I learned about accelerated
learning … and brain based learning things, I just started to cut back on it as much as
possible, so when I'm thinking about it, I don't do it … Consciously trying not to do.
When, but when I'm not, sometimes I'll make corrections, and … doesn't always work
or sometimes they don't even notice, as I've done it … so it's kind of a waste, so I just
try to avoid doing it, but sometimes I do.

The whole picture of his stated beliefs on CF became clear finally at this stage. He
used to give a lot of CF, but as he applied some of what he had learned on his own, he came to

243

believe that CF would not be helpful or effective for students. Thus, he had been trying to refrain
from giving CF. However, he deemed that the use of clarification requests and recasts was
“natural,” in the sense that they were used as natural responses in interactions, so he did not mind
using them even though he did not necessarily believe in their effectiveness as CF because he
thought that his students expected him to give the correct answers. The explicit types of CF,
though, were something that he would be strongly against using. Such stated beliefs on CF were
powerful and consistent, and in fact his classroom practices of CF were congruent to his stated
beliefs on CF in the sense that only 25.0% of error-filled AS-units were followed by CF, and his
use of CF was restricted only to clarification requests and recasts without explicit correction.
Similar to Jim, through his abundant English teaching experience, Tom’s stated beliefs
on CF were established, and as years went by they had become stable without any leeway for a
reconsideration. However, although they both showed their rigidity regarding CF, there was an
important difference between Jim and Tom. In Jim’s case, it seems that CF had never been one
of his primary concerns in his teaching career. It just happened to be there with him when he
taught, and had been using it without deep thoughts because it was a part of his job as a L2
teacher. In contrast, in Tom’s case CF was or had been one of his major concerns. Thus,
although he used to give CF extensively, as he gained knowledge of other disciplines, he
self-reflected on his own use of CF and gradually shifted to restrict the use of it. But his distress
lay in the fact that his “old habits die hard” in the sense that he still used the explicit types of CF
sometimes even now, which caused regrets in his mind.

6.5 Discussion

244

In this section, two major factors for the decision on CF, which became apparent from
the analyses of the four teachers in this chapter will be discussed. The present data showed that
one of the main factors that was indispensable in determining whether or not the four teachers
would give CF was communication. That is to say, one of the functions of CF that they perceived
was to help communication, and thus, in turn, CF must not hinder it. In other words,
comprehending what their students were trying to say, namely the content of their utterances and
not the language, was their main concern in their classes. In order to accomplish this goal,
clarification requests were utilized. Specifically speaking, when there was a communication
breakdown, because these teachers needed to understand what their students were saying, they
naturally used questions to elicit the students’ true meaning. Oftentimes, instead of simply saying
a phrase like “what?,” the teachers guessed what the student was trying to say, and said the
phrase with a rising intonation. When their guesses were right, clarification requests functioned
as input-providing, rendering the CF recasts. At the same time, even when the teachers noticed
an error in their students’ utterances, as long as the meaning was clear and no communication
breakdown occurred, and especially when CF was likely to interrupt the flow of communication,
they refrained from using CF. In Tom’s case, he was reluctant to give CF based on this logic that
CF would hinder communication, except clarification requests for the reason above, and his
natural use of recasts.
Another factor that was playing a major role in regards to CF, especially in the case of
recasts and explicit correction, was the function of CF for the benefit for other students, which
actually had twofold meanings: (a) other students’ comprehension and (b) positive evidence for
other learners. The former point was mentioned by Cecile and John. Unfortunately, although
Cecile expressed her concern about other learners’ comprehension, according to her recall, there

245

was no case when she was able to identify an instance in which she succeeded in giving CF
consciously for other students’ comprehension because in one particular case (Example 22), she
was unable to provide the correction as she also could not figure out what the student really
meant after her attempt with several clarification requests. In contrast, John’s class showed many
instances when he consciously used CF for the sake of other students’ comprehension. When he
suspected that there would be a communication breakdown between the speakers who made
errors and their listeners even if there was no communication breakdown between John and the
speakers who made errors, he often gave CF.
The latter point was also mentioned by John himself, but more explicitly and
recurrently by Jim. One of Jim’s remarks, “I didn’t want them to walk out of the room, thinking
that every odor was a fume” graphically represent his thoughts. Such comments also occurred
regardless of whether or not there was a communication breakdown. Jim seemed to favor recasts
as can be seen in Examples 32, 33, and 35 (the pronunciation of taste, magnetoception, and
nociception and proprioception, respectively). These two major factors for the decision on CF
explain why many of the errors that were followed by CF were meaning-based, rather than
form-based as can be seen in the examples above (e.g., Examples 19, 26, 36, for Cecile, John,
and Jim, respectively).
In summary, this chapter examined the four teachers’ stated beliefs and classroom
practices of CF posterior to the reading. It was found that their decision on CF was based upon
two factors: communication and the benefit for other students. For the former, CF was used
when there was a communication breakdown, but CF was not used when it would have been
likely to hinder communication. For the latter, CF was used in order to help other students
understand the utterances of the speakers as well as to let other students be exposed to the correct

246

input.

247

Chapter 7: Discussions and Conclusions

This final chapter consists of four sections: (a) discussions overarching the whole
study, (b) limitations of the present study and ideas for future research, (c) recommendations to
stakeholders, and (d) conclusions.

7.1 Discussions Overarching the Whole Study

The main focus of the present study was to investigate how the four ESL teachers
would respond when they read three research studies about CF. In relation to that, their teaching
philosophies, their stated beliefs on CF, their classroom practices of CF, as well as the
relationship among these three were examined.

7.1.1 Responses to Each Type of Research Study
In the present study, the four teachers read three research studies, which were
intentionally chosen so that each of them would represent a different type of research: a
meta-analysis, a single experimental study, and a position paper. Such a categorization seems to
be one of the factors that affected how the teachers responded to each of these research studies;
thus, I believe it worthwhile to discuss their responses from this perspective.
As discussed in Chapter 5, a single experimental study did not seem to attract these
four teachers’ attention. None of the four teachers criticized Mackey et al. (2000) from this
standpoint; however, the evidence for their lack on interest in Mackey et al. can be found from
the comments made by John and Jim on Lyster and Saito (2010), which was that a meta-analysis

248

would make the results and the claims of the study more convincing than a single study because
the former was a collection of studies. John even explicitly stated that he “could find holes” in a
single experimental study. However, such features of a meta-analysis were regarded as beneficial
only by John and Jim. Neither Cecile nor Tom seemed to notice or care about the difference
between a single experimental study and a meta-analysis. Therefore, the power of a
meta-analysis is not universal; it can exert itself only on certain teachers.
Interestingly, the position paper, Truscott (1999), was favored by both Cecile and Tom,
who did not acknowledge the power of meta-analyses. Neither of them seemed to care whether
or not the claims made in the study were supported empirically. They chose Truscott (1999)
based on the fact that his arguments resonated better in them because the claims that Truscott
made were basically in line with their stated beliefs on CF. In the meantime, it is important to
acknowledge that position papers without statistics are easy to read, which renders them more
accessible to teachers. Here, it should be remembered that none of the 13 teachers in Zeuli
(1994) understood the statistics used in the research study that they read. Furthermore, the fact
that Truscott was written from a teachers’ perspective may have even facilitated the accessibility
of the article. In Zeuli, while those teachers who had an experience of research believed in
indirect impact that research studies had on their stated beliefs and classroom practices, those
teachers who did not have such an experience believed in direct pedagogical implications. In this
sense, in the present study, Tom, who was the only teacher with an experience of publishing
research studies, should be expected to show a different pattern than the other teachers.
However, the results did not conform to this expectation. Although John and Jim did not
appreciate Truscott as much as Cecile and Tom did, John admitted that it was “the most
entertaining,” and “enjoyed [it] the most as far as just reading.” Being accessible to teachers may

249

be the most crucial, and yet unfulfilled task to be accomplished as the first step to close the gap
between researchers and teachers (S. Borg, 1998a; Ellis, 1997). In this sense, a position paper
can potentially be a great resource to have an impact on teachers.

7.1.2 Stated Beliefs on CF
The results showed that among the four teachers, only Cecile showed a fluctuation in
her stated beliefs on CF. This is in line with the findings in Mok (1994), in which the effect of a
practicum (not reading) was larger for the inexperienced teachers than the experienced teachers.
At the beginning, Cecile was not able to articulate her stated beliefs on CF. She seemed to lack
any firm thoughts on CF. She had two years of teaching experience in a foreign country, but this
not only did not help her establish her stated beliefs on CF, but also seems to have worked
against her building them up because she had been provided with conflicting perspectives on CF.
Nevertheless, she started to form her beliefs on CF while participating in the present study. Her
initial shift was indeed initiated by reading the three research studies for the present study in that
reading facilitated her reflection on her stated beliefs and classroom practices of CF. She mainly
adopted the view of Truscott (1999), and stated that she would give less CF, yet such stated
beliefs were still vulnerable. Accordingly, after a month, her stated beliefs on CF showed another
swing, which was caused by multiple sources outside the realm of the present study. Thus,
reading research studies as well as the coursework in her program had a large impact on her
stated beliefs on CF, similar to the teacher in Rankin and Becker (2006).
In contrast, in line with Shkedi’s (1998) findings and Clark (1988)’s claim, the stated
beliefs among the other three teachers largely remained intact, which is contrary to the claim
made by Bullough and Baughman (1993). John seemed to possess solid beliefs on CF from the

250

beginning of the present study, most of which were already implicit and unconscious to him at
the beginning. Through reading the three research studies, he raised his consciousness of CF, and
became able to articulate his beliefs on CF more clearly and explicitly than before, but the
reading itself did not seem to have a large impact on his stated beliefs. Jim also had a clear policy
of CF, which was to provide the correct answer possibly with an explanation. To him, CF
seemed to be a topic that he had not thought deeply about, and he was giving CF as a role of a
teacher. Reading the research studies only worked to support his stated beliefs in the
effectiveness of CF, following Lyster and Saito (2010). Finally, Tom also had strong stated
beliefs on CF, which was basically the negation of the explicit types of CF. This notion had been
formed through his previous L2 learning and teaching experience as well as his knowledge of
brain-based learning and child L1 learning. His stated beliefs on CF were so strong that reading
research studies did not have any effect on them. In Zeuli (1994), teachers were interested in
“trying to further understand concepts and evidence when a study’s conclusion did not match
their own beliefs” (p. 52) and “thinking more about whether their interpretation was warranted”
(p. 52). Although the three teachers who participated in the present study also showed such a
tendency to some extent, the influence of the reading on their stated beliefs on CF was not
extensive.
A factor that obviously divides Cecile from the three other teachers is the amount of
L2 teaching experience. Although she had had two years of English teaching experience
overseas, she was still relatively inexperienced and the other three teachers were much more
experienced than her, having at least 14 years of teaching experience. Thus, judging from the
difference in their responses to the three research studies, it seems legitimate to assume that the
amount of L2 teaching experience might play a role in how much impact reading research

251

studies can have on the stated beliefs on CF among ESL teachers. However, the amount of
teaching experience should not be counted as the direct cause of this effect. It is what the
teachers gain gradually through teaching experience, such as knowledge, skills, confidence, and
of course, beliefs, the last of which becomes robust and less amenable to fluctuation as time
passes, rendering the power of reading research studies weak (Nisbett & Ross, 1980; Pajares,
1992). In contrast, when teachers are relatively inexperienced, their stated beliefs on CF are still
vulnerable and maybe even amorphous; thus, the influence of reading research studies may tend
to be large.
Having said that, regardless of their amount of teaching experience, it is worthy to
remember that these teachers were highly selective as to which studies to identify with. In cases
of experienced teachers, they were basically showing their favor for the studies whose claims or
findings were in line with their stated beliefs prior to reading. John and Jim believed in CF, and
accordingly they chose Lyster and Saito (2010). Tom was reluctant to give CF, and accordingly
he chose Truscott (1999). In Cecile’s case, despite her calling herself a “sponge,” her “sponge”
had a filter and took only Truscott (1999) to her heart.
In conclusion, the effect of the reading on teachers’ stated beliefs on CF was limited; it
mainly helped them articulate their stated beliefs on CF in case when they were implicit and
unconscious. In this sense, what was observed in the present study was impact, rather than
change (S. Borg, 2011). It should be noted that it was the opposite direction of the effect that was
stronger; their choice of the research studies to identify with was largely affected by their stated
beliefs on CF prior to the reading.

7.1.3 Classroom Practices of CF

252

Because their stated beliefs on CF were stable, it is understandable that the teachers’
classroom practices of CF also stayed largely intact throughout the three observations. This is in
line with the results in Mackey et al. (2004) in the sense that, while the teachers’ awareness of
CF was raised through participation in a workshop, their classroom practices were stable.
Despite her fluctuating stated beliefs on CF, even Cecile consistently used recasts and
clarification requests. John and Jim used CF within the realm where it would not interrupt the
goals of their class. Tom did not let the reading affect him, and restricted his use of CF to the
natural use of recasts and clarification requests.
One reason why the reading did not have any effect on their classroom practices of CF
is that all four teachers had other things in mind to focus in their class as discussed in Chapter 4.
Therefore, CF was not the main focus of their class, but rather a teaching technique to deal with
errors. In the program for which they worked, CF was basically optional in the sense that it was
up to each teacher whether or not to use it and, if they decided to use it, how they would use it.
Therefore, CF was so marginal in their class that they did not see any necessity to reconsider its
use.

7.1.4 Participating in Research as an Opportunity for Self-Reflection
Although their stated beliefs and classroom practices of CF mostly remained the same,
there was one effect of the reading mentioned by the teachers; they became more conscious of
CF than before. This is in line with a statement in Freeman (1989), which was “change does not
necessarily mean doing something differently; it can mean a change in awareness” (p. 38).
Although their comments of heightened consciousness were garnered at Interview 2 as a
response to reading research studies, their consciousness of CF was probably raised also by

253

participating in the present study through such tasks as reading research studies, watching their
own teaching video, recalling their thoughts while teaching, and the most importantly, reflecting
their own stated beliefs and classroom practices of CF using these resources, which C. Kennedy
(1999) listed as one of the strategies to change the behaviors of teachers. For example, at
Stimulated Recall 1, when asked whether or not she noticed an error in Observation 2, Cecile
replied, “I think I probably noticed it, to be honest, maybe because you were in there, and you
were, I know, that you’re focusing on error correction, so maybe I was more aware of those
kinds of things.” As Weinstein (1999) noted, “[r]eviewing video- or audio-tapes of their own
teaching and engaging in stimulated recall, discussion, and analysis may promote an awareness
of the cognitive demands of teaching, as well as students’ own implicit premises” (p. 287).
Such a phenomenon was also observed in Duff and Uchida (1997), in which the
researchers were able to perceive some changes in cultural awareness among teachers caused by
the participation in their study. In Burns (1996) the ESL teacher whose stated beliefs and
classroom practices did not match became aware of this discrepancy through “close engagement
with data from her own classroom” (p. 174), which led her to successfully adjust her classroom
practices so that they would accord with her stated beliefs toward the end of the study. Burns
called this process as “an enlightening process and a significant means of professional growth”
(p. 174) and claims, “increased awareness through close and critical engagement with classroom
data can be a catalyst for the re-evaluation of thinking and beliefs and for the development of
alternative mode of instruction” (pp. 169-170).
Among all the components in the methods for this study, two of them probably
contributed the most to teachers’ reflections on their stated beliefs and classroom practices of
CF. The first one was the reading of the research studies. Especially, Cecile was the one whose

254

stated beliefs on CF were beginning to emerge at Interview 2 although she could not formulate
them at Interview 1. John’s beliefs on CF were mostly implicit at Interview 1, but he came to be
able to articulate them at Interview 2 possibly by raising his thoughts on CF from an unconscious
level to a conscious level. Jim was much more experienced, and did not seem to be influenced by
the reading, but he stated that he would be more conscious of CF, which Cecile and John also
mentioned. Tom stated that reading let him relearn and reevaluate his vested knowledge. In this
sense, the reading itself, probably not just on CF but also any kinds of research studies, seems to
have a potential for teachers’ reflection however subtle it may be.
The second component is, as Weinstein (1999) and Swain (2006) proposed, the
stimulated recalls, which seem to have facilitated teachers’ reflections on their own classroom
practices. Swain argues that stimulated recalls are “a process of comprehending and reshaping
experience – they are part of what constitutes development and learning” (p. 110). Cecile became
aware that she was focusing on content rather than language and that she was not providing her
students with the correct answers. John realized his general pattern of giving or not giving CF,
distinguishing the whole class situation and the small group situation. Neither Jim nor Tom made
such direct comments. Still, Jim stated that one of his purposes of giving CF was to provide the
correct input to other students for the first time at Stimulated Recall 1, but he did not mention
such a point at Interview 2. This may be because watching his own teaching had given him an
opportunity to observe his teaching somewhat objectively, facilitated his rethinking of his
rationale to give CF. Tom ascribed his use of recasts to their “naturalness” for the first time at
Stimulated Recall 1, possibly for the same reason as Jim. It is important to mention that I did not
mention any of these points; the teachers came to notice them on their own. Thus, it is possible
that through such reflections, stimulated recalls can be utilized as a self-professional

255

development tool.

7.2 Limitations and Ideas for Future Research

The present study has four conceivable limitations. The following paragraphs will
describe each of them along with some ideas for future research derived from them.
First, in the present study, teachers’ stated beliefs on CF prior to reading were explored
at Interview 1, whose purpose was to collect background information from the four teachers, and
CF was only one of the topics to be asked about. I had deliberately decided not to pursue the
topic of CF more deeply at Interview 1 because I did not want the teachers to become aware of
the focus of the present study until they read the research studies. However, through analyzing
the data, I found that this type of data collection limited both the quantity and the quality of the
data obtained regarding their stated beliefs on CF. Spending more time focusing on CF at
Interview 1 would have been an option to solve this problem. For example, introducing each type
of CF, I could have asked what the teachers would think of each of them. Also, because the data
after reading was collected through stimulated recalls, if another simulated recall had been
conducted prior to reading, the data between the three stimulated recalls would have been
comparable, which may have been more illuminating. This would have been especially helpful in
the case of recasts and clarification requests because these two types of CF can be provided as a
natural flow of a conversation without any intention to correct errors. Moreover, as shown in the
present study, stimulated recalls may elicit teachers’ stated beliefs in more detail than interviews
possibly owing to the use of the stimulus in the form of recording teachers’ own classroom
practices. Investigating teachers’ stated beliefs on CF prior to the reading using another

256

stimulated recall could have produced another set of data different from that obtained from
Interview 1 because their comments may become more detailed as watching their own classroom
practices of CF would have facilitated their deeper reflection upon their stated beliefs on CF.
Second, in the present study, the topic of the reading, namely CF, had been decided by
me prior to the data collection without the teachers’ input. I would like to claim that this was a
reasonable choice because reading the common research studies on the same topic made it easy
to compare and contrast the four teachers. However, it was found that CF was not the main
concern of the four teachers; thus, there was a mismatch between what the teachers really wanted
to know and what they were told to read about. Admittedly, this is one thing in the procedure that
needs to be considered for a possible change for future research. As Pennington (1996) states,
“attempts to influence teacher’s behaviour will have an impact only in areas where the input is
valued and salient to the individual, and where it is congruent with, and interpretable within, the
teacher’s own world of thought and action” (p. 340). Therefore, selecting a topic that the teacher
finds the most interesting would more likely yield the strongest impact on their stated beliefs and
classroom practices, rendering the findings of the studies accessible input which will turn into
intake in the teacher. For example, CF was the topic of the reading in Rankin and Becker (2006)
as well, but the decision to focus on CF was made through agreement between the researcher and
the teacher as they thought that CF “would be mutual interest and worthy of collaborative study”
(p. 356). This could be one way to approach this issue.
Third, among the four teachers who participated in the present study, Cecile showed a
quite different pattern from the other three teachers regarding stated beliefs on CF. I have
explained this from the perspective of her being an inexperienced teacher, but this can also be
approached from different viewpoints, and two of which are (a) occupation, and (b) status in the

257

program. Cecile was an exceptional case in these two points because (a) she was the only
student, and (b) she was the only TA. For (a), being a student, she had many occasions to be
exposed to different concepts about teaching in the classes that she was taking. This may have
made her more receptive to new ideas presented in the readings. For (b), being a TA may have
made her feel uncertain about her beliefs about teaching because she was treated as a trainee. It is
also possible that multiple factors like these are playing a role in Cecile’s uniqueness. In any
case, it is impractical to draw any conclusion from such a small pool of participants. This issue
awaits further investigations.
Finally, it is unknown to what extent these four teachers were of typical examples of
L2 teachers or even the teachers in the program for which they were working. These four
teachers decided to participate in the present study despite their busy schedule. Were they
interested in the participation in a research study by itself? Or, as Tom said, were they willing to
learn something new, possibly from research studies? Or was it simply due to their personal
attachment to the researcher? Regardless of the answers to this question, it is perhaps needless to
say that the majority of L2 teachers are not willing to participate in research studies, rendering
these four teachers rather exceptional in the sense that they were more dedicated to my research
than others.

7.3 Recommendations to Stakeholders

Jim ascribed his disagreement with Truscott (1999) to a different teaching context
between them. This is not only a legitimate argument to explain another contextual factor that
influences the use of CF, but also evidence that some teachers may dichotomize L2 classrooms

258

between teachers’ and researchers’, implying that researchers may never be able to understand
the L2 teaching contexts where the majority of L2 teachers are struggling. At Interview 1, John
stated:
My problem with research articles is they might be published by people who have very
little experience in an ESL class … they've been a professor for twenty-five years, But,
how long has it been since you've been in an ESL classroom. That dynamic is
changing. Even since you've taught here, man, that has changed. The dynamic is
different in our class, right? Completely different. So, I've a problem with that. And I
can sense that.

