As Errors Go By…

14 08 2012

Recent discussions on error correction on the iTDi blog and the Braz-TESOL Conference last month have sparked my desire to contribute to this debate by revisiting a pilot research my colleague Lúcia Santos and I conducted in our institution in 2004. I hope it will shed some light into the discussion.

The study culminated in the talk As Time Goes By: Corrective Feedback in the EFL Classroom , given in the 2004 Braz-TESOL Conference in Belo Horizonte, then in the 2005 TESOL Convention in San Antonio, and also about three or four times in our institution, where it has had a great impact on how corrective feedback has been done since then.

Our hypothesis at the time was that our advanced students’ oral production was not in keeping with their level because teachers might not be providing sufficient corrective feedback. Thus, we set out to investigate:

a) How corrective feedback was used in the classroom;

b) How students were responding to it;

c) What types of corrective feedback teachers were using;

d) If students were perceiving that they were being corrected;

e) If the teachers knew their students’ preferred types of corrective


To this end, we observed 11 groups of advanced students, using a grid where we marked the mistakes made by students and the type of corrective feedback (Lyster and Ranta, 1997) used by the teacher, if any, during a fifty-minute observation, in whole-class interactions (it would have been impossible for us to monitor all the group and pair work).

We also distributed questionnaires to teachers and students in 30 groups, to obtain a larger sample.  We chose the advanced level because, while in other levels one can claim that the focus on developing fluency should be greater, it is in the advanced level that students have the last chance to brush up their English in order to use it for higher-stakes academic and professional purposes. After all, a learner who only wants to communicate fluently and get their message across might not even be interested in advanced studies of the language.

If you don’t have time for reading or don’t care about the details of the study itself and the methodology used, you can skip this next part and go straight to the conclusions at the end, in bold. Otherwise, here it is:

Not surprisingly, given our hypothesis, out of the 62 errors we identified in our observations, 41 were not corrected at all. For the remaining 21, 11 were corrected by way of explicit correction, six by elicitation, and four by recasts. We were happy and sad at the same time. On the one hand, most errors had just gone by; on the other, when errors were treated, the most effective types of corrective feedback strategies were used (Lyster and Ranta, 1997).

As mentioned, besides the observations, we also administered questionnaires to students and their teachers. Our first question to students was whether they would like to have their spoken errors corrected by the teacher. An astounding 95.8% answered:

Then we asked students how frequently they perceived that errors were being corrected in their classroom: 49.6% said somewhat frequently; 30% said frequently; 7.8% said very frequently. Only 10.7% responded that it was rarely and 0.02% said there was no correction at all. Conversely, 16% of the teachers reported to correct errors very frequently; 50% reported doing it frequently; and 33% stated they did it somewhat frequently. When asked whether they wanted more or less corrective feedback, 68% of the students responded that they wanted more. They were asking for more! They wanted guidance, direction.

We also compared each group’s response to the teacher’s in order to determine if there was a convergence or divergence in this perception of the frequency of feedback provided, concluding that there was a convergence in 11 of the cases and a divergence in 13 (some teachers had two or three advanced groups, so we considered them as one). This made us wonder why some teachers thought they were providing a good amount of corrective feedback to their students, while these same students didn’t really think so or perhaps perceive so. But then we realized that perceptions of frequency can vary depending on expectations. For example, if a student wants all errors to be corrected and they are indeed treated 80% of the time, they might still think that it’s not frequent enough, while the teacher might find it so. After all, if there’s something that seems to be unanimous among us is that we shouldn’t correct all errors made by all students all the time!

When asked their preferred form of feedback, most students selected metalinguistic feedback, followed by repetition. For each type, we gave a clear example rather than the technical terminology. Teachers, on the other hand, demonstrated a more balanced approach, even though elicitation and recasts ranked higher on their list. Metalinguistic feedback, the students’ favorite, was marked by 13 of the 24 teachers who answered the questionnaire.

