125

Exploring the Effects of Teachers’ 
and Learners’ Conflicting Beliefs 
on the Provision of Corrective 
Feedback During Undisturbed 
Classroom Interactions1

Explorando los Efectos de las Creencias Conflictivas 
de Maestros y Alumnos Sobre la Provisión 
de Retroalimentación Correctiva Durante las 
Interacciones de Clase sin Perturbaciones

Edgar Emmanuell Garcia-Ponce, Irasema Mora-Pablo2*
Universidad de Guanajuato, México

Abstract
Extensive research literature suggests that corrective feedback (CF), when 
effective, has a beneficial impact on the development of learners’ interlanguage. 
This is because CF provides learners with language data concerning the 
correctness of their utterances and thus pushes their oral production towards 
greater clarity, accuracy and comprehensibility. However, CF has been found to 
be considerably scarce during classroom interactions. In an attempt to understand 
its scarcity, the present study investigates the interplay between the amount of 
CF provided by teachers and learner peers and the effects of their beliefs during 
uncontrolled classroom interactions at a Mexican university. By combining 
data collected from recorded classroom interactions, teacher interviews and 
learner focus groups, the findings show that there was a considerable number 
of errors which were avoided and not corrected by the teachers and learners 
during the classroom interactions. The findings also suggest that the scarcity of 
CF was in response to the teachers’ and learners’ conflicting beliefs about CF. 
This study provides a great opportunity to direct research towards the effects of 

1 Received: May 30th 2017/ Accepted: November 17th 2017
2  ee.garcia@ugto.mx , imora@ugto.mx 

LEARNERS’ VOICE IN ACADEMIC WRITING EFL Gist Education and Learning Research Journal. ISSN 1692-5777.
No.15. (July - December) 2017. pp. 125-148.

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



126

teachers’ and learners’ beliefs on providing and receiving CF during classroom 
interactions, and find ways through which a socio-affective climate can be 
promoted in the language classroom in order to facilitate the provision of CF.

Keywords: accuracy, classroom interactions, corrective feedback, 
English as a foreign language, teacher and learner beliefs 

Resumen
Un gran número de investigaciones han argumentado que la retroalimentación 
correctiva es de beneficio para el desarrollo del interlenguaje de los estudiantes. 
Esto se debe a que dicha retroalimentación facilita información sobre la precisión 
de sus estructuras lingüísticas y promueve que su producción oral sea más 
clara, precisa y comprensible. Sin embargo, estudios empíricos han encontrado 
que la retroalimentación correctiva es considerablemente escasa durante 
interacciones en el salón de clase. Con la intención de entender su escasez, el 
presente estudio indaga sobre la relación entre la cantidad de retroalimentación 
correctiva (facilitada por maestros y estudiantes) y sus creencias en salones 
de clase de una universidad mexicana. Al combinar datos recolectados de 
interacciones en salones de clase, entrevistas con maestros y grupos focales 
con alumnos, los resultados corroboran que hubo un número considerable de 
errores que no fueron atendidos y corregidos por los maestros y estudiantes 
durante las interacciones. Los hallazgos también sugieren que la escasez de 
la retroalimentación correctiva se debe a la influencia de sus creencias. Este 
estudio presenta una gran oportunidad de dirigir esfuerzos para investigar los 
efectos de las creencias de los maestros y estudiantes en la provisión de la 
retroalimentación correctiva y para encontrar alternativas que promuevan un 
ambiente socioafectivo en el salón de clase que permita la provisión efectiva de 
la retroalimentación correctiva.     

Palabras clave: creencias de maestros y estudiantes, retroalimentación 
correctiva, inglés como lengua extranjera, precisión, interacciones de 
clase.

Resumo
Um grande número de pesquisas tem argumentado que a retroalimentação 
corretiva é de benefício para o desenvolvimento da inter-linguagem dos 
estudantes. Isto se deve a que mencionada retroalimentação facilita informação 
sobre a precisão de suas estruturas linguísticas e promove que a sua produção 
oral seja mais clara, precisa e compreensível. Porém, estudos empíricos 
encontraram que a retroalimentação corretiva á consideravelmente escassa 
durante interações na sala de aula. Com a intenção de entender a sua escassez, 
o presente estudo indaga sobre a relação entre a quantidade de retroalimentação 
corretiva (facilitada por professores y estudantes) e suas crenças em salas 
de aula de uma universidade mexicana. Ao combinar dados recolhidos de 

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



127

interações em salas de aula, entrevistas com professores e grupos focais com 
alunos, os resultados corroboram que houve um número considerável de erros 
que não foram atendidos e corregidos pelos professores e estudantes durante as 
interações. As descobertas também sugerem que a escassez da retroalimentação 
corretiva se deve à influência das suas crenças. Este estudo apresenta uma 
grande oportunidade de dirigir esforços para pesquisar os efeitos das crenças 
dos professores e estudantes nas provisões da retroalimentação corretiva e para 
encontrar alternativas que promovam um ambiente sócio afetivo na sala de aula 
que permita a provisão efetiva da retroalimentação corretiva.     

Palavras chave: crenças de professores e estudantes, retroalimentação 
corretiva, inglês como língua estrangeira, precisão, interações de aula.

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



128

Introduction

The role of corrective feedback (henceforth CF) has been acknowledged in most second language theories and language pedagogy since it is viewed as language data that fosters 
linguistic accuracy and language learning (Ellis, 2009). Over the past 
three decades, extensive research has corroborated the beneficial effects 
of CF on learners’ language development (see Nicholas, Lightbown 
& Spada, 2001; Russel, 2009; Sheen, 2001). Based on this evidence, 
we can thus assume that there is no reason why teachers and learners 
should avoid providing CF in the language classroom. However, recent 
empirical studies suggest that the provision of CF is scarce during 
classroom interactions, despite teachers’ and learners’ stated value of 
it. It may seem possible that there is an interrelated set of instructional, 
interactional, practical and perceptual factors that compel teachers and 
learners to avoid providing and receiving this language data concerning 
the accuracy of their utterances.

By combining interactional (from recorded classroom 
interactions) and perceptual (from teacher interviews and learner focus 
groups) data, the present study aims to develop an understanding of the 
interplay between the amount of CF that is provided during classroom 
interactions led by teachers (teacher-led interactions) and learners (peer 
interactions) and the effects of teachers’ and learners’ beliefs. Due to 
practical constraints, the study is unable to encompass all the classroom 
factors that have effects on providing or avoiding CF. Rather, it intends to 
determine the impact of teachers’ and learners’ beliefs on the provision 
of CF, and thus provide new insights into how these perceptual factors 
can be oriented towards enhancing its provision during classroom 
interactions. The study is guided by three research questions (RQs):

RQ1: What is the amount of corrective feedback during 
uncontrolled classroom interactions led by the teachers and 
learner peers?

RQ2: To what extent do teachers’ and learners’ beliefs influence 
the provision of corrective feedback during these interactions?

RQ3: What can be learned from RQs 1 and 2 in order to enhance 
the provision of corrective feedback?

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



129

Literature review

In the classroom, there can be two types of feedback: positive and 
negative feedback. Positive feedback signals the veracity of the content 
of a learner’s response or the correctness of an utterance (Ellis, 2009). 
Negative feedback, on the other hand, is language information provided 
by teachers or learner peers for learners to signal that their utterance 
lacks veracity or linguistic correctness (Ellis, 2009; Walsh, 2011). In 
other words, this latter feedback is other-initiated repair and corrective 
in nature. CF consists of:

• an indication that an error has been committed,
• provision of the correct target language form, and
• metalinguistic information about the nature of the error, or any 

combination of these (Ellis, Loewen & Erlam, 2006; cited in 
Ellis, 2009).

Due to the asymmetrical roles in classroom interactions, the 
provision of CF is a ritual that is mostly initiated by teachers (Walsh, 
2006, 2011, 2013). Both language teachers and researchers have paid 
careful attention to CF because CF promotes language learning (see 
Ellis, 2009; Ellis, Loewen & Erlam, 2006; García Mayo & Pica, 2000; 
McDonough, 2004; Walsh, 2006, 2011, 2013; to name just a few). 
In particular, CF is claimed to provide learners with opportunities to 
metalinguistically reflect on the clarity, accuracy, and comprehensibility 
of their output (Martínez-Flor, 1999; McDonough, 2004), as well as 
opportunities to correct wrong language hypotheses and prevent errors 
from being fossilised (Swain, 2005). It has been also argued that CF can 
be beneficial – when its initiation and moves to provide it are embedded 
in a collaborative interaction during which teachers and learners provide 
jointly owned affordances to solve linguistic problems (Rassaei, 2014; 
Swain & Susuki, 2008). 

Despite arguments that there is no reason why erroneous 
utterances should not be corrected in L2 classrooms, language teachers 
and researchers frequently disagree on the following conflicting actions 
regarding the provision of CF:

1. What errors to correct.
2. How and when to correct errors.
3. Whether to correct errors, interrupt the interaction and avoid 

interlanguage fossilization.
4. Whether to omit the error, continue with the interaction and 

maintain learners’ face.

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



130

It has been found that the fourth action is motivated when learners 
perceive CF as face-threatening (Yoshida, 2013a), evaluative (Allwright 
& Bailey, 1991), or a communication failure (Tsui, 1995). Learners’ 
perceptions of CF as face-threatening information, in influencing 
emotions, have significant effects on their self-concepts and perceived 
self-efficacy (e.g., learners’ self-perceptions of limited linguistic 
competence, poor pronunciation, limited vocabulary, etc.), which in 
some cases may deter them from fully participating and thus developing 
the target language (Tsui, 1995; Wesely, 2012). In light of the possibility 
that CF during classroom interactions may be perceived by learners as 
face-threatening and thus limit their oral production, research literature 
has suggested alternative techniques for providing learners with CF 
or information concerning their accuracy. For example, Hendrickson 
(1978) suggests that teachers should only correct those errors that 1) 
hinder communication significantly; 2) have highly stigmatising effects; 
and 3) occur frequently in learners’ speech. Tsui (1995) warns that 
teachers should not correct every error since it may discourage learners 
from answering questions and participating in future interactions. 
However, the immediate issue that emerges from an avoidance approach 
to providing CF is that learners’ opportunities to develop metalinguistic 
knowledge and push their utterances towards greater accuracy would 
be limited. In particular, this approach would also limit negotiations for 
meaning during which (implicit or explicit) CF is facilitated.

It seems possible that the provision of CF in the language classroom 
is enhanced if research efforts are oriented towards understanding the 
perceptual factors that motivate learners’ perceptions of CF as face-
threatening and thus limitations in its provision. Therefore, the present 
study attempts to gain insights into the interplay between the amount of 
CF in uncontrolled classroom interactions and teachers’ and learners’ 
beliefs about CF with a view to enhancing the opportunities to provide 
learners with more effective CF in the language classroom.