Overall, it seems that these teachers were perceiving the two worlds to be disparate:
the teachers’ and the researchers’. Whether or not we can bring these worlds together depends on
our efforts. In the following sections, some recommendations for the four stakeholders,
researchers, pre-service teacher educators, in-service teachers, and in-service teacher educators,
based on the results of the present study will be argued.

7.3.1 Researchers
Although many researchers are still teachers in the sense that they still teach
something, which may not be L2s, I believe teachers and researchers are two distinct occupations
working in different fields. If so, it is quite natural for each of them to live in a world where they
communicate in a way that is effectual only in that community. Teachers talk in the language
that teachers can understand whereas researchers talk in the language that researchers can
understand. I deem that this is necessary in order for each of the professions to prosper and grow

259

because knowledge in profession is usually cultivated on the accumulation of prior knowledge.
Such prior knowledge is shared by those people in the field and is indispensable to create new
theories and practices; however, those people outside the field are not familiar with such prior
knowledge. If those people in the field avoid using such prior knowledge, the efficiency of their
work will be severely hampered. Thus, for example, in a professional research journal, it would
be more fruitful to write findings and discussion than to explain what k sample is because most
of the readers should know what k sample is.
But at the same time, it would be unfortunate if all the efforts that researchers are
making have no impact on teachers’ development and better teaching. It is true that not all of the
findings in SLA research can be applied to pedagogy. Furthermore, even many of those that can
be applied to pedagogy have only indirect implications. Yet, hopefully, anybody, including the
researchers who are not mindful of pedagogical implications of their own studies, wishes for
better education for all learners. The question is how. The small section dedicated for the
pedagogical implications affixed at the end of an article seems to be of little value for now as L2
teachers do not normally read research studies in the first place (e.g., S. Borg, 2007a; Crookes &
Arakaki, 1999; Shkedi, 1998) or find reading research studies helpful for their professional
development (Pennington & Urmston, 1998; J. Richards et al., 2001).
Researchers are often educators of pre-service teachers and in-service teachers. I would
save my recommendations for those teacher educators for the following sections; therefore, in
this section, I would like to present an idea that is applicable for all researchers. My
recommendation would be to find a way of providing simplified summaries of major research
studies that are relevant for teaching. In the first place, if teachers do not try reading research
studies, the impact of the research studies on practices would be quite limited. Yet, teachers are

260

often too busy to read them (S. Borg, 2007a; Crookes & Arakaki, 1999). In order to make
research studies accessible to teachers, the simplified summaries of research studies that are
relevant for teaching should be provided for teachers. A possible solution is to compile the
abstract of each research study into a document. As teachers read the document, they may find
some research studies interesting, and decide to read the original articles.

7.3.2 Pre-Service Teacher Educators
Pre-service teachers, such as MA TESOL students, need a great amount of support
from their mentors in the program to which they belong. These teacher educators decide the
curriculum that they follow and guide them through the program; thus, I believe pre-service
teacher educators play an important role of bridge between teachers and researchers for
pre-service teachers.
The time when they are still pre-service teachers is potentially the best period in their
lifetime to find the connection between teaching and researching. That is because they must not
only teach in the practicum but also read research studies constantly. Also, as Cecile showed in
the present study, in the first place, pre-service teacher educators need to realize that research
studies may have a large impact on pre-service teachers’ stated beliefs on CF and possibly on
other areas as well. However, reading by itself may not facilitate self-reflection; pre-service
teacher educators need to set up tasks in which pre-service teachers produce their thoughts either
orally or in a written form while trying to see connections between the research studies that they
read and classroom practices (of their own, if possible). Such tasks can be incorporated into
assignments, exams, and portfolios. If time is secured, it would be fruitful for pre-service
teachers to have a discussion to share their teaching philosophy, past experience of L2 learning

261

and teaching, stated beliefs, and their relationships in response to the research studies that they
read. Moreover, in addition to reading research studies, perhaps some other procedures for
self-reflection, especially watching their own classroom practices while recalling their thoughts
should be implemented. Lastly, these data should be recorded periodically and stored throughout
their career as a graduate student. Looking back on them longitudinally also provides pre-service
teachers with materials to add to their teaching portfolios and another opportunity for
self-reflection, which may facilitate professional development.

7.3.3 In-Service Teacher Educators
The present study showed that participating in a research study that focuses on
teachers’ cognition can potentially facilitate their reflection on their stated beliefs and classroom
practices. But in-service teachers rarely come across such a chance unless they conduct action
research on their own. Thus, my recommendation would be that when in-service teacher
educators have an opportunity to work with in-service teachers, in-service teacher educators can
set up a task, ideally in the form of action research as shown in Rankin and Becker (2006), in
which in-service teachers can reflect on their stated beliefs and classroom practices using
research studies, whose procedures were described in the previous section for pre-service
teachers. In the first place, without trying reading research studies, in-service teachers may not
be able to realize the importance of research studies. However, teachers may have difficulty in
finding time to read them. In such a case, providing a summary of the article, as I did in the
present study, is one way to render the findings more accessible in a short period of time.
Furthermore, as written above, in the present study, there was a mismatch between the topic of
the readings and the teachers’ interested fields. Action research can authorize teachers to choose

262

the topic of their interest, which is more likely to motivate them to read the articles more
earnestly. Still, even if they read them, because the teaching contexts are usually quite distinct
between their own and the studies’, they cannot see any applicability to their teaching (Ellis,
2005). This is quite true, and that is why, as Kagan (1992a) claims, self-reflection on their own
beliefs and classroom practices is indispensable for them to connect these two. Especially, the
present study showed a potential usefulness of stimulated recalls even for experienced L2
teachers. Stimulated recalls are probably something that teachers never experience unless they
participate in a research study. Still, watching their own teaching via video recording is one of
the common components in pre-service teacher training programs. Once they earn a degree and
begin the career as full-time teachers, however, they seldom have such opportunities or
incentives. Being given a great resource for self-reflection, the teachers may be able to (a) raise
their consciousness of their classroom practices, and/or (b) render their implicit stated beliefs
explicit, which possibly contributes to their professional development as well.

7.3.4 In-Service Teachers
Finally, without waiting for such an opportunity to participate in a research study led
by a researcher to come, in-service teachers can conduct action research in which they utilize
research studies for self-reflection on their own beliefs and classroom practices. They may
choose to collaborate with colleagues so that they can find the topic of their common interest,
read the same research studies, and have discussions on them. Should they choose to work on
their own, they may want to keep reflective journals. Furthermore, as discussed in the previous
section, they can conduct stimulated recalls with their own teaching videos either alone or with
colleagues. They may need some guidance by professionals (e.g., researchers) regarding the

263

choice of the research studies to be read, but as the topic would be something that they are most
interested in, their motivation to read them is likely to be much higher than that in the present
study. As described in the previous section, through conducting such action research, teachers
may be able to raise their consciousness of their classroom practices and/or render their implicit
beliefs explicit.

7.4 Conclusions

The present study had three research questions: (a) the four ESL teachers’ teaching
philosophies, stated beliefs on CF, classroom practices of CF, and their relationships; (b) their
responses to the reading of the three research studies; and (c) the effect of the reading on their
stated beliefs and classroom practices of CF. Let me briefly summarize the answers to these
questions in relation to some of the previous studies in the following paragraphs.
Regarding the first research question, in line with the previous studies (Bailey et al.,
1996; M. Borg, 2005; S. Borg, 1999a, 2009b; Calderhead & Robson, 1991; Duff & Uchida,
1997; Eisenstein Ebsworth, & Schweers, 1997; Farrell, 1999; Freeman, 1992; Golombek, 1998;
Goodman, 1988; Holt-Reynolds, 1992; Hsiao-Ching, 2000; Johnson, 1994; Kagan, 1992a; M.
Kennedy, 1990; Legutke & Ditfurth, 2009; Mok, 1994; Nespor, 1987; Numrich, 1996; Powell,
1994; Sato & Kleinsasser, 1999), the present study showed that, in addition to the four teachers’
teaching philosophies, their prior language learning and teaching experience had a great impact
on their stated beliefs on CF. The former was what Lortie (1975) called the “apprenticeship of
observation” (p. 61) and mentioned in Duff and Uchida (1997), and the latter was mentioned in
Duff and Uchida (1997) and Nespor (1987). The findings in regards to whether or not stated

264

beliefs and classroom practices correspond to each other differ in each previous study, but the
four teachers in the present study showed consistency between their stated beliefs and classroom
practices as shown in some of the previous studies (Bailey, 1996; Brumfi et al., 1996;
Hsiao-Ching, 2000; Johnson, 1992a; Mitchell et al., 1994a, 1994b).
Regarding the second and third research questions, the small number of previous
studies regarding the impact of reading research studies have depicted a fairly strong influence of
research studies on teachers’ stated beliefs and classroom practices (Farrell,1999; Rankin &
Becker, 2006; Vásquez & Harvey, 2010; however, it should be noted that in Rankin and Becker,
the teacher had already committed to changing his way of providing CF at the time of the
reading the research studies). The present study showed that, although the reading research
studies (not only the ones assigned in the present study but also another assigned in her MA
class) had a great impact on the novice teacher (i.e., Cecile), the effect on the three experienced
teachers was not as large as that for the novice teacher. There is no study yet that investigated
this topic taking teachers’ teaching experience into account; however the results in Bartels
(2003) and Shkedi (1998) imply that the effect of research studies diminishes as the amount of
teaching experience accumulates.
Another noteworthy finding in the present study was that the relationship between the
reading of the research studies and the four teachers’ stated beliefs was bidirectional, rather than
unidirectional, in the sense that the teachers’ responses to the studies was heavily influenced by
their stated beliefs prior to reading. This is in line with the claim made by S. Borg (1999a), who
called it a “reciprocal relationship” (p. 122). Such a phenomenon was observed even in the
novice teacher (i.e., Cecile) who did not seem to have any firm stated beliefs on CF at the time of
the reading.

265

The findings of the present study, which sought a way to connect research and practice,
poses a paradoxical question to the field of applied linguistics, namely, whether or not
pedagogical implications, that are often criticized for their impracticality (Han, 2007; Lightbown,
1985; Magnan, 2007), should still have a place in research studies. Even with those pedagogical
implications, owing to the different teaching contexts between the teachers’ (i.e., readers’) and
the researchers’ (i.e., studies’; Ellis, 2005), “[teachers] are unable to see what published research
means for their classroom practice” (S. Borg, 2007a, p. 744). In this sense, I agree with Ellis
(1997) who claims that, although SLA research may not have a direct effect on teachers’
behaviors, it has a potential to indirectly influence them “either by helping them to make explicit
their existing principles and assumptions, thereby opening these up to reflection, or by helping
them to construct new principles” (p. 82). Ellis made a similar claim once again in an expansive
literature review for L2 teachers (Ellis, 2005). Borrowing Ellis (1997)’s terminologies, what was
mostly observed in the present study was transformation, rather than transmission, through
self-reflection on stated beliefs and classroom practices. If this is what we can and should expect,
a tiny section dedicated for pedagogical implications attached to the end of a research study often
for formality’s sake may not have a large impact on the readers who are L2 teachers. What they
need is a chance to read the research and to reflect on their own teaching in regards to it.
The bridge between teachers and researchers needs to be constructed from both sides
of the continents. But I believe that the responsibility to draw the design is largely on
researchers’ shoulders because they are the ones who are creating the building material, namely
research studies. As described at the very beginning of the present paper, this movement to
connect teachers and researchers as well as practice and research is nothing new, but as the field
of SLA is growing, there should be a greater chance than ever to carry out the construction.

266

Thirty-six years ago, Tarone et al. (1976) claimed, “[s]econd language acquisition research is
still in its infancy, and hasty pedagogical applications should not be made on the basis of its
findings” (p. 19), and listed seven limitations as follows:
(1) the restricted linguistic scope of studies to date, (2) lack of data on cognitive
processes and learning strategies, (3) limited information about the role of individual
variables in second language acquisition, (4) insufficient information about the role of
social and environmental variables in second language acquisition, (5) undeveloped
methodology for data collection and (6) for data analysis, and (7) the limited number
of replicated studies to date. (p. 19)

Referring to this list, Lightbown (1985) stated:
Nearly ten years later we could quite reasonably use the same list. This is not to say
that there has been no progress. But our understanding of acquisition is still far from
complete, and our ability to make recommendations is still very restricted. (p. 184)

Approximately the same time, Chaudron (1988) also deplored that “[t]heories and
claims about language teaching methods … have rarely been based on actual research in
language classrooms” (p. xv). Further twenty-seven years later from Lightbown (1985), if we
reexamine this list now, most SLA researchers would confidently acknowledge that the field of
SLA has grown up expansively in the past decades and successfully approached all of the seven
limitations above, probably except the last one. Notwithstanding, it is still unpredictable whether
there will be a seamless connection between the two continents where people can easily go
across the bridge or people still need to swim fiercely in order to reach the other side. The jury is

267

still out there. Still, I hope the present study shed a light on this issue and draw a line on the
blueprint for the bridge.

268

APPENDICES

269

Appendix A
The outline of the PowerPoint file given to the teacher
For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this dissertation.

1. The summary of three research studies
2. First of all…
If you have any questions about any part in this slide, please send me e-mail. I will reply

•

as soon as I can.
3. Error correction
•

When a student makes an error in a classroom, the teacher may give feedback on it.

•

This is called error correction, corrective feedback, or reactive Focus on Form. In
this slide, error correction will be used.
Error correction may be oral or written, but in this slide, we will focus on the oral one.

•

4. Two types of error correction
There are two types of error correction.

•

♦

Input-providing (reformulations)
²

♦

The teacher gives the correct answer.

Output-prompting (prompt)
²

The teacher tries to elicit the correct answer from the student.

I will introduce some examples of these two types of error correction.

•

5. Input-providing (Reformulations)
1.

Explicit correction

270

•

Provides learners with a correct form with a clear indication of what is being
corrected.
S: and three pear. (sounds like beer)
S2: three beer.
T: not beer. Pear. ← explicit correction

6. Input-providing (Reformulations)
2.

Recasts
•

Reformulates the whole or part of learner’s erroneous utterance without
changing its meaning.
S: Any person who is very great poet, I would be.
T: Oh, okay. All right. A great poet? You would be a great poet? ← recast

7. Output-prompting (Prompt)
3.

Clarification requests
•

Signals to learners that their utterances were either not understood or were ill
formed.
S: I want practice today, today.
T: I’m sorry? ← clarification request

8. Output-prompting (Prompt)
4.

Metalinguistic clues
•

Contains technical information regarding the student’s erroneous utterance
without explicitly providing the correct answer.
S: There are influence person who.
T: Influential is an adjective. ← metalinguistic feedback

271

S: Influential person - (unintelligible) - because of his power.
9. Output-prompting (Prompt)
5.

Elicitation
•

Involves at least three techniques for eliciting the correct form from the
students:
(a) Elicit completion (e.g., It is a …)
(b) Elicitative question (e.g., How do you say X in French?)
(c) Reformulation request (e.g., Can you say it another way?)

10. Output-prompting (Prompt)
5.

Elicitation
S: In a fast food restaurant, how much do you tip?
S: No money.
T: What’s the word? ← elicitation
S: five…four…

11. Output-prompting (Prompt)
6.

Repetition
•

Occurs when the teacher repeats learners’ ill-formed utterances without any
change.
S: Oh my God, it is too expensive, I pay only 10 dollars.
T: I pay? ← repetition
S: okay let’s go.

12. Output-prompting (Prompt)
7.

Multiple feedback

272

Refers to the combination of more than one type of feedback.

•

13. Three research studies
•

Now you are ready to read the three research studies on error correction.
1.

Lyster & Saito (2010)

2.

Mackey, Gass, & McDonough (2000)

3.

Truscott (1999)

14. Lyster & Saito (2010)
•

Lyster & Saito conducted a meta-analysis of 15 classroom-based studies of error
correction.

•

In a meta-analysis, the results of multiple studies are combined to see their overall trend.

15. Lyster & Saito (2010)
•

They categorized error correction into three types: prompts, recasts and explicit
correction, based on the figure on the next slide.

16. Lyster & Saito (2010): Figure 1 on p. 278
17. Lyster & Saito (2010)
•

Results:
1.

Error correction was more effective than no error correction, and the effectiveness
lasted for a few weeks.

2.

The effectiveness of recasts, prompts, and explicit correction was all significant.

3.

Prompts was more effective than recasts.

4.

The effectiveness of explicit correction was not significantly different from those
of recasts and prompts.

18. Mackey, Gass, & McDonough (2000)

273

•

Mackey et al. investigated how accurately learners can recognize the types of error
correction using a method called stimulated recall.

•

In stimulated recall, participants watch the video of their own interactions and make
comments about what they were thinking at some moment, but not what they are
thinking while they are watching the video.

19. Mackey, Gass, & McDonough (2000)
•

10 English as a second language (ESL) learners and 7 learners of Italian as a foreign
language (IFL) participated.

•

Each learner carried out a communicative task with a native (English) or near-native
(Italian) interviewer.

•

During the tasks, the interviewer provided error correction (mainly recasts) when the
learner made an error in morphosyntax, phonological, semantics, and lexis.

20. Mackey, Gass, & McDonough (2000)
•

After the tasks, each learner watched the video of their interaction.

•

The learner paused the video when they wanted to describe their thoughts at that
moment.

•

The researcher also paused the video when error correction was provided and asked the
learner to express her/his thoughts at that moment.

21. Mackey, Gass, & McDonough (2000): Table 2 on p. 483
22. Mackey, Gass, & McDonough (2000): Figure 1 on p. 487
23. Mackey, Gass, & McDonough (2000)
•

Results:
1.

ESL learners accurately recognized error correction generally for lexis and

274

phonology (83%, 60%), but not for morphosyntax (13%).
2.

IFL learners accurately recognized error correction generally for lexis (66%), but
not for phonology or morphosyntax (21%, 24%).

24. Truscott (1999)
•

Truscott basically claims that teachers should not correct students’ errors.

•

He provided the reasons from three perspectives: teachers, students, and others.

25. Truscott (1999)
•

From teachers’ perspective:
1.

It is difficult for teachers to understand the nature of errors correctly.

2.

It is difficult for teachers to present the correction clearly to students.

3.

It is difficult for teachers to make error correction consistent.

4.

Teachers may fail to notice errors because they are busy in the classroom.

26. Truscott (1999)
•

From students’ perspective:
1.

Students may not notice error correction.

2.

Students may not understand error correction.

3.

Students may misunderstand error correction. For example, they may
misunderstand that error correction is teachers’ understanding or approval, and
vice versa.

4.

Students may not take error correction seriously.

5.

Students may not want to accept error correction.

27. Truscott (1999)
6.

An error correction for a student may not be appropriate for other students.

275

7.

Error correction is humiliating for some students, so they may stop expressing
themselves.

8.

Adjusting error correction to each student, students may feel confused or inferior
because only some students receive error correction whereas others do not.

9.

Students may not be ready to learn from error correction because there is a fixed
order in grammar acquisition.

10. Students need to rehearse and make use of error correction in order to incorporate
it.
28. Truscott (1999)
From other perspectives:

•

1.

Error correction interrupts communication.

2.

Oral error correction is fleeting, and there is no written record to recall later.

3.

Many previous studies show that error correction is ineffective.

4.

Error correction facilitates explicit knowledge, but does not facilitate implicit
knowledge.

29. For reading
•

Now you are familiar with each of these three studies.

•

If you have time, I would like you to actually read these studies (but even if you can’t,
please don’t worry).

30. If you are going to read
1.

You do not need to read all of the three studies. You can start reading from the one in
which you are most interested. There is no order in reading.

2.

Except Truscott (1999), you can read only the results section. If you like, you can also

276

read the discussion section. You can skip the remaining sections.
3.

For Truscott (1999), it is probably better to read from the beginning.

4.

You can underline, highlight, and/ or take notes in the text or another sheet of paper, if
you like.

31. If you finish reading
•

If you finish reading all the three studies, please contact me by e-mail, and I will make
an appointment with you for an interview.

•

If I don’t hear from you for two weeks, I will contact you by e-mail.

277

Appendix B
The questions for the Interview 1

L2 learning experience
1. Tell me about your language learning experience.
2. What do you remember about the experience of studying that language?
3. How has this experience influenced how you teach?
(a) What L2s have you studied so far?
(b) How did you decide to study that language?
(c) Where did you learn that language?
(d) How long did you learn that language?
(e) What methods were used in the classroom?
(f)

How did you like learning that language?

(g) How much did you use that language outside the classroom?
(h) Tell me about your experiences of traveling abroad related to language use and
learning.
(i)

What do you remember about the experience of traveling abroad?
①

Where did you go?

②

When did you go there?

③

How long did you stay there?

④

What did you do over there?

⑤

How much L2 did you use over there?

278

L2 teacher training experience
1. Tell me about any teacher training that you have had thus far.
2. What do you remember about the experience of pre-service and in-service training to teach
L2?
3. How has this experience influenced how you teach L2?
(a) How have you decided to learn how to teach ESL?
(b) Where did you learn to teach L2?
(c) How did you learn to teach L2?
(d) How long did you learn to teach L2?
(e) What qualifications and certifications do you have?
(f)

Did you take a practicum?
①

What do you remember about the experience of the practicum?

②

What did you teach?

③

Where did you do it?

④

How did you do it?

⑤

How long did you do it?

⑥

How did you like it?

L2 teaching experience
1. Tell me about your language teaching experience.
2. How would you describe your teaching style?
3. What do you think has been the greatest influence on your teaching?
4. Describe a recent lesson that you think went well.

279

5. How has this experience influenced how you teach English?
6. How has your teaching changed from the past?
7. What are your principles and philosophy of teaching?
(a) What do you remember about the experience of teaching English?
(b) How did you decide to teach English?
(c) How long have you taught English?
(d) Where have you taught English?
(e) Whom have you taught English?
(f)

What skills have you taught in English?

(g) What levels of students have you taught English?
(h) How do you like teaching English?

CF
1. Tell me about how you deal with students’ errors in class.
2. What factors affect your error correction practices?
3. What do you think has been the greatest influence on your CF?
4. How has this experience influenced how you teach English?
(a) What kinds of CF do you use?
(b) How do you choose CF among all choices?
(c) Is there any particular way to provide CF that you like?
(d) What is your policy about CF?
(e) How much do you think CF is effective?

280

Research studies
1. From what do you learn new ideas, techniques, and theories of teaching?
2. Do you read research studies? If so, how often do you read them? What kind of research
studies do you read?
3. How does reading research studies influence your teaching?

Current teaching context
1. Tell me about the current class that you are teaching.
2. What king of activities do you do in class?
3. What is your goal for this class?
4. What are the external factors that affect your teaching?
5. Could you give me a copy of the syllabus of your class?
(a) What are some of the strengths and weaknesses of this class?
(b) What do you like/ dislike about this class?
(c) What do you think of your students?
(d) What do you think of the proficiency level of your students?
(e) How much of your speech do you think they understand?
(f)

How does the ELC affect your teaching?

(g) How does the curriculum affect your teaching?
(h) What other factors do you think are affecting your teaching?

281

Appendix C
The questions for the Interview 2

1. Did you go through the PowerPoint file?
2. Did you have time to read any of the three articles?
3. Which one(s) did you read? Which part(s) did you read? When did you read them?
4. What made you decide to read it/ them?
5. What did you think of these articles?
(a) Do you have any questions?
(b) How did you enjoy them?
(c) How did you like them?
6. How do you think will these articles affect your teaching?
(a) Is there any article that you found particularly interesting?
(b) Is there any article that you found particularly useful?
(c) Is there any article that you found particularly uninteresting?
(d) Is there any article that you found particularly useless?

282

Appendix D
The transcript convention

T

Teacher

S

Student (when appropriate, a number is assigned to each student to identify each
one of them, such as S1 and S2)

Ss

Students

R

Researcher

NS

Native speaker of L2

NNS

Non-native speaker of L2

(word)

Uncertain transcript

(?)

Unable to hear

(( ))

Explanation of nonverbal action

[]

Transcript hidden purposefully (e.g., anonymity)

//

Pronunciation

/

Overlapping speech

=

Connected speech

.

Falling intonation

,

Continuous (low-rising) intonation

?

Rising intonation

…

Utterance continues (i.e., omitted; may include utterances of the researcher in case
of Interviews and Stimulated Recalls)

CAPITAL

Emphasis

283

Blank line

A pause lasting over 10 seconds

Underline

An error

Italics

CF

284

REFERENCES

285

REFERENCES
Akbari, R., & Dadvand, B. (2011). Does formal teacher education make a difference? A
comparison of pedagogical thought units of B.A. versus M.A. teachers. The Modern
Language Journal, 95(1), 44-60. doi: 10.1111/j.1540-4781.2010.01142.x
Alcón, E. (2007). Incidental focus on form, noticing and vocabulary learning in the EFL
classroom. International Journal of English Studies, 7(2), 41-60.
Allison, D., & Carey, J. (2007). What do university language teachers say about language
teaching research? TESL Canada Journal, 24(2), 61-81.
Allwright, D. (2005). From teaching points to learning opportunities and beyond. TESOL
Quarterly, 39(1), 9-31.
Almarza, G. G. (1996). Student foreign language teacher's knowledge growth. In D. Freeman &
J. C. Richards (Eds.), Teacher learning in language teaching (pp. 50-78). Cambridge, UK:
Cambridge University Press.
Angelova, M. (2005). Using Bulgarian mini-lessons in an SLA course to improve the KAL of
American ESL teachers. In N. Bartels (Ed.), Researching applied linguistics and language
teacher education (pp. 27-42). New York, NY: Springer Science+Business Media.
Bailey, K. M. (1992). The processes of innovation in language teacher development: What, why
and how teachers change. In J. Flowerdew, M. Brock & S. Hsia (Eds.), Perspectives on
second language teacher education (pp. 253-282). Hong Kong: City Polytechnic.
Bailey, K. M. (1996). The best laid plans: Teachers' in-class decisions to depart from their lesson
plans. In K. M. Bailey & D. Nunan (Eds.), Voices from the language classroom:
Qualitative research in second language education (pp. 15-40). Cambridge, UK:
Cambridge University Press.
Bailey, K. M., Bergthold, B., Braunstein, B., Fleischman, N. J., Holbrook, M. P., Tuman, J., . . .
Zambo, L. J. (1996). The language learner's autobiography: Examining the "apprenticeship
of observation". In D. Freeman & J. C. Richards (Eds.), Teacher learning in language
teaching (pp. 11-29). Cambridge, UK: Cambridge University Press.
Bartels, N. (2003). How teachers and researchers read academic articles. Teaching and Teacher
Education, 19(7), 737-753. doi: 10.1016/j.tate.2003.06.001
Basturkmen, H., Loewen, S., & Ellis, R. (2004). Teachers' stated beliefs about incidental focus
on form and their classroom practices. Applied Linguistics, 25(2), 243-272.
286

Belcher, D. (2007). A bridge too far? TESOL Quarterly, 41(2), 396-399.
Berliner, D. C. (1987). Knowledge is power: A talk to teachers about a revolution in the teaching
profession. In D. C. Berliner & B. V. Rosenshine (Eds.), Talks to teachers (pp. 3-33). New
York, NY: Random House.
Bolitho, R. (1991). A place for second language acquisition in teacher development and in
teacher education programmes. In E. Sadtono (Ed.), Language acquisition and the second/
foreign language classroom (pp. 25-37). Singapore: SEAMEO Regional Language Centre.
Borg, M. (2001). Teachers' beliefs. ELT Journal, 55(2), 186-188.
Borg, M. (2005). A cases study of the development in pedagogic thinking of a pre-service
teacher. TESL-EJ, 9(2), 1-30.
Borg, S. (1998a). Data-based teacher development. ELT Journal, 52(4), 273-281.
Borg, S. (1998b). Teachers' pedagogical systems and grammar teaching: A qualitative study.
TESOL Quarterly, 32(1), 9-38.
Borg, S. (1999a). The use of grammatical terminology in the second language classroom: A
qualitative study of teachers' practices and cognitions. Applied Linguistics, 20(1), 95-126.
Borg, S. (1999b). Teachers' theories in grammar teaching. ELT Journal, 53(3), 157-167.
Borg, S. (1999c). Studying teacher cognition in second language grammar teaching. System, 27,
19-31.
Borg, S. (2003a). Teacher cognition in language teaching: A review of research on what
language teachers think, know, believe, and do. Language Teaching, 36(2), 81-109. doi:
10.1017/s0261444803001903
Borg, S. (2003b). Teacher cognition in grammar teaching: A literature review. Language
Awareness, 12(2), 96-108.
Borg, S. (2007a). Research engagement in English language teaching. Teaching and Teacher
Education, 23(5), 731-747. doi: 10.1016/j.tate.2006.03.012
Borg, S. (2007b). English language teachers' views of research: Some insights from Switzerland.
ETAS Newsletter, 24, 15-18.

287

Borg, S. (2009a). English Language Teachers' Conceptions of Research. Applied Linguistics,
30(3), 358-388. doi: 10.1093/applin/amp007
Borg, S. (2009b). Language teacher cognition. In A. Burns & J. C. Richards (Eds.), The
Cambridge guide to second language teacher education (pp. 163-171). Cambridge, UK:
Cambridge University Press.
Borg, S. (2011). The impact of in-service teacher education on language teachers’ beliefs. System.
doi: 10.1016/j.system.2011.07.009
Borko, H., & Niles, J. (1982). Factors contributing to teachers' judgments about students and
decisions about grouping students for reading instruction. Journal of Reading Behavior,
14(2), 127-140.
Bramald, R., Hardman, F., & Leat, D. (1995). Initial teacher trainees and their views of teaching
and learning. Teaching and Teacher Education, 11(1), 23-31.
Breen, M. P., Hird, B., Milton, M., Oliver, R., & Thwaite, A. (2001). Making sense of language
teaching: Teachers' principles and classroom practices. Applied Linguistics, 22(4),
470-501.
Brinton, D., & Holten, C. (1989). What novice teachers focus on: The practicum in TESL.
TESOL Quarterly, 23(2), 343-350.
Brousseau, B. A., Book, C., & Byers, J. L. (1988). Teacher beliefs and the cultures of teaching.
Journal of Teacher Education, 39(6), 33-39. doi: 10.1177/002248718803900607
Brown, C. A., & Cooney, T. J. (1982). Research on teacher education: A philosophical
orientation. Journal of Research and Development in Education, 15(4), 13-18.
Brumfit, C., Mitchell, R., & Hooper, J. (1996). 'Grammar', 'language' and classroom practice. In
M. Hughes (Ed.), Teaching and learning in changing times (pp. 70-87). Oxford, UK:
Blackwell.
Buchmann, M. (1987). Teaching knowledge: The lights that teachers live by. Oxford Review of
Education, 13(2), 151-164.
Buchmann, M. (1991). Making new or making do: An inconclusive argument about teaching.
American Journal of Education, 99(3), 279-297.
Buchmann, M., & Schwille, J. (1983). Education: The overcoming of experience. American
288

Journal of Education, 92(1), 30-51.
Bullough, R. V., & Baughman, K. (1993). Continuity and Change in Teacher Development: First
Year Teacher after Five Years. Journal of Teacher Education, 44(2), 86-95. doi:
10.1177/0022487193044002003
Burns, A. (1996). Starting all over again: From teaching adults to teaching beginners. In D.
Freeman & J. C. Richards (Eds.), Teacher learning in language teaching (pp. 154-177).
Cambridge, UK: Cambridge University Press.
Busch, D. (2010). Pre-service teacher beliefs about language learning: The second language
acquisition course as an agent for change. Language Teaching Research, 14(3), 318-337.
doi: 10.1177/1362168810365239
Cabaroglu, N., & Robers, J. (2000). Development in student teachers' pre-existing beliefs during
a 1-year PGCE programme. System, 28, 387-402.
Calderhead, J., & Robson, M. (1991). Images of teaching: Student teachers' early conceptions of
classroom practice. Teaching and Teacher Education, 7(1), 1-8.
Carter, K., & Doyle, W. (1989). Classroom research as a resource for the graduate preparation of
teachers. In A. Woolfolk (Ed.), The graduate preparation of teachers (pp. 51-68).
Englewood Cliffs, NJ: Prentice-Hall.
Chandler, J. (2003). The efficacy of various kinds of error feedback for improvement in the
accuracy and fluency of L2 student writing. Journal of Second Language Writing, 12(3),
267-296. doi: 10.1016/s1060-3743(03)00038-9
Chaudron, C. (1988). Second language classrooms. Cambridge, UK: Cambridge University
Press.
Cheung, M., & Pennington, M. C. (1994). Cultivating product-oriented teachers to adopt process
writing: An example of an innovation-related in-service programme. In D. C. S. Li, D.
Mahoney & J. C. Richards (Eds.), Exploring second language teacher development (pp.
209-231). Hong Kong: City Polytechnic.
Clark, C. M. (1988). Asking the right questions about teacher preparation: Contributions of
research on teacher thinking. Educational Researcher, 17(2), 5-12. doi:
10.3102/0013189x017002005
Clarke, M. A. (1994). The dysfunctions of the theory/ practice discourse. TESOL Quarterly,
28(1), 9-26.
289

Collie Graden, E. (1996). How language teachers' beliefs about reading instruction are mediated
by their beliefs about students. Foreign Language Annals, 29(3), 387-395.
Crandall, J. J. (2000). Language teacher education. Annual Review of Applied Linguistics, 20,
34-55.
Crookes, G., & Arakaki, L. (1999). Teaching idea sources and work conditions in an ESL
program. TESOL Journal, 8(1), 15-19.
Doughty, C. J. (2003). Instructed SLA: Constraints, compensation and enhancement. In C. J.
Doughty & M. H. Long (Eds.), The handbook of second language acquisition (pp.
256-310). Malden, MA: Blackwell.
Duff, P. A. (2006). Beyond generalizability: Contextualization, complexity, and credibility in
applied linguistics research. In M. Chalhoub-Deville, C. A. Chapelle & P. A. Duff (Eds.),
Inferences and generalizability in applied linguistics (pp. 65-95). Amsterdam, The
Netherlands: John Benjamins.
Duff, P. A., & Uchida, Y. (1997). The negotiation of teachers' sociocultural identities and
practices in postsecondary EFL classrooms. TESOL Quarterly, 31(3), 451-486.
Dunkin, M. J. (1996). Types of errors in synthesizing research in education. Review of
Educational Research, 66(2), 87-97. doi: 10.3102/00346543066002087
Egi, T. (2004). Verbal reports, noticing, and SLA research. Language Awareness, 13(4),
243-264.
Eisenstein Ebsworth, M., & Schweers, C. W. (1997). What researchers say and practitioners do:
Perspectives on conscious grammar instruction in the ESL classroom. Applied Language
Learning, 8(2), 237-260.
Eisner, E. W. (1984). Can educational research inform educational practice? The Phi Delta
Kappan, 65(7), 447-452.
Ellis, R. (1994). Second language acquisition research and teacher development: The case of
teachers' questions. In D. C. S. Li, D. Mahoney & J. C. Richards (Eds.), Exploring second
language teacher development (pp. 175-193). Hong Kong: City Polytechnic.
Ellis, R. (1997). SLA and language pedagogy: An educational perspective. Studies in Second
Language Acquisition, 20, 69-92.

290

Ellis, R. (1998). Teaching and research: Options in grammar teaching. TESOL Quarterly, 32(1),
39-60.
Ellis, R. (2005). Instructed second language acquisition: A literature review. Wellington, New
Zealand: Research Division.
Ellis, R. (2009). Second language acquisition: Theory, research and language pedagogy. Paper
presented at the AAAL 2009, Denver, CO.
Ellis, R., & Barkhuizen, G. (2005). Analyzing learner language. Oxford, UK: Oxford University
Press.
Ellis, R., Basturkmen, H., & Loewen, S. (2001a). Preemptive focus on form in the ESL
classroom. TESOL Quarterly, 35(3), 407-432.
Ellis, R., Basturkmen, H., & Loewen, S. (2001b). Learner uptake in communicative ESL lessons.
Language Learning, 51(2), 281-318.
Ellis, R., Loewen, S., & Erlam, R. (2006). Implicit and explicit corrective feedback and the
acquisition of L2 grammar. Studies in Second Language Acquisition, 28, 339-368.
Ellis, R., Loewen, S., & Erlam, R. (2009). Implicit and explicit corrective feedback and the
acquisition of L2 grammar. In R. Ellis, S. Loewen, C. Elder, R. Erlam, J. Philp & H.
Reinders (Eds.), Implicit and explicit knowledge in second language learning, testing and
teaching (pp. 303-332). Bristol, UK: Multilingual Matters.
Eraut, M. (1994). Developing professional knowledge and competence. London: Falmer.
Erlam, R. (2008). What do you researchers know about language teaching?: Bridging the gap
between SLA research and language pedagogy. Innovation in Language Learning and
Teaching, 2(3), 253-267. doi: 10.1080/17501220802158958
Erlam, R., Sakui, K., & Ellis, R. (2006). Instructed second language acquisition: Case studies.
Wellington, NZ: Learning Media.
Eykyn, L. B. (1987). Confessions of a high school language teacher or why I never (used to) read
foreign language annals. Foreign Language Annals, 20(3), 265-266.
Fang, Z. (1996). A review of research on teacher beliefs and practices. Educational Research,
38(1), 47-65.

291

Farrell, T. S. C. (1999). The reflective assignment: Unlocking pre-service English teachers'
beliefs on grammar teaching. RELC Journal, 30(2), 1-17. doi:
10.1177/003368829903000201
Feiman-Nemser, S., & Melnick, S. (1992). Introducing teaching. In S. Feiman-Nemser & H.
Featherstone (Eds.), Exploring teaching: Reinventing an introductory course (pp. 1-17).
New York, NY: Teachers College Press.
Fogle, L., Jacobsen, N., Friginal, E., Grieve, J., & Uccelli, P. (2010). AAAL at TESOL. Paper
presented at the TESOL 2010, Boston, MA.
Foster, P., Tonkyn, A., & Wigglesworth, G. (2000). Measuring spoken language: A unit for all
reasons. Applied Linguistics, 21(3), 354-375.
Freeman, D. (1989). Teacher training, development, and decision making: A model of teaching
and related strategies for language teacher education. TESOL Quarterly, 23(1), 27-45.
Freeman, D. (1992). Language teacher education emerging discourse, and change in classroom
practice. In J. Flowerdew, M. Brock & S. Hsia (Eds.), Perspectives on second language
teacher education (pp. 1-21). Hong Kong: City Polytechnic.
Freeman, D. (1993). Renaming experience/ reconstructing practice: Developing new
understandings of teaching. Teaching and Teacher Education, 9(5, 6), 485-497.
Freeman, D. (1996a). Redefining the relationship between research and what teachers know. In
K. M. Bailey & D. Nunan (Eds.), Voices from the language classroom: Qualitative
research in second language education (pp. 88-115). New York, NY: Cambridge
University Press.
Freeman, D. (1996b). Renaming experience/ reconstructing practice: Developing new
understandings of teaching. In D. Freeman & J. C. Richards (Eds.), Teacher learning in
language teaching (pp. 221-241). Cambridge, UK: Cambridge University Press.
Freeman, D., & Johnson, K. E. (1998). Reconceptualizing the knowledge-base of language
teacher education. TESOL Quarterly, 32(3), 397-417.
Gass, S. (2003). Input and interaction. In C. J. Doughty & M. H. Long (Eds.), The handbook of
second language acquisition (pp. 224-255). Oxford, UK: Blackwell.
Gass, S., & Mackey, A. (2000). Stimulated recall methodology in second language research.
Mahwah, NJ: Lawrence Erlbaum.

292

Golombek, P. R. (1998). A study of language teachers' personal practical knowledge. TESOL
Quarterly, 32(3), 447-464.
Goodman, J. (1988). Constructing a practical philosophy of teaching: A study of preservice
teachers' professional perspectives. Teaching and Teacher Education, 4(2), 121-137.
Guskey, T. R. (1986). Staff development and the process of teacher change. Educational
Researcher, 15(5), 5-12. doi: 10.3102/0013189x015005005
Hall, G. E., & Loucks, S. F. (1982). Bridging the gap: Policy research rooted in practice. In A.
Liberman & M. W. McLaughlin (Eds.), Policy making in education (pp. 133-158).
Chicago: University of Chicago Press.
Han, Z. (2007). Pedagogical implications: Genuine or pretentious? TESOL Quarterly, 41(2),
387-393.
Harris, K., Hellermann, J., Keck, C., Olsher, D., & Santos, M. (2011). Bridging the
research-practice divide through collaborative research on video-based classroom
discourse. Paper presented at the AAAL 2011, Chicago, IL.
Hatch, E. (1978). Apply with caution. Studies in Second Language Acquisition, 2, 123-143.
Hollingsworth, S. (1989). Prior beliefs and cognitive change in learning to teach. American
Educational Research Journal, 26(2), 160-189. doi: 10.3102/00028312026002160
Holt-Reynolds, D. (1992). Personal history-based beliefs as relevant prior knowledge in course
work. American Educational Research Journal, 29(2), 325-349. doi:
10.3102/00028312029002325
Holten, C. A., & Brinton, D. M. (1995). "You shoulda been there": Charting novice teacher
growth using dialogue journals. TESOL Journal, 4, 23-26.
Horwitz, E. K. (1988). The beliefs about language learning of beginning university foreign
language students. The Modern Language Journal, 72(3), 283-294.
Howatt, A. P. R. (1984). A history of English language teaching. Oxford, UK: Oxford University
Press.
Hsiao-Ching, S. (2000). The interplay of a biology teacher's beliefs, teaching practices and
gender-based student-teacher classroom interaction. Educational Research, 42(1), 100-111.
doi: 10.1080/001318800363953
293

Hymes, D. (1971). Competence and performance in linguistic theory. In R. Huxley & E. Ingram
(Eds.), Language acquisition: Models and methods (pp. 3-28). London: Academic Press.
Johnson, K. E. (1992a). The relationship between teachers' beliefs and practices during literacy
instruction for non-native speakers of English. Journal of Reading Behavior, 24(1),
83-108.
Johnson, K. E. (1992b). Learning to teach: Instructional actions and decisions of preservice ESL
teachers. TESOL Quarterly, 26(3), 507-535.
Johnson, K. E. (1994). The emerging beliefs and instructional practices of preservice English as
a second language teachers. Teaching and Teacher Education, 10(4), 439-452.
Johnson, K. E. (1996a). The role of theory in L2 teacher education. TESOL Quarterly, 30(4),
765-771.
Johnson, K. E. (1996b). The vision versus the reality: The tensions of the TESOL practicum. In
D. Freeman & J. C. Richards (Eds.), Teacher learning in language teaching (pp. 30-49).
Cambridge, UK: Cambridge University Press.
Johnson, K. E., & Golombek, P. R. (2002). Inquiry into practice: Teachers’ personal and
professional growth. In K. E. Johnson & P. R. Golombek (Eds.), Teachers’ narrative
inquiry as professional development (pp. 1-14). Cambridge, UK: Cambridge University
Press.
Kagan, D. M. (1990). Ways of evaluating teacher cognition: Inferences concerning the
goldilocks principle. Review of Educational Research, 60(3), 419-469. doi:
10.3102/00346543060003419
Kagan, D. M. (1992a). Implications of research on teacher belief. Educational Psychologist,
27(1), 65-90.
Kagan, D. M. (1992b). Professional growth among preservice and beginning teachers. Review of
Educational Research, 62(2), 129-169. doi: 10.3102/00346543062002129
Karavas-Doukas, E. (1996). Using attitude scales to investigate teachers' attitudes to the
communicative approach. ELT Journal, 50(3), 187-198.
Kennedy, C. (1999). Introduction: Learning to change. In C. Kennedy, P. Doyle & C. Goh (Eds.),
Exploring change in English language teaching (pp. iv-viii). Oxford, UK: Macmillan
Heinemann.

294

Kennedy, D. (1999). The foreign trainer as change agent and implications for teacher education
programmes in China. In C. Kennedy, P. Doyle & C. Goh (Eds.), Exploring change in
English language teaching. Oxford, UK: Macmillan Heinemann.
Kennedy, M. M. (1990). Policy issues in teacher education National Center for Research on
Teacher Learning. East Lansing: Michigan State University.
Kennedy, M. M. (1991). An agenda for research on teacher learning National Center for
Research on Teacher Learning. East Lansing: Michigan State University.
Kennedy, M. M. (1997). The connection between research and practice. Educational Researcher,
26(7), 4-12. doi: 10.3102/0013189x026007004
Kettle, B., & Sellars, N. (1996). The development of student teachers' practical theory of
teaching. Teaching and Teacher Education, 12(1), 1-24.
Kramsch, C. (1995). The applied linguist and the foreign language teacher: Can they talk to each
other? In G. Cook & B. Seidlhofer (Eds.), Principle & practice in applied linguistics (pp.
43-56). Oxford, UK: Oxford University Press.
Krashen, S. (1981). Second language acquisition and second language learning. Oxford, UK:
Pergamon.
Krashen, S. (1983). Second language acquisition theory and the preparation of teachers: Toward
a rationale. In J. E. Alatis, H. H. Stern & P. Strevens (Eds.), Georgetown University Round
Table on Languages and Linguistics 1983 (pp. 255-263). Washington, D.C.: Georgetown
University Press.
Krashen, S. (1989). Language acquisition and language education. Englewood Cliff, NJ:
Prentice Hall.
Kubanyiova, M. (2006). Developing a motivational teaching practice in EFL teachers in
Slovakia: Challenges of promoting teacher change in EFL contexts. TESL-EJ, 10(2), 1-17.
Kumaravadivelu, B. (1993). Maximizing learning potential in the communicative classroom.
ELT Journal, 47(1), 12-21.
Kumaravadivelu, B. (2001). Toward a postmethod pedagogy. TESOL Quarterly, 35, 537-560.
Lamb, M. (1995). The consequences of INSET. ELT Journal, 49(1), 72-80.

295

Lamie, J. M. (2004). Presenting a model of change. Language Teaching Research, 8(2),
115-142.
Larsen-Freeman, D. (2000). Techniques and principles in language teaching. Oxford, UK:
Oxford University Press.
Lasley, T. J. (1980). Preservice reacher beliefs about teaching. Journal of Teacher Education,
31(4), 38-41. doi: 10.1177/002248718003100410
Leeman, J. (2007). Feedback in L2 learning: Responding to errors during practice. In R.
DeKeyser (Ed.), Practice in a second language: Perspectives from applied linguistics and
cognitive psychology (pp. 111-137). Cambridge, UK: Cambridge University Press.
Legutke, M. K., & Ditfurth, M. S.-v. (2009). School-based experience. In A. Burns & J. C.
Richards (Eds.), The Cambridge guide to second language teacher education (pp.
209-217). Cambridge, UK: Cambridge University Press.
Leung, C., Hawkins, M., & Canagarajah, S. (2010). TESOL at AAAL. Paper presented at the
AAAL 2010, Atlanta, GA.
Li, S. (2010). The effectiveness of corrective feedback in SLA: A meta-analysis. Language
Learning, 60(2), 309-365. doi: 10.1111/j.1467-9922.2010.00561.x
Lightbown, P. M. (1985). Great expectations: Second-language acquisition research and
classroom teaching. Applied Linguistics, 6(2), 173-189.
Lightbown, P. M. (2000). Anniversary article: Classroom SLA research and second language
teaching. Applied Linguistics, 21(4), 431-462.
Lightbown, P. M., & Spada, N. (2006). How languages are learned (3rd ed.). Oxford, UK:
Oxford University Press.
Lo, Y.-H. G. (2005). Relevance of knowledge of second language acquisition. In N. Bartels (Ed.),
Researching applied linguistics in language teacher education (pp. 135-158). New York,
NY: Springer Science+Business Media.
Loewen, S. (2003). Variation in the frequency and characteristics of incidental focus on form.
Language Teaching Research, 7(3), 315-345.
Loewen, S. (2011a). Focus on form. In E. Hinkel (Ed.), Handbook of research in second
language teaching and learning (Vol. 2, pp. 576-592). New York, NY: Routledge.
296

Loewen, S. (2011b). The role of feedback. In S. Gass & A. Mackey (Eds.), The Routledge
Handbook of second language acquisition (pp. 24-40). New York, NY: Routledge.
Loewen, S., & Philp, J. (2006). Recasts in the adult English L2 classroom: Characteristics,
explicitness, and effectiveness. The Modern Language Journal, 90(4), 536-556.
Long, M. H. (1990). The least a second language acquisition theory needs to explain. TESOL
Quarterly, 24(4), 649-666.
Long, M. H. (1991). Focus on form: A design feature in language teaching methodology. In K.
de Bot, R. B. Ginsberg & C. Kramsch (Eds.), Foreign language research in cross-cultural
perspective (pp. 39-52). Amsterdam: John Benjamins.
Long, M. H. (1996). The role of the linguistic environment in second language acquisition. In W.
C. Ritchie & T. K. Bhatia (Eds.), Handbook of second language acquisition (pp. 413-468).
San Diego, CA: Academic Press.
Long, M. H. (2007). Problems in SLA. Mahwah, NJ: Lawrence Erlbaum.
Long, M. H., & Robinson, P. (1998). Focus on form: Theory, research, and practice. In C.
Doughty & M. H. Long (Eds.), Focus on form in classroom second language acquisition
(pp. 15-41). Cambridge, UK: Cambridge University Press.
Lortie, D. C. (1975). School teacher. Chicago: University of Chicago Press.
Lyster, R. (1998a). Negotiation of form, recasts, and explicit correction in relation to error types
and learner repair in immersion classrooms. Language Learning, 48(2), 183-218.
Lyster, R. (1998b). Recasts, repetition, and ambiguity in L2 classroom discourse. Studies in
Second Language Acquisition, 20, 51-81.
Lyster, R., & Ranta, L. (1997). Corrective feedback and learner uptake. Studies in Second
Language Acquisition, 20, 37-66.
Lyster, R., & Saito, K. (2010). Oral feedback in classroom SLA: A meta-analysis. Studies in
Second Language Acquisition, 32(2), 265-302. doi: 10.1017/s0272263109990520
MacDonald, M., Badger, R., & White, G. (2001). Changing values: What use are theories of
language learning and teaching? Teaching and Teacher Education, 17, 949-963.
Mackey, A., & Gass, S. (2005). Second language research: Methodology and design. Mahwah,
297

NJ: Lawrence Erlbaum.
Mackey, A., Gass, S., & McDonough, K. (2000). How do learners perceive interactional
feedback? Studies in Second Language Acquisition, 22, 471-497.
Mackey, A., & Goo, J. (2007). Interaction research in SLA: A meta-analysis and research
synthesis. In A. Mackey (Ed.), Conversational interaction in second language acquisition
(pp. 408-452). Oxford, UK: Oxford University Press.
Mackey, A., & Philp, J. (1998). Conversational interaction and second language development:
Recasts, responses, and red herrings? The Modern Language Journal, 82(3), 338-356.
Mackey, A., Polio, C., & McDonough, K. (2004). The relationship between experience,
education and teachers' use of incidental focus-on-form techniques. Language Teaching
Research, 8(3), 301-327.
Magnan, S. S. (2007). Gauging the scholarly value of connecting research to teaching. TESOL
Quarterly, 41(2), 400-404.
Markee, N. (1997). Second language acquisition research: A resource for changing teachers'
professional cultures? The Modern Language Journal, 81(1), 80-93.
Mattheoudakis, M. (2007). Tracking changes in pre-service EFL teacher beliefs in Greece: A
longitudinal study. Teaching and Teacher Education, 23(8), 1272-1288. doi:
10.1016/j.tate.2006.06.001
McDonough, J., & McDonough, S. (1990). What's the use of research? ELT Journal, 44(2),
102-109.
McNamara, O. (2002). Evidence-based practice through practice-based evidence. In O.
McNamara (Ed.), Becoming an evidence-based practitioner (pp. 15-26). London:
RoutledgeFalmer.
Mitchell, R., Brumfit, C., & Hooper, J. (1994a). 'Knowledge about language': Policy, rationales
and practices. Research Papers in Education, 9(2), 183-205.
Mitchell, R., Brumfit, C., & Hooper, J. (1994b). Perceptions of language and language learning
in English and foreign language classrooms. In M. Hughes (Ed.), Perceptions of teaching
and learning (pp. 53-65). Clevedon, UK: Multilingual Matters.
Mitchell, R., & Myles, F. (1998). Second language learning theories. London: Arnold.
298

Mok, W. E. (1994). Reflecting on reflections: A case study of experienced and inexperienced
ESL teachers. System, 22(1), 93-111.
Mori, J., & Zuengler, J. (2008). Conversation analysis and talk-in-interaction in the classrooms.
In M. Martin-Jones, A. M. de Mejia & N. H. Hornberger (Eds.), Encyclopedia of language
and education (2nd ed., Vol. 3: Discourse and education, pp. 15-26). New York: Springer.
Murray, H. (2003). Tracing the development of language awareness: An exploratory study of
language teachers in training. Bern, Switzerland: University of Bern.
Nespor, J. (1987). The role of beliefs in the practice of teaching. Journal of Curriculum Studies,
19(4), 317-328.
Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social
judgment. Englewood Cliff, NJ: Prentice Hall.
Norris, J., & Ortega, L. (2000). Effectiveness of L2 instruction: A research synthesis and
quantitative meta-analysis. Language Learning, 50(3), 417-528.
Numrich, C. (1996). On becoming a language teacher: Insights from diary studies. TESOL
Quarterly, 30(1), 131-153.
Nunan, D. (1987). Communicative language teaching: Making it work. ELT Journal, 41(2),
136-145.
Nunan, D. (1992). The teacher as decision-maker. In J. Flowerdew, M. Brock & S. Hsia (Eds.),
Perspectives on second language teacher education (pp. 135-165). Hong Kong: City
Polytechnic.
Oakeshott, M. (1962). Rationalism in politics: And other essays. London: Methuen.
Okazaki, H. (1996). Effects of a method course for pre-service teachers on beliefs on language
learning. Journal of Japanese Language Teaching, 89, 25-38.
Ortega, L. (2005). For what and for whom is our research? The ethical as transformative lens in
instructed SLA. The Modern Language Journal, 89(3), 427-443.
Ortega, L. (2012). Epistemological diversity and moral ends of research in instructed SLA.
Language Teaching Research, 16(2), 206-226. doi: 10.1177/0267658311431373
Pajares, M. F. (1992). Teachers' beliefs and educational research: Cleaning up a messy construct.
299

Review of Educational Research, 62(3), 307-332. doi: 10.3102/00346543062003307
Peacock, M. (1999). Beliefs about language learning and their relationship to proficiency.
International Journal of Applied Linguistics, 9(2), 247-265.
Peacock, M. (2001). Pre-service ESL teachers' beliefs about second language learning: A
longitudinal study. System, 29, 177-195.
Pennington, M. C. (1996). When input becomes intake: Tracing the sources of teachers’ attitude
change. In D. Freeman & J. C. Richards (Eds.), Teacher learning in language teaching (pp.
320-348). Cambridge, UK: Cambridge University Press.
Pennington, M. C., & Richards, J. C. (1997). Reorienting the teaching universe: the experience
of five first-year English teachers in Hong Kong. Language Teaching Research, 1(2),
149-178. doi: 10.1177/136216889700100204
Pennington, M. C., & Urmston, A. W. (1998). The teaching orientation of graduating students on
a BATESL course in Hong Kong: A comparison with first-year students. Hong Kong
Journal of Applied Linguistics, 3, 17-46.
Phipps, S., & Borg, S. (2007). Exploring the relationship between teachers' beliefs and their
classroom practice. The Teacher Trainer, 21(3), 17-19.
Phipps, S., & Borg, S. (2009). Exploring tensions between teachers’ grammar teaching beliefs
and practices. System, 37(3), 380-390. doi: 10.1016/j.system.2009.03.002
Pica, T. (1994). Questions from the language classroom: Research perspectives. TESOL
Quarterly, 28(1), 49-79.
Pica, T., & Long, M. H. (1986). The linguistic and conversational performance of experienced
and inexperienced teachers. In R. Day (Ed.), Talking to learn: Conversation in second
language acquisition (pp. 85-98). Rowley, MA: NewBury House.
Powell, R. R. (1994). Case studies of second-career secondary student teachers. International
Journal of Qualitative Studies in Education, 7(4), 351-366.
Rankin, J., & Becker, F. (2006). Does reading the research make a difference? A case study of
teacher growth in FL German. The Modern Language Journal, 90(3), 353-372.
Richards, J. C. (1998). What's the use of lesson plans? In J. C. Richards (Ed.), Beyond training
(pp. 103-121). Cambridge, UK: Cambridge University Press.
300

Richards, J. C., Gallo, P. B., & Renandya, W. A. (2001). Exploring teachers' beliefs and the
processes of change. The PAC Journal, 1(1), 43-64.
Richards, J. C., Ho, B., & Giblin, K. (1996). Learning how to teach in the RSA Cert. In D.
Freeman & J. C. Richards (Eds.), Teacher learning in language teaching (pp. 242-259).
Cambridge, UK: Cambridge University Press.
Richards, J. C., Li, B., & Tang, A. (1998). Exploring pedagogical reasoning skills. In J. C.
Richards (Ed.), Beyond training (pp. 86-102). Cambridge, UK: Cambridge University
Press.
Richards, J. C., & Lockhart, C. (1994). Reflective teaching in second language classrooms.
Cambridge, UK: Cambridge University Press.
Richards, J. C., & Nunan, D. (1990). Second language teacher education. Cambridge, UK:
Cambridge University Press.
Richards, J. C., & Pennington, M. C. (1998). The first year of teaching. In J. C. Richards (Ed.),
Beyond training (pp. 173-190). Cambridge, UK: Cambridge University Press.
Richards, K. (2003). Qualitative inquiry in TESOL. New York, NY: Palgrave Macmillan.
Richardson, V. (1996). The role of attitude and beliefs in learning to teach. In T. J. Buttery, E.
Guyton & J. P. Sikula (Eds.), Handbook of research on teacher education (2nd ed., pp.
102-119). New York, NY: Macmillan.
Roberts, M. A. (1995). Awareness and the efficacy of error correction. In R. W. Schmidt (Ed.),
Attention and awareness in foreign language learning (pp. 163-182). Manoa, HI: Second
Language Teaching Curriculum Center.
Rokeach, M. (1968). Beliefs, attitudes, and values: A theory of organization and change. San
Francisco: Jossey-Bass.
Russell, J., & Spada, N. (2006). The effectiveness of corrective feedback for the acquisition of
L2 grammar. In J. Norris & L. Ortega (Eds.), Synthesizing research on language learning
and teaching (pp. 133-164). Amsterdam, The Netherlands: John Benjamins.
Sato, K., & Kleinsasser, R. C. (1999). Communicative language teaching (CLT): Practical
understandings. The Modern Language Journal, 83(4), 494-517.
Schwartz, B. D. (1993). On explicit and negative data effecting and affecting competence and
301

linguistic behavior. Studies in Second Language Acquisition, 15, 147-163.
Scott, R., & Rodgers, B. (1995). Changing teachers' conceptions of teaching writing: A
collaborative study. Foreign Language Annals, 28(2), 234-246.
Sendan, F., & Roberts, J. (1998). Orhan: A case study in the development of a student teacher's
personal theories. Teachers and teaching: Theory and practice, 4(2), 229-244.
Sheen, R. (2005). Focus on forms as a means of improving accurate oral production. In A.
Housen & M. Pierrad (Eds.), Investigations in instructed second language acquisition (pp.
271-310). Berlin, Germany: Mouton de Gruyter
Sheen, Y. (2004). Corrective feedback and learner uptake in communicative classrooms across
instructional settings. Language Teaching Research, 8(3), 263-300.
Shkedi, A. (1998). Teachers' attitudes towards research: A challenge for qualitative researchers.
International Journal of Qualitative Studies in Education, 11(4), 559-577.
Skehan, P. (1998). A cognitive approach to language learning. Oxford, UK: Oxford University
Press.
Smith, D. B. (1996). Teacher decision making in the adult ESL classroom. In D. Freeman & J. C.
Richards (Eds.), Teacher learning in language teaching (pp. 197-216). Cambridge, UK:
Cambridge University Press.
Stenhouse, L. (1981). What counts as research? British Journal of Educational Studies, 29(2),
103-114.
Stewart, T. (2006). Teacher-researcher collaboration or teachers' research? TESOL Quarterly,
40(2), 421-429.
Swain, M. (2006). Verbal protocols: What does it mean for research to use speaking as a data
collection tool? In M. Chalhoub-Deville, C. A. Chapelle & P. A. Duff (Eds.), Inference
and generalizability in applied linguistics: Multiple research perspectives (pp. 97-113).
Amsterdam, The Netherlands: John Benjamins.
Tarone, E., & Allwright, D. (2005). Second language teacher learning and student second
language learning: Shaping the knowledge base. In D. J. Tedick (Ed.), Second language
teacher education: International perspectives (pp. 5-23). Mahwah, NJ: Lawrence Erlbaum.
Tarone, E., Swain, M., & Fathman, A. (1976). Some limitations to the classroom applications of
302

current second language acquisition. TESOL Quarterly, 10(1), 19-32.
Tatto, M. T. (1998). The influence of teacher education on teachers' beliefs about purposes of
education, roles, and practice. Journal of Teacher Education, 49(1), 66-77. doi:
10.1177/0022487198049001008
TESOL International Association. (2012). TESOL Quarterly. Retrieved from Author guidelines
website:
http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1545-7249/homepage/Society.html
Thomas, D. R. (2006). A General inductive approach for analyzing qualitative evaluation data.
American Journal of Evaluation, 27(2), 237-246. doi: 10.1177/1098214005283748
Truscott, J. (1996). The case against grammar correction in L2 writing classes. Language
Learning, 46(2), 327-369.
Truscott, J. (1999). What's wrong with oral grammar correction. The Canadian Modern
Language Review, 55(4), 437-456.
Tsui, A. B. M. (1996). Learning how to teach ESL writing. In D. Freeman & J. C. Richards
(Eds.), Teacher learning in language teaching (pp. 97-119). Cambridge, UK: Cambridge
University Press.
Urmston, A. W. (2003). Learning to teach English in Hong Kong: The opinions of teachers in
training. Language and Education, 17(2), 112-137.
Urmston, A. W., & Pennington, M. C. (2008). The beliefs and practices of novice teachers in
Hong Kong: Change and resistance to change in an Asian teaching context. In T. S. C.
Farrell (Ed.), Novice language teachers: Insights and perspectives for the first year (pp.
89-103). London: Equinox.
Van Fleet, A. (1979). Learning to teach: The cultural transmission analogy. Journal of Thought,
14(4), 281-290.
Vásquez, C., & Harvey, J. (2010). Raising teachers' awareness about corrective feedback through
research replication. Language Teaching Research, 14(4), 421-443. doi:
10.1177/1362168810375365
Weinstein, C. S. (1990). Prospective elementary teachers' beliefs about teaching: Implications for
teacher education. Teaching and Teacher Education, 6(3), 279-290.

303

Williams, J. (1999). Learner-generated attention to form. Language Learning, 49(4), 583-625.
Woods, D. (1996). Teacher cognition in language teaching: Beliefs, decision-making and
classroom practice. Cambridge, UK: Cambridge University Press.
Yorozuya, R. (1999). The types and effects of the teacher's error treatments in EFL classes The
report of a research project supported by grant-in-aid for scientific research (c). Tokyo:
The ministry of education, science, sports and culture, Japan.
Yoshida, R. (2010). How do teachers and learners perceive corrective feedback in the Japanese
language classroom? The Modern Language Journal, 94(2), 293-314.
Zeuli, J. S. (1994). How do teachers understand research when they read it? Teaching and
Teacher Education, 10(1), 39-55.
Zhao, S. Y., & Bitchener, J. (2007). Incidental focus on form in teacher–learner and learner–
learner interactions. System, 35(4), 431-447. doi: 10.1016/j.system.2007.04.004

304