Edith Hernández Méndez^*
María del Rosario Reyes Cruz^**
Universidad de Quintana Roo, Mexico
^*edith@uqroo.mx
^**rosreyes@uqroo.mx

This article was received on December 6, 2011, and accepted on May 14, 2012.

Corrective feedback has been discussed mainly in second language acquisition contexts, but less has been done concerning corrective feedback in foreign language settings. In this descriptive study, conducted at a Mexican university, our aims were to identify the perceptions of instructors of English as a foreign language about corrective feedback and its actual practice in their classrooms. A semi-structured interview and a questionnaire were used to collect the data. The results show that teachers in general have a positive perception of oral corrective feedback. However, some consider it as optional because instructors are very concerned with students' feelings and emotions. Unfocused oral corrective feedback and implicit strategies are predominant in practice. Corrective feedback provided by the instructor is preferred to that provided by peers. Self-correction is the least popular.

Key words: Corrective feedback, EFL, perceptions, practice.

La retroalimentación correctiva se ha discutido principalmente en contextos de adquisición de segundas lenguas, pero poco se ha hecho en el área de lenguas extranjeras. Esta investigación descriptiva, realizada en una universidad mexicana, tuvo como objetivo identificar las percepciones de profesores de inglés como lengua extranjera sobre retroalimentación correctiva y su práctica. Para la recolección de datos se usaron una entrevista semiestructurada y un cuestionario. Los resultados muestran que si bien los profesores en general tienen una percepción positiva sobre la retroalimentación correctiva oral, algunos la consideran opcional, pues les preocupan los sentimientos y emociones de los estudiantes. En la práctica predominan la retroalimentación correctiva oral no enfocada y las estrategias implícitas. Asimismo, se prefiere la retroalimentación correctiva que ofrece el docente y la autocorrección es la menos común.

Palabras clave: inglés lengua extranjera, percepciones, práctica, retroalimentación correctiva.

Errors in most cultures are seen as something we should avoid or prevent, as errors can be the cause even of unfortunate events. To deal with them, then, is not easy. When talking about errors in language learning or language acquisition, we cannot help but become part of a very controversial topic, either on the theoretical or methodological (pedagogical) side.

Han (2008) suggests that error correction implies an evident and direct correction, whereas corrective feedback is a more general way of providing some clues, or eliciting some correction, besides the direct correction made by the teacher. For the sake of clarity, we will refer to correction as corrective feedback in this paper.

Although the role of corrective feedback has been discussed from both theoretical and methodological viewpoints, one wonders: What occurs in practice in real foreign language classrooms? How are these theories and methodologies translated and implemented with real language learners? These questions have been around for some decades now, and problems with regard to the use of corrective feedback or its absence in the language classroom have been identified, to wit: a) the inconsistency, ambiguity, and ineffectiveness of teachers' corrections (Allwright, 1975; Chaudron, 1977; Long, 1977); b) ambiguous, random and unsystematic feedback on errors by teachers (Lyster & Mori, 2006); c) acceptance of errors for fear of interrupting communication; and d) a wide range of learner error types addressed as corrective feedback (Lyster & Ranta, 1997).

Corrective feedback (CF) has been discussed mainly in second language acquisition contexts, but less has been done in foreign language settings. Therefore, this paper reports the findings from a study conducted at a Mexican university where English as a Foreign Language is taught to all undergraduate students who have, as a graduation requirement, a need to cover four prescribed levels of English (from Introduction to Intermediate). Our aims were to identify the EFL (English as a Foreign Language) instructors' perceptions about CF and its actual practice in their classrooms. Our specific questions were: What are the teachers' perceptions about corrective feedback? What are the teachers' self-reported ways of implementing corrective feedback in their classrooms?

This paper is organized into three sections. First, an overview of CF in literature is presented. We discuss mainly the changing viewpoints with regard to CF, and then we describe strategies employed to provide oral corrective feedback, considering the provider, the frequency of provision, the type of error, and the type of strategy. The next section includes a description of the method used to conduct this descriptive study. The research findings as well as a discussion and interpretation make up section 3. Data from both the questionnaire and the interview are integrated in the discussion. Finally, a conclusion and some suggestions are offered for EFL teaching.

The term corrective feedback has been defined at different times in a very similar way. One of the earliest definitions is that of Chaudron (1977), who considers it as "any reaction of the teacher which clearly transforms, disapprovingly refers to, or demands improvement of the learner utterance" (p. 31). More recently, Ellis, Loewen and Erlam (2006) stated that:

Corrective feedback takes the form of responses to learner utterances that contain error. The responses can consist of (a) an indication that an error has been committed, (b) provision of the correct target language form, or (c) meta-linguistic information about the nature of the error, or any combination of these. (p. 340)

Although all these definitions include the learners' and teacher's participation, and thus, a classroom as the setting where CF takes place, this can also occur in naturalistic settings where native or non-native speakers can provide it.

Interestingly, in the foreign language contexts, Sheen (2011) points out that not all CF occurs because of a communication breakdown; teachers can use it to draw the learners' attention to form even in those situations where they comprehend each other. This means that CF can carry negotiation of meaning and negotiation of form as well.

The role and importance of CF in EFL pedagogy can vary from teacher to teacher. This may depend on their previous education and training, teaching experience, and their own experience as language learners, amongst others. CF is a very controversial issue in this regard. Perspectives toward errors have gone from the extreme of non acceptance and preventing them at all cost, to more permissive perspectives in which errors are seen as part of the language development.

Next, we present a summary of the main issues concerned with the provision of oral CF.

Types of Oral Corrective Feedback Strategies

Sheen (2011) classifies CF strategies into seven types; Yao (2000) added body language as another strategy. Table 1 illustrates this and a more detailed study follows.

Recasts

"A recast is a reformulation of the learner's erroneous utterance that corrects all or part of the learner's utterance and is embedded in the continuing discourse" (Sheen, 2011, p. 2). Recasts can be partial or whole (only a part or the whole utterance is reformulated, respectively). They can be didactic or conversational. The former is a partial or whole reformulation that draws the learner's attention to the error made. The purpose is merely pedagogical. On the other hand, the conversation recasts take place when there is a breakdown in communication, and the corrector reformulates to verify if he comprehends what is intended. The following dialogs illustrate this strategy:

S: I have 20 years old.¹

T: I am

(Partial didactic recast)

S: I can lend your pen?

T: What?

S: Can I lend your pen?

T: You mean, Can I borrow your pen?

(Conversation recast)

Explicit Correction

The correct form is provided by the instructor. Sheen (2011) indicates that phrases such as "It's not X but Y", "You should say X", "We say X not Y" usually accompany this treatment. Example:

S: She go to school every day.

T: It's not "she go", but "she goes".

(Sample of our own)

Explicit Correction with Meta-Linguistic Explanation

The correct form and a meta-linguistic comment on the form are provided. Let us see the following example:

S: Yesterday rained.

T: Yesterday it rained. You need to include the pronoun "it" before the verb. In English we need "it" before this type of verb related to weather.

(Sample taken from Sheen, 2011)

Repetition

In order to elicit the correct form, the wrong utterance is repeated (partially or entirely). We suggest that this repetition is generally accompanied by some intonation change emphasizing the error or in a question form. Example:

S: I eated a sandwich.

T: I EATED a sandwich?

(Sample of our own)

This strategy takes place when there is a repetition of the learners' erroneous utterance up to the point when the error occurs. This way self- correction is promoted. Example:

S: When did you went to the market?

T: When did you...?

(Sample of our own)

Meta-Linguistic Cue

This strategy is similar to "explicit correction with meta-linguistic explanation" to some extent, but it differs in that there is a meta-linguistic comment by the corrector, but the correct form is not provided. Self-correction is then encouraged. Example:

S: There were many woman in the meeting.

T. You need plural.

(Sample of our own)

Body Language

The corrector uses either a facial expression or a body movement to indicate that what the student said is incorrect. A frown, head shaking, or finger signaling "no" can be observed (Yao, 2000). Example:

S: She doesn't can swim.

T: Mmm. (T. Shakes her head= no).

(Sample of our own)

These strategies can be classified into those which provide some input (correct form is provided) or the learner is prompted to generate some output by himself (correct form is elicited).

Clarification Requests

When the learner's utterance has an error and a clarification is requested: "Sorry?", "Pardon me?" I don't understand what you just said. Example:

S: How many years do you have?

T: Sorry?

(Sample of our own)

Another useful categorization of strategies is that which divides them into explicit CF and implicit CF. With explicit, there is an overt linguistic signal in the correction; with implicit the correction is prompted or elicited without an overt linguistic signal. The preference for one type or the other may depend on the teacher.

A very important factor to consider when choosing the CF strategy is its effect on learner uptake, which is defined by Lyster and Ranta (1997) as "a student utterance that immediately follows the teacher's feedback with the intention of drawing attention to some aspect of the student's initial utterance" (p. 49). In other words, it's the learner's response to the CF received. He has to choose: to repair or not to repair. Lyster and Ranta call these actions: repair and needs repair. In the former, the learner corrects after receiving CF; in the latter, the learner may acknowledge the correction (but without any correction) or just continue talking.

Focused and unfocused CF is another way of providing correction in the classroom setting. The former refers to the "intensive corrective feedback that repeatedly targets one or a very limited number of linguistic features"; unfocused CF is "extensive corrective feedback that targets a range of grammatical structures" (Sheen, 2011, p. 8). We can also understand unfocused CF as that feedback that targets any feature of a language level: pronunciation, grammar, semantics, pragmatics; and many structures, phonemes, and categories at the same time.

Another issue regarding how to provide CF has to do with the dichotomy individual vs. group correction. Some instructors consider individual correction as an activity that may prevent further participation in the classroom because they see CF as an inhibitor, as something that may damage learners' feelings; therefore, they favor group correction. A differing view is that individual correction seems more effective, as the learners addressed becomes aware of their errors, notices the error, and corrects. When using group correction, many students do not even acknowledge the errors they made and there is no repair at all.

Although the literature on corrective feedback generally does not discuss the possibility that the strategies to provide CF can vary, depending on the learners' language proficiency and meta-linguistic vocabulary, in practice this is something that can occur. For instance, it may be difficult to provide explicit correction with meta-linguistic explanation to beginners in the target language, and probably more time would be wasted than that required for another strategy such as body language. This is another important decision for the language instructor, who needs not only a range of strategies, (examples provided previously), but also the experience of how to put them into practice with real language learners and their particular individual differences.

In the theoretical and pedagogical grounds CF has been a very controversial topic. Loewen, Li, Fei, Thomson, Nakatsukasa, Ahn, and Chen (2009) claim that this controversy can be better understood in terms of meaning-focused instruction versus form-focused instruction. The former assumes that second language (L2) acquisition occurs unconsciously and implicitly like first language acquisition (L1) does. Advocates of this view claim that overt attention to linguistic form is not needed, and they see corrective feedback as ineffective (e.g., Krashen, 1982; Newmark & Reibel, 1968; Schwartz, 1993; Terrell, 1977; Truscott, 1999, all cited by Loewen et al. 2009). Krashen (1982), one of its proponents, suggests that CF is useless and potentially harmful.

The meaning-focused instruction has been questioned with regard to its effectiveness. Research suggests that learners' production shows grammatical inaccuracy even after years of exposure to the target language. This situation has been associated with a lack of noticing and practicing linguistic forms on behalf of the learners. Findings suggest, therefore, that form-focused instruction can benefit language learners. Form-focused instruction is defined by Ellis (2001, p. 1) as "any planned or incidental instructional activity that is intended to induce language learners to pay attention to linguistic forms." This last instruction supports the use of CF in language learning.

When to Use CF?

CF can be provided immediately after the error has been made, or it can be delayed until later, after the communicative activity the learners are engaged in is finished. The main distinction many instructors make is between fluency and accuracy or if the activity involves negotiation of meaning or negotiation of form. Instructors who practice a focus on meaning instruction and encourage fluency in their classrooms prefer to delay CF. However, if their instruction follows a focus on form and they want to encourage accuracy, then both immediate and delayed CF are encouraged.

Important also to consider in instructional settings is the frequency with which teachers use CF in their classes. Too much correction can sometimes have a negative effect on the learners' attitudes or performances; whereas too little feedback can also be perceived by learners as a hindrance for efficient and effective language learning. Finding the right balance as regards the amount of CF is, therefore, not an easy task.

Error Types

When correcting, it is paramount to identify the type of error the learners make because it is not always the case teachers want or need to correct everything. Errors have been categorized by Mackey, Gass and McDonough (2000) and Nishita (2004), cited by Yoshida (2008) as:

• Morphosyntactic error: Learners incorrectly use word order, tense, conjugation and particles.

• Phonological error: Learners mispronounce words (or we suggest it could also include suprasegmental errors such as stress and intonation).

• Lexical error: Learners use vocabulary inappropriately or they code-switch to their first language because of their lack of lexical knowledge.

• Semantic and pragmatic error: The misunderstanding of a learner's utterance, even if there are no grammatical, lexical or phonological errors.

When dealing with errors, language instructors have to make many decisions and one of them is the type of error to correct. However, sometimes some types of errors are neglected to some extent, or only the most "serious" errors are corrected. That is, there are errors that probably do not hinder comprehension between the language instructor and the learner, but they are errors that in a real world setting might affect communication with other speakers who are not familiar with foreign accents, or who are not tolerant with nonnative speakers. Thus, identifying and targeting the types of errors that are relevant and essential to become a successful EFL learner is another complex task for the instructor.

CF Providers in the Classroom Setting

Considering the participant(s) in the corrective feedback interaction, the following possibilities can be observed:

Self-correction is possible when the learner realizes that s/he has committed an error and repairs it by providing a correct form. Self-correction seems to be preferred to correction provided by others because it is face-saving and allows the learner to play an active role in the corrective event. Self-correction plays a central role in the promotion of autonomous learning nowadays.

Peer correction occurs when one learner corrects another one. Its most important advantages are that both learners are involved in face-to-face interaction; the teacher obtains information about learners' current abilities; learners co-operate in language learning and become less teacher-dependent; peer correction does not make errors a public affair, which protects the learners' egos and increases their self-confidence.

Teacher correction occurs, of course, when the person to correct the errors is the teacher. He or she knows the problem and the solution, and can define and put things simply so that the learner can understand the error.

As shown in the previous pages, CF is a very complex phenomenon in EFL which has its own peculiarities that distinguish it from ESL contexts. It is not only that the classroom is the setting where learners mainly receive language input, but also where they receive their provision of CF. With this limitation, thinking about CF and its role in language learning in this particular context becomes a relevant issue. Practicing CF in EFL settings is therefore a complex task in which many factors meet and intertwine. Teachers have to ask themselves: Why to include CF? How to provide CF? What to correct? How much CF and how frequently? Who is to correct? And then make decisions. Additionally, teachers have also to be concerned with the individual differences. This is something that will be discussed in the findings in the next section.

This is a descriptive study conducted at a Mexican university located in the southeast region of the country. This university offers English as Foreign Language (EFL) courses from beginners to advanced level to all undergraduate students. However, they only have to cover, as a graduation requirement, four levels of English (from Beginner to Intermediate). The population of students taking EFL courses totals 600 every term, approximately, and there are about 40 instructors teaching these courses.

For this study, a semi-structured interview and a questionnaire were used to collect the data. Five language instructors, with ages from 25 to 60, were interviewed. Their teaching experience ranged from 4 to 20 years. The interviews were recorded and analyzed considering variables such as types of errors, the CF provider, frequency of correction, CF techniques, perceptions of students' attitudes, training, and perceptions about CF.

A questionnaire was designed and distributed among 40 instructors. Unfortunately, only 15 gave us back the questionnaire. The instructors were teaching courses from introductory to interme-diate levels at that time. The questionnaire consisted of five sections intended to obtain data about those instructors' ideas on CF and its practice in the classroom; perceptions about their learners' reactions and attitudes toward CF; attention paid to the different language levels (i.e. phonetics phonology, syntax, morphology, semantics, and pragmatics); the different strategies used to provide oral CF and its frequency of use; and instructors' perceptions on the strategies most preferred by their students.

To analyze the data, descriptive statistics were used with SPSS v. 18 (Statistical Package for the Social Sciences), and a qualitative analysis of the variables was conducted and interpreted by both researchers.

The Role and Importance of Corrective Feedback in the Classroom

From the questionnaires there is a strong tendency (80%) to agree on the need to correct learners so that they gain fluency and accuracy. This is concurrent with the idea that CF has a positive impact on language learning in which 87.7% of the instructors agreed. However, 3 out of 15 teachers believe that CF does not play a relevant role in the acquisition of fluency and accuracy. In the interviews, 4 out of 5 instructors agreed with the need to provide oral CF in the classroom, but it seems they do not believe in the benefits of CF, or the impact it can have on the learners. They consider CF to be only necessary to develop accuracy.

Overall, these instructors have positive beliefs and attitudes toward oral CF, as they consider it necessary for language learning. Nonetheless, in the interviews most of the teachers associate CF to focus on form (limited to accuracy). It seems that they favor the focus on meaning instruction (and fluency), and therefore, they cannot accept CF completely. This lack of total acceptance may have to do with their academic profile and teaching experience; their previous knowledge and education.

Effects on Learners: Reactions, Emotions and Feelings

13.3% consider that CF inhibits students' participation, 60% partially agreed with this statement and 20% disagreed. These varied answers probably have to do with the consideration of other variables that can influence this outcome. If teachers consider, for example, that individual correction affects more than group correction, and it does prevent further participation, then that is why they either agreed or partially agreed. They may also have thought that the amount of CF could be another variable affecting participation. These results coincided with those of the interviews, in which instructors mentioned that it was important to get to know their students very well in order to know if CF could be used or not with some students. They said that learners had different attitudes toward CF and teachers should be aware of this and decide whether or not to consider it for the provision of CF. 3 out of 5 teachers said that at the beginning of their courses they asked their students if they wanted to be corrected. This leads one to think that teachers perceive CF as an activity with many intricate variables to control, and if this is not done tactfully, then it may be detrimental to class participation.

33.4% consider frequent CF as a cause of frustration or demotivation; 46.7% partially agreed, and 20% disagreed. Again, it seems that teachers think of CF as a potential cause of these emotions or feelings if some other issues (such as personal traits) are not controlled or considered for the provision of CF. This was also manifested in the interviews where instructors expressed that CF could damage the learners' feelings and the process of learning if used very frequently and regardless of the personality or emotions of the students.

A contradiction, however, was identified in the results. When asked if shyness or low motivation should be factors to consider in the provision of CF, 66.7% did not think so, and the rest partially agreed with this statement. This complements the question of whether correction should be used only with more open/receptive learners. 60% of the instructors disagreed on correcting only this type of student; 13.3% partially agreed and 26.7% agreed. There is inconsistency by some in conceiving error correction as an inhibitor of participation or a factor of de-motivation. Yet they think that shyness and low motivation should not prevent CF from taking place in the classroom. In the interviews, teachers expressed being very concerned with learners' personality traits, preferences, and attitudes. It appears then that there is not complete agreement in this respect.

Teachers in general (80%) perceive that learners do not get angry or feel bothered when provided with oral CF. In the interviews they agreed that anger and annoyance are not emotions manifested by students in their classroom, but that anxiety, shyness, and introversion were thought of as factors to consider for CF provision.

A paramount reaction to CF is the learners' uptake and repair, that is, what do they do with this CF? Do they correct? 60% agreed that learners do repair their utterances frequently and always; 33.3% said learners do it sometimes, and 6.7% agreed that learners never correct their errors. In the interviews, 3 out of 5 instructors agreed that their students repair; 2 said that learners do it sometimes. Therefore, in general, there is this perception that repair takes place in their classroom with regular frequency. This belief can actually be one of the factors causing some negative or cautious attitude toward CF. Why should teachers bother to provide CF if students do not respond to it? If CF is not helping, why include it in the teaching practice?

The Role of Learners' Proficiency in CF

60% of instructors in this institution consider that the use of certain CF strategies depends on the learners' proficiency. Similar results were obtained through the interviews (3 out of 5 teachers). They said that body language, for example, could be more exploited with beginners, whereas meta-linguistic explanations should be used more with advanced students who have better proficiency in the L2. They also suggested that peer correction is more suitable for advanced students, but not for beginners because they are not used to this type of strategy and they do not trust their classmates for this endeavor. 20%, however, do not think that proficiency in the target language influences the choice of one or another strategy, and 20% partially agreed.

These perceptions could be based on other factors besides proficiency level. Typically students at this institution begin to learn English during their first year of college, and as they progress in the English courses (and therefore gain proficiency), they also get to know their classmates better since they are often placed in the same classroom and have the same schedule. So, while beginners have sometimes complete strangers as classmates, in other higher level English courses, students are acquainted with many of their classmates. This ensures a more trusting environment. In some classrooms the instructor is the only newcomer.

It is also interesting to examine these results if we consider that some students are not familiar with all CF strategies or with meta-language. Two teachers mentioned in the interview that beginners have no idea of what a verb is, an adjective or a pronoun. Consequently, using meta-linguistic feedback is unthinkable at the beginning levels. However, advanced learners do know these terms and are then able to understand the explanations, and repair. The same occurs with some students who have not experienced the different CF strategies, which usually occurs in the first English courses.

When to Correct

Regarding the distinction between immediate or delayed CF, 40% agreed that teachers should provide CF just immediately after the learner has made an error, but without interruption; and 53% partially agreed with this statement. This partial agreement is probably rooted in the purpose teachers have with CF e.g. if the teachers' focus is on accuracy, then they will probably engage in CF immediately; or if it is fluency, they can delay correction.

Most professors (60%) prefer to provide the whole class with CF at the end of the class time. 33% partially agreed with this practice. This trend can be understood as teachers' concern with learners' feelings and emotions and their fear of interrupting and inhibiting participation. The above interpretation (accuracy and fluency distinction) can apply to this result as well. This preference on behalf of the instructors, however, differs with regard to "what should be done". Concerning the statement, "Not only general errors made by the whole class should be corrected, but also individual errors", 73% agreed with it. In the interviews, most of the teachers showed a preference for CF to be provided for the whole class.

Types of Errors to Correct

With regard to correcting only errors that interfere with meaning and with getting the message across, 46% partially agreed and 46% disagreed. It seems then that there is a tendency not to favor this practice in the classroom, or that this may depend on the activities involved and on the focus (meaning vs. form).

There is also a clear tendency on behalf of the instructors to direct CF toward morphosyntactic errors (86.7%), followed by pronunciation (73.3%), lexicon (66.7%) and pragmatics (53.3%). Interestingly, as percentages get lower, more diverse answers to the amount of attention are observed; that is lexicon and pragmatics, for example, had answers such as "some" and "little", respectively, whereas morphosyntax and pronunciation got "a lot" by most teachers. These findings are similar to the answers provided by the teachers in the interview, in which most of them (4 out of 5) emphasized they indeed corrected pronunciation and morphosyntactic errors. None mentioned pragmatic errors. Unfocused correction was manifested through the different examples used to show the type of corrective feedback in their classrooms.

These findings (pronunciation and morphosyntactic errors as main targets) suggest that these instructors pay more attention to language structures rather than meanings when providing CF. They see CF as a way to prevent or correct structure errors. On the contrary, they care less about semantic or pragmatic meanings. This does not correspond to a focus on meaning instruction.

Who Corrects

In the questionnaire, 86.7% consider that the teacher is not the only one who can and must correct errors. This coincides with 73.3% in agreement with the statement that the learners should engage in self-correction with the instructor's help. Although there seems to be a positive attitude toward self-correction, their perceptions about the effectiveness of CF considering the corrector are not consistent among all the teachers. 40% agreed that self-correction is more effective than teachers' CF, and 33% partially agreed. Thus, other variables are apparently seen as intervening in this effectiveness.

Peer correction, on the other hand, is not perceived as a positive activity in the classroom by most teachers (86.7%). The rest partially agreed with having peer correction in the classroom, but none agreed on this strategy as something positive. When asked about the effectiveness of teachers' correction and peer correction, 53.4% do not consider the former to be more effective than the latter; the rest agreed and partially agreed with this statement.

All interviewees agreed that the teacher is the authority for providing CF in the classroom. The instructors do not think that peers are good at correcting their classmates; actually, they said that sometimes peer correction can be harmful for the relationships among students. Generally speaking, teachers seem to favor more teachers' CF, followed by self-CF and then by peer CF. They perceive the former to be the most effective as well. This is probably a result of the traditional and paternalistic education we have had in Mexico for many years. Learners' autonomy has been included in the schools' curricula very recently, and teachers and students are still trying to integrate this into the classrooms, but it has not been easy.

CF Strategies and Their Frequency of Use

The favored strategy was to ask for clarification or confirmation, which was reported to be used always and most frequently by 86.6% of the teachers, although the remaining 13.4% report periodic use. Gestures and mimicry, as well as recasting, were favored next by 80% of the teachers, who reported using them always, and the remaining 20% periodically. 67.7% of teachers emphasize the error so that the learner makes the correction. They use this strategy always and frequently; 20% rarely use this strategy. This emphasis is made mainly with a change in intonation when uttering the erroneous part, or putting the error into a question. (For example, a student says: Where did you went last weekend? Teacher replies: Went?). Most teachers in the interview pointed out that this was one of the main strategies they used and they thought it worked very well with students.

Regarding peer correction, 60% reported using it rarely and 20% only sometimes. This supports what was found regarding who carries out the CF in the classroom as mentioned above. Three interviewees argued that they do not use peer correction very frequently because they thought that this type of feedback could create negative attitudes among the students toward their classmates; many times, they claimed, the "corrector" is seen as superior or a more knowledgeable person than the rest, and this can create a hostile environment which prevents proper camaraderie among students.

Finally, concerning grammatical explanation as a CF strategy, results from the questionnaires show that 46.6% always and frequently provide grammatical explanations; 33% do it sometimes, and 6.7% rarely. In the interviews some teachers (3 out of 5) also manifested some aversion to this strategy, mainly the youngest instructors who insisted that other strategies could be used instead with more positive outcomes. These answers suggest that teachers do not seem to favor explicit correction.

In sum, these trends in strategies used show a higher frequency and preference toward indirect and implicit CF (clarification requests, confirmation checks, gestures), followed by direct and explicit strategies (emphasizing the error and grammatical explanations). Peer correction seems not to be promoted by teachers, but self-correction instead. In the interviews, similar answers were reported, although self-correction was the least promoted by the professors, who highlighted the lack of language awareness on behalf of students correcting their errors.

Learners' Preferences According to Teachers

70% think that their students prefer a teacher's CF rather than a peer's. This concurs with their perception (84.6%) that students would prefer that their instructors do not ask a classmate to help with their corrections. All interviewees agreed that students prefer teachers' feedback rather than their classmates'. They added this was rooted in the following perceptions: That the teacher is the authority in the class and an expert and their classmates do not seem to be very reliable. As such, peers do not rely on their classmates' CF. Also, peer correction could cause a negative impact on the students' relationships because, for example, a student could be corrected by someone he does not like and this could cause some kind of unconstructive attitudes or undesirable reactions.

As to the time when CF is provided, 76.9% consider that students prefer for the instructor to provide CF immediately just after the error has been made. The same percentage believes that students favor group correction rather than individual CF. This seems contradictory because if learners wanted to be corrected immediately after the error, this would imply individual correction. However, in the interviews teachers mentioned that learners prefer both (depending on the type of error, or in order to have some variety in strategies). 53.8% agreed with their perception that students like personal and individual CF.

Regarding the students' favorite oral CF strategies, according to the teachers' perceptions, 61.5% suggested recasting as number one, followed by grammatical explanations, provision of further examples (60%), gestures, and finally clarification requests (53.8%). Although implicit strategies seem to be the most preferred, grammatical explanations came in second place. This finding, interestingly, does not match the teachers' practice of this strategy since they reported lower use of this one in particular and a clear tendency to favor indirect strategies. A possible interpretation of this is that teachers probably reported what they believed or thought as regards how oral CF should be provided, but not how they actually provide it.

In general, teachers at this institution have a positive perception of oral corrective feedback. However, they need to know more about its effects and role in interlanguage development because they look at CF only as a technique to improve accuracy in the language, particularly in pronunciation and morphosyntax. Some teachers actually consider CF as optional (mainly individual CF) because they are very concerned with students' feelings and emotions. In this regard, these instructors in particular have such a respect for the individual differences such as personality, attitudes, motivation, and beliefs, that this affects -sometimes positively and other times negatively- their practice with regard to oral CF.

Unfocused oral CF is predominant in the instructors' practices and this situation may need to be reconsidered as it probably inhibits the learners' noticing their errors and subsequent pursuit of repair. With many aspects covered at the same time, students might not engage in as much correction as desired. Teachers should make it clear to their students what they need to correct and pay more attention to it so that repair does indeed occur.

With regard to the use of strategies, the implicit ones are more favored by this group of teachers. Teachers should know the effectiveness of both explicit and implicit strategies and choose the ones proven to be more effective. As a matter of variability, many possible strategies should be exploited in the classroom.

For the instructors, the most suitable person to provide CF is the teacher, followed by the learner doing self-correction; peer correction is the least favored. However, fostering autonomous learning is a paramount task in the teachers' agenda as is collaborative learning. Teachers should be aware of the advantages that self and peer correction have, as they can raise or increase language awareness and help learners to test hypotheses in the target language.

In brief, this research in the Mexican context provides, in general, evidence of similar problems found in previous studies (Allwright, 1975; Chaudron, 1977; Long, 1977; Lyster & Ranta, 1997; Lyster & Mori, 2006): inconsistency; ambiguity of teachers' corrections; random and unsystematic feedback on errors by teachers; acceptance of errors for fear of interrupting the communication; and a wide range of learner error types addressed as corrective feedback. The first step then is, as language teachers, to learn more about CF and to share it with the learners; to manage individual differences in a way that they do not interfere with the language learning; to put into practice new and more effective strategies; to organize and systematize corrective feedback; and to set clear and feasible goals in this respect.

¹ S= student; T= teacher. Samples are of our own.

Allwright, R. L. (1975). Problems in the study of the language teacher's treatment of error. In M. K. Burt & H. D. Dulay (Eds.), New directions in second language learning, teaching, and bilingual education. Selected papers from the Ninth Annual TESOL Convention. Washington, D.C: TESOL.         [ Links ]

Chaudron, C. (1977). A descriptive model of discourse in the corrective treatment of learners' errors. Language Learning, 27, 29-46.         [ Links ]

Ellis, R. (2001). Investigating form-focused instruction. Language Learning, 51(1), 1-46.         [ Links ]

Ellis, R., Loewen, S., & Erlam, R. (2006). Implicit and explicit corrective feedback and the acquisition of L2 grammar. Studies of Second Language Acquisition, 28, 339-368.         [ Links ]

Han, Z. H. (2008). Error correction: Towards a differential approach. Paper presented at The Fourth QCC Colloquium on Second Language Acquisition. New York, NY. Retrieved from http://www.tc.columbia.edu/academics/?facid=zhh2        [ Links ]

Krashen, S. D. (1982). Principles and practice in second language acquisition. New York, NY: Pergamon Institute of English.         [ Links ]

Long, M. (1977). Teacher feedback on learner error: Mapping cognitions. In H. Brown, C. Yorio & R. Crymes (Eds.), TESOL`77. Teaching and learning English as a second language: Trends in research and practice (pp. 278-294). Washington D.C.: TESOL.         [ Links ]

Loewen, S., Li, S., Fei, F., Thomson, A., Nakatsukasa, K., Ahn, S., & Chen, X. (2009). L2 learners' beliefs about grammar instruction and error correction. The Modern Language Journal, 93(1), 91-104.         [ Links ]

Lyster, R., & Mori, H. (2006). Interactional feedback and instructional counterbalance. Studies in Second Language Acquisition, 28, 269-300.         [ Links ]

Lyster, R., & Ranta, L. (1997). Corrective feedback and learner uptake. Studies in Second Language Acquisition, 19, 37-66.         [ Links ]

Mackey, A., Gass S., & McDonough, K. (2000). How do learners perceive interactional feedback? Studies in Second Language Acquisition, 2, 471-497.         [ Links ]

Sheen, Y. (2011). Corrective feedback, individual differences and second language learning. Dordrecht: Springer.         [ Links ]

Yao, S. (2000). Focus on form in the foreign language classroom: EFL college learners' attitudes toward error correction. Buffalo: State University of New York at Buffalo.         [ Links ]

Yoshida, R. (2008). Teachers' choice and learners' preference of corrective feedback types. Language Awareness, 17(1), 78-93.         [ Links ]

Edith Hernández Méndez holds a PhD in Hispanic Linguistics from Ohio State University. Full time professor at Universidad de Quintana Roo and member of the National System of Researchers in Mexico. Research interests: language acquisition, language learning and teaching, and sociolinguistics.

María del Rosario Reyes Cruz holds a PhD in International Education from Universidad Autónoma de Tamaulipas. Full time professor at Universidad de Quintana Roo and member of the National System of Researchers in Mexico. Research interests: pedagogical beliefs, epistemological beliefs, beliefs about language learning, and educational technology.