Methodology

The present study resided in an exploratory and naturalistic 
inquiry which adopts a multiple data-gathering approach with a view 
to developing an understanding of the participants’ classroom practices 
and meanings (Lillis, 2008). The data collection instruments were:

For interactional data: 1) recorded classroom interactions

For perceptual data: 2) teacher interviews and 3) learner focus 
groups 

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



131

The rationale behind the use of the above instruments is that 
interactional and perceptual data provide insights into the interplay 
between classroom interactional behaviour and beliefs (Wesely, 2012). 
Moreover, these research instruments not only allow a thick description 
of what may prove to be potentially significant, but also help researchers 
maintain an openness to what may be important to the participants 
(Lillis, 2008).

Research context and participants

The present study took place at a university located in the centre 
of Mexico. In this teaching and learning context, learners take subjects 
which train them to become language teachers or translators after a 
five-year BA programme. However, it is common in this context that 
most learners opt to major in EFL teaching. Learners are also expected 
to learn English throughout the programme.

Specifically, the study was conducted in three classrooms at 
basic, intermediate, and advanced levels. Learners in courses at basic 
and intermediate levels practise English during six hours per week. 
At advanced levels, learners study English during five hours per 
week. The reason why this university decided to reduce the number 
of hours at advanced levels was to encourage learner autonomy 
outside the classroom. The total number of learners who voluntarily 
accepted to participate in the study was 63 (17 at the basic level; 26 at 
the intermediate level; and 20 at the advanced level). All the learners 
were Mexican, their age ranged from 18 to 24 years, and their L1 
was Spanish. Most of the learners had studied English before starting 
their university studies. Some of them came from state schools, where 
language exposure ranges from four to five hours per week. A low 
number of learners had studied English in private schools, where the 
language is practiced from 15 to 20 hours per week. The teachers were 
also invited to participate in the study and accepted under no obligation. 
The three teachers were women, born and raised in Mexico, and their 
mother tongue was Spanish. They all stated that they had been learning 
English for 18 or more years, and teaching it for 11 or more years. 

Data collection instruments

Classroom interactions at the three proficiency levels are claimed 
to provide a detailed and comprehensive description of teachers’ and 
learners’ interactional behaviour in a naturally-occurring way (Larsen-
Freeman & Long, 1991). Following this claim, the interactions were 

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



132

recorded in two sessions of 100 minutes approximately at each 
proficiency level. In total, 600 minutes of classroom interactions were 
recorded, 200 minutes approximately from each proficiency level. 
After having recorded the interactions, the data were then transcribed 
completely, and segmented into teacher-led interactions (TLIs) and 
peer interactions (PIs).

Teacher interviews were included in this study in order to 
understand how the teacher participants make sense of the provision of 
CF in relation to the context which they inhabit (Snape & Spencer, 2003). 
The teacher interviews were conducted by one of the researchers, after 
the recorded classroom interactions. In order to guide the interviews, 
a list of ten questions was used to understand the teachers’ beliefs and 
practices regarding CF. In the case of the learner focus groups, Gibbs 
(1997) maintains that they allow insights into people’s beliefs, attitudes 
and values from individual as well as group perspectives. This was of 
particular relevance for the study which seeks to investigate the extent to 
which learners’ beliefs influence the provision of CF. Thus, three focus 
groups were carried out with five learners from each proficiency level. 
They were selected randomly from the teachers’ attendance list, and 
invited to participate free of obligation. They all agreed to participate 
in the focus groups which were arranged at their convenience. A list 
of 10 questions was used to facilitate and guide the oral interactions 
between the researcher and learners. The teacher interviews and learner 
focus groups lasted around 20 minutes. They both were conducted in 
Spanish so as to avoid the learners’ anxiety about the correctness of 
their utterances in the L2. For analysis purposes, the oral interactions 
during the interviews and focus groups were recorded and transcribed 
in their entirety.

All participants were informed of their right to withdraw at 
any time, and provided consent to participate. Complying with their 
right to be protected from identification, the learners’ names and 
identities were carefully anonymised in the data. Instead, abbreviations 
and pseudonyms are used. The word ‘Learner’ (or letter ‘L’) and an 
identification number, e.g., Learner 5 are used to refer to learners in 
the extracts. In the case of teachers, pseudonyms are also used to refer 
to the teachers: María for the teacher at the basic level, Tanya for the 
teacher at the intermediate level, and Aranza for the teacher at the 
advanced level.

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



133

Data analysis and interpretation

According to Ellis (2006, 2009) and Lyster (2004), moves to 
provide CF can be classified in the following way:

Table 1. Taxonomy of CF moves
 

 Input-providing Recast

 Output-prompting Repetition

   Clarification request

Therefore, the study investigates the incidence of recasts, 
repetitions and clarification requests that were initiated during the 
TLIs and PIs at the three proficiency levels. In order to identify these 
moves in the uncontrolled TLIs and PIs, it was necessary to find the CF 
episodes which involve a trigger, feedback and (optionally) uptake (see 
Ellis, 2009), as shown in the following example:

Trigger L11: //The woman … call a taxi//
Feedback T: the woman?
Potential Uptake L11: //Calls a taxi//

Once identified, we classified each feedback move following the 
specifications summarised in Table 2.

Table 2. CF moves

Move

1.Clarification 
requests

2. Repetitions

3. Recasts

Specification

These are mostly wh- or bipolar questions which are initiated 
to elicit clarification or new information of the interlocutor’s 
preceding utterance(s) (Long, 1980; Ellis, 2009; Tedick & 
Gortari, 1998).

Language information which is provided to reshape another 
speaker’s utterance (Ellis, 2009; Tedick & Gortari, 1998). These 
are the most common types of feedback moves which usually 
contain an additional feature, for example, stress or lengthening 
of a segment, questioning intonation, etc. (Chaudron, 1988).

These are reformulations which are initiated to reshape or refine 
all or part of others’ utterances (Walsh, 2006). Recasts need to 
1) contain content words of a preceding incorrect utterance; 2) 
reshape utterances in a phonological, syntactic, morphological 
or lexical way (Braidi, 2002); and 3) focus on meaning rather 
than form (Long & Robinson, 1998). 

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



134

Because the purpose of the study was not to test hypotheses but 
to explore the incidence of CF, the interactional data were calculated 
using simple totals and ratios. Firstly, the total numbers were obtained 
by counting the instances when CF was provided and the moves that 
triggered them. Secondly, ratios were calculated by dividing the total 
number of each move per the total minutes of each TLI or PI. In order 
to understand the amount of errors that were attended during the 
interactions, we also counted the total number of errors, and calculated 
ratios by dividing the total number of errors per the total minutes of 
each TLI and PI. Prior to this analysis, we needed to establish what 
constituted an error. The following criteria were then coded for 
identifying and counting errors:
• Errors in word selection
• Errors in morphology
• Errors in syntax

• Errors in pronunciation

• False starts, hesitations and self-corrections were excluded. 

The perceptual data from the interviews and focus groups were 
analysed following a meaning categorisation which is believed to 
facilitate the identification of patterns, themes and meaning (Berg, 
2009). This involved identifying extracts manually, and attributing 
them to theme categories and sub-categories which emerged from the 
data:

Perceptions about CF
• Positive attitudes towards CF
• Perceived benefits of CF
• Negative attitudes towards CF
• Teachers avoiding corrections

Perceived nature of CF
• Types of CF moves
• CF at word level
• No perceived need for CF
• CF moves not initiated by learners
• Face-threatening CF
• More CF episodes in PIs

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



135

Results

In order to address RQ1 (i.e., what is the amount of corrective 
feedback during uncontrolled classroom interactions led by the teachers 
and learner peers?) and RQ2 (i.e., to what extent do teachers’ and 
learners’ beliefs influence the provision of corrective feedback during 
these interactions?), this section firstly discusses the findings into 
the amount of CF at the three proficiency levels. It then explores the 
influence that the teachers’ and learners’ beliefs exerted on this amount. 
Overall, the interactional evidence shows a considerable proportion of 
errors which went unnoticed or omitted by the teachers and learner peers, 
and a low number of CF moves to address them during the interactions 
at the three proficiency levels. The perceptual data suggests that the 
teachers’ and learners’ beliefs about CF are conflicting and important 
perceptual factors that heavily influenced the provision of CF.   

Results of interactional data

The findings of the interactional data are summarised in tables, 
and are presented by kind of interaction (TLI or PI) at each proficiency 
level. The first two tables show the number of errors per minute and 
ratios of CF moves at the basic level:

Table 3. Number of errors and CF moves during TLIs at basic level

Table 3 shows that the number of CF moves per minute tends 
to be low compared to the number of errors per minute that the basic 
learners committed during the five TLIs. It can also be seen in this table 
that there is a trend towards clarification requests, ranging from 0.1 to 
1.3 clarification requests per minute. However, in some TLIs, there is 
an absence of some moves. For example, in TLIs 1 and 3, there is a 
lack of repetitions. In the case of TLIs 1, 2 and 4, there is omission of 
recasts. It was only during TLIs 2-4 that the basic teacher and learners 
attended to all the errors. However, as we will see in the remainder of 

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



136

this section, the interactional data indicate that there was a high number 
of errors which were unnoticed or omitted, but a considerable scarcity 
of CF moves during the TLIs and PIs at the intermediate and advanced 
levels. This was also the case of the basic PIs:  

Table 4. Number of errors and CF moves during PIs at basic 
level 

  

As in the TLIs, Table 4 shows that there is a high proportion 
of errors per minute to which a low amount of CF was provided, as 
indicated by the CF moves per minute during these interactions led by 
learner peers. For example, during PI 1, 3.6 errors per minute were 
committed, but only 0.6 clarification requests per minute were initiated 
to address them. Moreover, there is a tendency towards a greater number 
of clarification requests (a range of 0.6 to 1.5 clarification requests per 
minute) than repetitions and recasts. It can also be seen from this table 
that there is a scarcity of repetitions, and a tendency of recasts in some 
PIs (PIs 2-5). 

Similar to the TLIs and PIs at the basic level, Tables 5 and 6 show 
that despite a high number of errors, there was a considerable scarcity 
of some CF moves across the intermediate interactions.  

Table 5. Number of errors and CF moves during TLIs at intermediate 
level 

  

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



137

In Table 5, it is evident that the number of CF moves per minute is 
lower than the number of errors per minute, indicating again that there 
was a high number of errors which were omitted by the intermediate 
teacher and learners. This table also shows that there is a low proportion 
of CF moves across the intermediate TLIs, with a clear tendency 
towards clarification requests. In the case of the PIs, the data indicates 
a greater scarcity of CF moves than in the TLIs: 

Table 6. Number of errors and CF moves during PIs at intermediate 
level 

  

As in previous interactions, Table 6 indicates that there is a low 
number of CF moves compared to the number of errors that were 
committed during these learner-led discussions. Due to the high 
number of errors during these and previous teacher- and learner-led 
interactions, it seems possible that there are more pressing factors that 
compel these teachers and learners to avoid providing CF during the 
interactions despite the high number of errors. It can also be seen in 
Table 6 that there was a low number of CF moves per minute across the 
six PIs. Again, clarification requests tended to be initiated more than 
repetitions and recasts. 

The following table shows the absence of errors and therefore the 
initiation of CF moves at the advanced TLIs: 

 Table 7. Number of errors and CF moves during TLIs at advanced level

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



138

In exploring these two TLIs, it was possible to observe the lack of 
learners’ opportunities to contribute to the classroom discourse with freer 
and more elaborate utterances. It was the teacher who controlled these 
interactions; learners only had opportunities to respond to the teacher’s 
elicitations centred on grammar practice. This lack of opportunities to 
contribute to the classroom discourse with more creative utterances may 
explain the absence of errors and thus CF moves. This suggestion is 
supported by the advanced PIs, during which learners were responsible 
of the discourse, and had more opportunities to experiment with the 
language, having perhaps an impact on the number of errors that they 
committed.

Table 8. Number of errors and CF moves during PIs at advanced 
level 

As shown in Table 8, there is a considerable amount of errors 
for which a limited number of CF moves was initiated. For example, 
7.7 and 7.6 errors per minute were committed during PIs 1 and 2, 
respectively, but only 0.1 and 0.5 CF moves per minute were initiated 
to attend to these errors. As at previous proficiency levels, the advanced 
learners during these interactions opted to initiate clarification requests.

Overall, the above interactional data firstly showed a high 
proportion of errors which was not attended. In other words, there was a 
high number of errors which went unnoticed or avoided by the teachers 
and learner peers. The issue that emerges from this evidence is that the 
learners had fewer opportunities to notice erroneous utterances and thus 
push their oral production to be more accurate and comprehensible. 

The data also indicated that there was a varied, but low number 
of CF moves initiated by the teachers and learner peers across the 
interactions at the three proficiency levels. Among the CF moves, it 
was clarification requests which tended to be initiated. It is possible 
that the teachers and learner peers relied more on clarification requests 
than repetitions and recasts because clarification requests were used as 
moves which provided CF but in an indirect way. That is, when an 

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



139

error was committed, the teachers and learner peers tended to initiate 
elicitations to signal that there was the presence of an erroneous 
utterance, but without modifying or correcting the utterance as in the 
case of repetitions and recasts. 

It seems possible that there are more pressing concerns or 
factors that compelled these teachers and learners not to initiate CF 
moves to attend to the high number of errors. Particularly, the teachers’ 
and learners’ beliefs as important perceptual factors may reveal the 
motivation behind this avoidance strategy, and possible ways through 
which the teachers and learners can be assisted in providing CF and thus 
benefitting from greater opportunities to develop the target language. 
The following section attempts to address this.

Teachers’ and learners’ beliefs about CF

During the teacher interviews and learner focus groups, various 
points concerning CF were suggested by the participant teachers and 
learners. In general, the three teachers claimed that they embrace the 
value of CF as a strategy for teaching and learning the target language, 
for example:

Extract 1 Quote by María (basic level)

“It [CF] may be significant for them, like having an alarm 
to correct. Then, they can produce the same sentence 
and if they make the same mistake, they will be able to 
correct it.”

Extract 2 Quote by Tanya (intermediate level)

“It is a matter of giving you my [corrective] feedback so 
that you in the future see which one is the standard. Then, 
making for the whole class, you realise that the learners 
are aware and say: ‘I can use this in this situation, and the 
other in another situation’ and all the class benefits from 
this [feedback].”

Maria’s and Tanya’s statements clearly reveal their positive 
attitudes towards providing CF during classroom interactions. These 
statements also indicate perceived benefits for learners’ self-corrections. 
Interestingly, they perceived that CF was beneficial not only for the 
learners to whom corrections are directed, but also to the whole class 

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



140

(for a discussion about this, see Havranek, 2002; Muranoi, 2000). The 15 
learners during the focus groups also suggested responses which reveal 
positive attitudes towards providing and receiving CF. For example, 
Learner 5 (basic level) said: “I think it is good that she corrects us.”

However, the responses of the three teachers, one learner at the 
basic level and the five learners at the advanced level suggested that CF 
was scarce or absent during classroom interactions, as borne out by the 
interactional data. For example, in “we need that the teacher starts to 
correct us,” Learner 1’s (basic level) suggestion points to a perceived 
scarcity of CF during the classroom discussions. This perceived scarcity 
is corroborated by Aranza’s statement: “I seldom correct while they are 
speaking, […] I rarely correct them during classroom discussions.” 
The teachers’ responses point to one main reason that motivated this 
avoidance:

Extract 3 Quote by Aranza (advanced level)

“Maybe they are fluent but with many mistakes. Thus, I 
have decided not to correct them so as not to affect”

Extract 4 Quote by Tanya (intermediate level)

“It is give them something positive, something not 
very positive and not tell them that their speaking was 
wrong, you may inhibit them and you could spoil the 
interactions.”

Aranza’s and Tanya’s statements suggest the feeling that correcting 
learners’ oral mistakes inhibited them from speaking, as indicated in 
“not tell them that their speaking was wrong, you may inhibit them, and 
you could spoil the interactions.” The feeling that CF inhibited learners 
from interacting was shared by the learners during the focus groups, for 
example:

Extract 5 Quote by Learner 4 (advanced level)

“Some people may feel pressed while talking to the 
teacher for fear of being corrected or something like 
that.”

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



141

Again, Learner 4’s statement suggests a feeling that CF had 
a negative impact on learners. It thus appears that the teachers’ and 
learners’ beliefs about CF were conflicting. That is, the teachers’ and 
learners’ beliefs about the importance of CF appear to have conflicted 
with their beliefs about negative effects of it on learners’ oral production, 
as suggested in Aranza’s statement: “it is funny because everybody 
agrees to be corrected, but when you do correct them, they [learners] 
don’t like it that much.” As indicated in “I have decided not to correct 
them so as not to affect [speaking]” (Aranza, Extract 3), it seems that 
these conflicting beliefs influenced Aranza’s teaching decisions not to 
correct learners’ oral mistakes. The other two teachers’ responses also 
suggest teaching decisions influenced by these conflicting beliefs:

Extract 6 Quote by Tanya (intermediate level)

“They perceive it negatively and take it personal, like 
exposing them. There are people who take it (corrections) 
personal […] you need to find like tactics, it is a delicate 
topic.”

Tanya’s explanation again points to a perception that CF had 
negative effects on learners, even at a personal level. As suggested in 
“you need to find like tactics, it is a delicate topic,” we see a perception 
that the conflicting beliefs around CF influenced her teaching decisions. 
The following two extracts suggest how the teachers’ and learners’ 
conflicting beliefs around CF influenced the basic and intermediate 
teachers’ teaching and interactional behaviour:

Extract 7 Quote by María (basic level)

“Depending on the intimacy for them to express, interact 
and tell them at the end [of the classroom discussion] 
where they were wrong.”

Extract 8 Quote by Tanya (intermediate level)

“The provision of feedback is personalised and without 
other learners […] Then, you have the freedom to tell 
them their mistakes and advise them.” “[…] I now do it 
in a personalised way so as to avoid peer criticisms.”

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



142

As suggested in María’s and Tanya’s statements, we again 
see beliefs that CF was perceived as face-threatening (as implied in 
“depending on the intimacy for them to express”), and had a negative 
impact on learners (as indicated in “I now do it in a personalised way 
so as to avoid peer criticisms”). These beliefs appear to have influenced 
the teachers’ teaching decisions to avoid providing CF during the 
classroom discussions, as indicated in “I now do it in a personalised 
way” and “tell them at the end [of the classroom discussions] where 
they were wrong.”

Discussions

In exploring the interactional data, it was found that there was a 
high number of errors that were unnoticed or omitted by both the teachers 
and learner peers during the classroom interactions. The number of CF 
moves was considerable low during the discussions led by the teachers 
and learner peers across the three proficiency levels. Among the three 
CF moves explored in this study, it was clarification requests which 
the teachers and learners tended to initiate during the discussions. The 
study was unable to determine the extent to which these CF moves were 
effective in providing feedback which leads to uptake. However, we put 
forward the argument that the teachers and learners decided to initiate 
mostly clarification requests since they involve less face-threatening 
interactional work which, implicitly, encourage learners to reshape 
their own erroneous or unclear utterances. It is possible that the other 
two CF moves were perceived by the teachers and learner peers as face-
threatening moves. 

Based upon the evidence that CF moves were scarce during 
classroom discussions, the study found that the teachers’ and learners’ 
beliefs about CF were conflicting. That is, the teachers and learners 
valued the role of CF, but it was perceived by both teachers and learners 
to inhibit learners and thus limit their oral production. This thus implies 
that providing information concerning the correctness of learners’ 
utterances and thus push them towards greater accuracy may have been 
perceived as face-threatening or as a sign of incompetence to speak 
the target language. This is in accord with Cathcart & Olsen (1976) 
and Allwright & Bailey (1991), who also found classroom perceptions 
of oral corrections as face-threatening, despite the fact that learners 
claimed to value them (Cathcart & Olsen, 1976). However, the findings 
of this study suggest that it was actually the teachers’ and learners’ 
conflicting beliefs about CF which influenced teachers’ and learners’ 
behaviour, resulting in an avoidance strategy to save the learners’ face.

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



143

Based on the above interactional and perceptual data, and in 
addressing RQ3 (i.e., what can be learned from RQs 1 and 2 in order 
to enhance the provision of corrective feedback?), it thus seems that 
the frequency and effectiveness of CF and its moves reside not only in 
teachers’ and learners’ value of them as opportunities which promote 
language learning, but also in their willingness to provide and receive 
strategic and supportive language data concerning the correctness 
of their utterances and thus promote others’ language development. 
Drawing on the evidence that the teachers and learners of the study were 
able to reflect on their teaching, learning and interactional behaviour 
during classroom discussions, it may seem possible that this reflection 
ability may be directed towards the socio-affective climate between 
teachers and learners and learner peers, and the ways through which 
they can be encouraged to develop a positive attitude and behaviour 
towards initiating and receiving CF. This suggestion is supported by 
Naughton (2006), who contends that the most relevant classroom 
discussions to interlanguage development is that in which teachers and 
learners share a need and desire to understand each other, and learn 
from them. As suggested by Naughton (2006), classroom discussions 
can be exploited through a classroom climate in which challenging or 
modifying others’ utterances is not social taboo. Under these conflicting 
circumstances, Yang and Kim (2011) raise the need to align teachers’ 
and learners’ beliefs with interactional behaviour that is more effective 
for classroom practices. In line with this, Yoshida (2013) contends that 
teachers and learners can be assisted in breaking away from classroom 
behaviour influenced by their conflicting beliefs in order to promote 
the development and appropriation of new beliefs consistent with more 
effective teaching and learning practices. We might thus explore the 
possibility that opportunities to provide CF can be enhanced if teachers 
and learners are assisted in mediating their beliefs and other cognitive 
factors through awareness-raising processes (e.g., advice from tutors on 
more effective interactional behaviour, or reflective procedures) (see, 
for example, Navarro & Thornton, 2011; Yang & Kim, 2011; Yoshida, 
2013). These processes can assist them in raising an awareness of 
the interplay between actions and beliefs, resulting in co-constructed 
beliefs which have a beneficial impact on the socio-affective climate 
of classrooms and teachers’ and learners’ interactional behaviour 
(Li & Lian, 2012; Yang & Kim, 2011; Yoshida, 2013), in this case, 
opportunities to provide and receive CF in both TLIs and PIs. 

In the language classroom, teachers need not abandon the 
provision of CF; its use during classroom interactions maximises 
learners’ opportunities to be exposed to information concerning the 

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



144

accuracy of their utterances (Rassaei, 2014). Teachers should make 
a conscious use of feedback in relation to the pedagogic goal of the 
moment (Tsui, 1995; Walsh, 2013). That is, teachers need to be aware 
of the effects of these moves, and use them depending on the aim of the 
teaching practice. In order to avoid learners’ loss of face, Rassaei (2014) 
suggests that the provision of CF needs to be performed collaboratively, 
in a way that encourages learners to produce language and assists them 
in negotiating and solving their erroneous utterances.

Conclusions and further research

The primary aim of the present study was to explore the interplay 
between the amount of CF and teacher and learner beliefs around this 
corrective language data. The study resided in a naturalistic as well as 
exploratory enquiry in order to understand better teaching and learning 
practices without controlling classroom variables.

The study firstly found that the amount of CF, indicated by 
clarification requests, repetitions and recasts, was scarce or absent 
during the classroom discussions led by the teachers and learner peers, 
despite the high number of errors that were identified in the interactional 
data. In an attempt to understand this avoidance, the study found that 
the teachers’ and learners’ beliefs concerning CF were conflicting, and 
influential on teaching behaviour by avoiding these moves perceived by 
the teachers and learners as face-threatening. Based on this evidence, 
the study puts forward the argument that in cases of low amount of 
CF, teachers’ and researchers’ attention should be directed towards 
the socio-affective climate in the language classroom where CF is 
seen not only as beneficial, but also as necessary to promote learners’ 
interlanguage development.

As in any exploratory study, further research should be conducted 
in order to generalise from the findings of this study. Research should 
explore the interplay between the nature of CF and the influence of 
teachers’ and learners’ beliefs in teaching practices with several aims. It 
would be interesting to know whether awareness-raising processes have 
a beneficial impact on the amount and quality of CF during classroom 
interactions. However, it is hoped that this study is useful for teachers 
and learners who endorse CF and its moved, but struggle to initiate 
them during classroom discourse.

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



145

References

Allwright, D. & Bailey, K. M. (1991). Focus on the language learner. 
Cambridge: Cambridge University Press.

Braidi, S. M. (2002). Reexamining the role of recasts in native-speaker/
nonnative-speaker interactions. Language Learning, 52, 1-42

Cathcart, R. L. & Olsen J. W. B. (1976). Teachers’ and students’ 
preferences for correction of classroom conversation errors. In J. F. 
Fanselow & R. H. Crymes (Eds.), On TESOL ’76, 41-45. Washington, 
D.C.: TESOL.

Chaudron, C. (1988). Second language classrooms. New York: 
Cambridge University Press.

Ellis, R. (2006). Researching the effects of form-focused instruction on 
L2 acquisition. In K. Bardovi-Harlig & Z. Dornyei (Eds.), Themes in 
SLA Research. AILA, 19, 18-41. Amsterdam: John Benjamins.

Ellis, R. (2009). CF and teacher development. L2 Journal, 1, 3-18.

Ellis, R., Loewen, S., & Erlam, R. (2005). Implicit and explicit CF 
and the acquisition of L2 grammar. Studies in Second Language 
Acquisition, 28, 339-368.

García Mayo, M. P. & Pica, T. (2000). L2 learner interaction in a foreign 
language setting: Are learning needs addressed? IRAL: International 
Review of Applied Linguistics, 38(1), 35-58.

Gibbs, A. (1997). Focus Groups. Social Research Update, 19, Winter, 
Department of Sociology, University of Surrey. Retrieved January 
12, 2016, from http://sru.soc.surrey.ac.uk/SRU19.html 

Havranek, G. (2002). When is negative feedback likely to succeed? 
International Journal of Educational Research, 37(3–4), 255–270.

Hendrickson, J. M. (1978). Error correction in foreign language 
teaching: recent theory, research, and practice. Modern Language 
Journal, 62, 387-398.

Larsen-Freeman, D. & Long, M. H. (1991). An Introduction to Second 
Language Acquisition Research. London: Pearson.

Li, S., & Liang, W. (2012). The dynamic nature of learner beliefs: the 
relationship between beliefs about EFL learning and proficiency in a 
Chinese context. System, 14(1), 177-211.

Lillis, T. M. (2008). Ethnography as Method, Methodology, and “Deep 
Theorizing”: Closing the Gap between Text and Context in Academic 
Writing Research. Written Communication, 25(3), 353–388.

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



146

Long, M. H. & Robinson, P. (1998). Focus on form: Theory, research 
and practice. In C. Doughty & J. Williams (Eds.), Focus on form 
in second language acquisition (pp. 15-41). Cambridge, England: 
Cambridge University Press.

Long, M. H. (1980). Input, interaction, and second language acquisition. 
Unpublished doctoral dissertation, UCLA, Department of Applied 
Linguistics and TESL.

Lyster, R. (2004). Differential effects of prompts and recasts in form-
focused instruction. Studies in Second Language Acquisition, 26, 
399-432.

Martínez-Flor, A. (1999). Does Language Learners Benefit from 
Classroom Interaction? Jornades de Foment de la Investigació, 
Universitat Jaume I.

McDonough, K. (2004). Learner-learner interaction during pair and 
small group activities in a Thai EFL context. System, 32(2), 207–224.

Muranoi, H. (2000). Focus on form through interaction enhancement: 
Integrating formal instruction into a communicative task in EFL 
classrooms. Language Learning, 50(4): 617–673.

Naughton, D. (2006). Cooperative strategy training and oral interaction: 
Enhancing small group communication in the language classroom. 
The Modern Language Journal, 90, 169-184.

Navarro, D. & Thornton, K. (2011). Investigating the relationship 
between belief and action in self-directed language learning. System, 
39(3), 290–301.

Nicholas, H., Lightbown, P. M. & Spada, N. (2001). Recasts as feedback 
to language learners. Language Learning 51(4): 719.758.

Philp, J. & Tognini, R. (2009). Language acquisition in foreign language 
contexts and the differential benefits of interaction. IRAL, 47(3/4): 
245–66.

Rassaei, E. (2014). Scaffolded feedback, recasts, and L2 development: 
A sociocultural perspective. The Modern Language Journal, 98, 
417–431.

Russell, V. (2009). Corrective feedback, over a decade of research 
since Lyster and Ranta (1997): Where do we stand today? Electronic 
Journal of Foreign Language Teaching 6(1): 21-31.

Sheen, Y. (2010). Differential effects of oral and written corrective 
feedback in the ESL classroom. Studies in Second Language 
Acquisition 32(2): 203-234.

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



147

Snape, D. & Spencer, L. (2003). The foundations of qualitative research. 
In J. Ritchie & J. Lewis (Eds.), Qualitative research practice. A guide 
for social science students and researchers (pp. 1–23). London: Sage 
Publications Ltd.

Swain, M. & Suzuki, W. (2008). Interaction, output, and communicative 
language learning. In B. Spolsky & F. M. Hult (Eds.), The handbook 
of educational linguistics (pp. 557–568). Oxford, UK: Blackwell.

Tedick, D. & de Gortari, B. (1998). Research on error correction and 
implications for classroom teaching. The ACIE Newsletter, 1(3).

Tsui, A. B. M. (1995). Introducing classroom interaction. London: 
Penguin English.

Walsh, S. (2006). Investigating classroom discourse. London: 
Routledge.

Walsh, S. (2011). Exploring classroom discourse: language in action. 
London: Routledge.

Walsh, S. (2013). Classroom discourse and teacher development. 
Edinburgh: Edinburgh University Press.

Wesely, P. (2012). Learner attitudes, perceptions, and beliefs in language 
learning. Foreign Language Annals, 45(S1), 98-117.

Yang, J. S. & Kim, T.Y. (2011). Sociocultural analysis of second 
language learner beliefs: A qualitative case study of two study-
abroad ESL learners. System, 39(3), 325–334.

Yoshida, R. (2013). Conflict between learners’ beliefs and actions: 
Speaking in the classroom. Language Awareness, 22(4), 371-388.

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS GARCÍA-PONCE & MORA-PABLOEFL

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)



148

Authors 

*Edgar Emmanuell García holds an MA in Applied Linguistics 
(UAEMex, Mexico), and a PhD in English Language Teaching 
and Applied Linguistics (University of Birmingham, UK). His 
research interests are centred on the interplay between classroom 
interactions and teacher and learner beliefs. He currently 
teaches in BA in TESOL and MA in Applied Linguistics of 
English Language Teaching at the Departamento de Lenguas, 
Universidad de Guanajuato.

Irasema Mora Pablo holds an MA in Applied Linguistics 
(Universidad de las Américas-Puebla) and a PhD in Applied 
Linguistics (University of Kent, UK). She coordinates the the 
Master’s in Applied Linguistics of English Language Teaching at 
the Universidad de Guanajuato. She is a member of the National 
System of Researchers (Sistema Nacional de Investigadores) 
and her research interests are bilingual studies, identity and 
nativespeakerism.

CORRECTIVE FEEDBACK DURING CLASSROOM INTERACTIONS

                No. 15 (July - December, 2017)     No. 15 (July - December, 2017)