We also asked teachers to summarize their approach to corrective feedback and, triangulating the data, we noticed a certain tendency:  students  whose teachers had a more fluency-focused approach stated they wanted more corrective feedback, while in the case of teachers who reported using a more balanced approach, there were more students demonstrating satisfaction with the amount of feedback provided.

Delving into the subject, we asked students what types of feedback they thought their teachers used the most. The vast majority of students (143) said the most common type of corrective feedback provided by their teachers was recasts, followed by repetition (84), metalinguistic feedback (80), elicitation (52), explicit correction (38), and clarification request (17). Teachers, on the other hand, reported using a balanced approach and resorting to all types of feedback, elicitation being the most marked option (19), followed by recasts (17), explicit correction (15), clarification request (14), metalinguistic feedback (13) and repetition (11). The relative discrepancy in the teachers’ stated feedback preference and the students’ perceived feedback practices puzzled us and led to the following questions and reflections:

–          Why did so many students report that recasts were the most common type when neither did the teachers report such a great preference for them, nor did our observations corroborate such response? Could it be that every time the teacher repeated what they or a classmate said, the student thought it was some sort of correction when, in some cases, the teacher could just be reacting to the message rather than the form?

–          Could students be confusing repetition as a drilling technique rather than a corrective feedback one? Teachers clearly showed little preference for it, while it came in second on students’ list.

–          The type of feedback selected by most teachers was elicitation; on the other hand, only 52 students marked elicitation as their teacher’s most used strategy. Could it be that they don’t always perceive when elicitation is used for corrective purposes as opposed to prompting purposes?

Anyhow, with all the limitations and discrepancies in our pilot study, here are some conclusions that have helped us reflect on the topic and reshape some of our teaching since then:

  • Teachers were not providing enough corrective feedback to  students, a least from the students’ perspective.
  • Teachers tend to use more indirect ways of correction.
  • Students tend to prefer more student-generated correction.
  • Many times students don’t notice if the teacher is correcting them or how the teacher is correcting them.
  • Advanced students want and need corrective feedback. Many teachers have reported correcting their students’ pronunciation of “ed” in regular verbs or certain word choices such as “depend of” (rather than “on”) and students reacting by saying, “Why didn’t anyone tell me this before!”
  •  In a strategy training and learner autonomy perspective, it might be beneficial to discuss this topic with your students and even use a questionnaire asking whether they want their errors to be corrected and how. While it’s virtually impossible for the teacher to remember each student’s preference, it will give them an idea of the group’s general preferences
  • Teachers should steer clear from recasts. They are definitely confusing to students as corrective feedback, as Lyster and Ranta (1997) have pointed out.


Again, we are talking about ADVANCED-LEVEL STUDENTS. We are not talking about students who are initiating their studies and need to feel comfortable with the language first, or who still need to lower their affective filters. We are not talking about how fluency should come before accuracy in initial stages of language learning. We are talking about students’ LAST CHANCE. We are talking about EMPOWERING them to surpass the uncountable gatekeepers they will encounter – proficiency tests, job interviews, college-application  essays, to name a few. One might say,  “Who cares if they say ‘depend of’ instead of ‘depend on’? They’re getting their message across anyway!” Well, THEY might care. Someone else out there who’s going to make decisions about their English might care.  My suggestion: AT LEAST ASK THE STUDENTS!


Leaving language instruction at an intuitive and ‘mystical’ level of ‘natural language acquisition’ may be easy for the teacher and may make some students feel good, but it leads to disempowerment. (Grabe, 2002:279)


Free images courtesy of


Grabe, W. (2002). Dilemmas for the Development of Second Language Reading Abilities. In J. C. Richards and W. A. Renandya (Eds). Methodology in Language Teaching – An Anthology of Current Practice.  New York, NY: Cambridge University Press.

Lyster, R. and Ranta, L. (1997). Corrective feedback and learner uptake: Negotiation or form in communicative classrooms. Studies in Second Language Acquisition, 10/1: 37-61.

%d bloggers like this: