Please note: this post was co-authored with my dear colleague Dylan Vinales

I have written extensively about Error Correction on this blog, often reiterating the point that whilst there is some evidence pointing to its effectiveness in enhancing L2 writing accuracy (e.g. Ashwell, 2000 ; Chandler, 2003), the gains obtained do not justify the enormous amount of time and effort invested by teachers in the process.

Take Chandler (2003)’s findings: she calculated that teachers’ marking time amounts on average to around 1 minute per 100 words, the time being slightly less ( around 48”) if one simply underlines errors. Then consider that

(a)   correction of whole texts can achieve significant results (i.e. 10% reduction in error rate) provided the students write in the region of 5,000 words a semester on first drafts

(b)   many types of error are resistant to eradication (Alroe,2011). In other words, improvements accompany large amounts of writing and consequently large amounts of correction.

and do the maths: is this modest benefit worth the effort?

Whilst the answer is probably ‘No’, teachers often do not have much choice, as surveys of student and parent opinion clearly show (Conti, 2001) , both groups of stakeholders expect errors to be corrected. So, chances are that your boss will demand that you correct.

Why the effects of Error Correction are so limited

Learning a language is not merely about accruing intellectual knowledge of the target language. As Truscott (1996) pointed out, learning to master a specific grammar rule doesn’t occur as a sudden revelation resulting from the information a teacher passes to a student through oral or written feedback.

L2 acquisition is much more complex than that: it is a long and painstaking process which may start with the understanding of how a given language item works, but requires extensive practice in the deployment of that item across a wide range of linguistic contexts before it can be said to have been brought to completion. Think about the mastery of Imperfect usage in L2 French, Spanish or Italian; how many years does your average student require in order to fully acquire it? Three? Four? Five years?

There are many factors which undermine commonly practised error-treatment methodologies (thoroughly discussed here). In previous blogs I identified the following ones as key:

(a) Corrective intervention neglects the intentionality dimension of learning from one’s mistake, i.e. it is not proactive in arousing students’ desire to eradicate errors. Yet, this affective dimension is vital to the success of error correction (James, 1998; Conti, 2001).

(b)   As many studies have pointed out, L2 students do not invest sufficient cognitive effort in the corrective process. They do not process the corrections deeply enough; at best, they make a mental note of the mistakes and the relative corrections and move on (Cohen and Cavalcanti, 1990; Conti, 2001; Alroe, 2011).

(c)    The corrective treatment rarely involves a long-term, sustained effort to eradicate mistakes. Yet, this is crucial to the success of error remediation.

(d)   For corrective feedback on item X to be effective, students need to process it many items over. This doesn’t always happen with less frequently occurring L2 items. Item X may occur two or three times in Essay 1 in September but never re-occur until Essay 5 later on in the year. This is a common shortcoming of error correction; one that is way too often overlooked, especially in contexts where highly structured writing tasks and translation are not assigned for fear of stifling creativity and/or negatively impacting student motivation. Yet, corrective feedback, like any other form of instruction, benefits greatly from frequent spaced (distributed) practice.

(e)   usually, teachers are concerned with fixing errors but not with training students in becoming effective independent editors of their own written output. Yet, several studies (see Macaro, 2003) have shown the benefits of enhancing L2-learner ability to self-monitor through a synergy of awareness-raising (what mistakes do I make more often?), error-targeting (what mistakes am I going to eradicate?) and editing strategies (what strategies work best with this error type?)

(f)    L2 writing (see picture 1 below) poses enormous cognitive challenges to novice L2 student-writers attentional resources . As I have discussed in detail here, errors are often the result of Working Memory processing inefficiency rather than lack of knowledge; i.e. the brain is juggling so many demands arising from writing a given sentence that mistakes are made not because the student doesn’t know the grammar rules governing the use of the items in that sentence but rather because of lack of fluency in the application of those rules. Corrective intervention through Indirect Correction (teacher underlines and student self-correct) does not address this very important issue; underlining errors and asking students to self-correct does not prepare them for the cognitive juggling they have to perform as they write in real time; in other words, it does not enhance their editing fluency.

This post is about how I have attempted to address the above issues in my own teaching practice over the last seven months or so.

Picture 1 – Hayes and Flower’s (1981) Cognitive model of the writing process

Enter Focused Correction

Another common (potential) weakness of corrective feedback, which I did not discuss in previous posts, but which is extremely relevant to the present one, is the fact that Error Correction is rarely, if ever, intentionally and systematically closely tied in with the current objectives of the curriculum. So, if in term 1 your students are learning about the communicative function ‘talking about a past event’ in the context of the topic ‘media and leisure’ many teachers would typically correct an essay assigned on that topic not merely on errors relating to the wrong use of time markers and past tenses, but also on other errors they find in that essay.

In my own research a decade ago, for instance, I found that both at university and secondary level, the approach adopted by most of the teachers I studied, tended to be more comprehensive than selective and when it was indeed selective, it was not consistently based on a specific set of criteria (e.g. frequency, perceived gravity or incomprehensibility) but it appeared to be somewhat ‘random’; moreover, and more importantly, it didn’t focus on the same areas systematically over a longer period of time.

My findings were that less proficient student writers appear to be more likely to benefit from correction when:

(a)   it concentrates on only two or three major target areas (so that their attentional resources can be used more efficiently);

(b)   such areas are perceived as relevant to their current learning (so that they feel more motivated to address them);

(c)    it provides frequent feedback on the same target areas week in week out (so that it enhances their understanding and keeps them constantly focused on the same error types week in week out);

(d)   students have the declarative knowledge for self-correcting the errors with minimal prompting by the teacher (so that students are self-reliant in the process );

(e)   the students’ levels of accuracy in the target areas has a substantive bearing on the grading system adopted (so that students are more motivated to be more accurate).

 

The criteria described in (a) to (e) above refer to an approach described in the AL literature as ‘Focused correction’ (henceforth FC). A number of studies have indicated that FC is significantly more effective than unfocused error correction (Alroe, 2011). Sheen (2007), for instance, compared the effect of FC on the writing accuracy of ESL learners with those of unfocused correction and no correction. She found that FC resulted in a significant reduction of mistakes in the five target areas (i.e., article, copular ‘be’, regular past tense, irregular past tense and preposition). Ellis et al. (2008) studied the effects of FC on indefinite/definite article use in writing. A control group was contrasted with two experimental groups one of which received focused treatment (only article errors were corrected) and one of which received unfocused treatment (other errors were also corrected).  The experimental groups were significantly better and the trend in the experimental groups suggested that focused feedback may be more effective in the long run.

 

How I implement Focused Correction

Two years ago I was  required by my previous head of department to adopt a corrective approach (strongly criticised in this Language Gym classic), which involved underlining randomly selected errors categories, coding them using a ludicrously lengthy and laborious coding system and asking students to self-correct.

Having found this methodology ineffective for the reasons discussed above, this year my current head of Department gave me ‘carte blanche’, and after discussing the issue at length, I was given the go-ahead to experiment with FC with my beginner to lower-intermediate classes (years 7 to 9 in the English system).

This is how I have implemented it:

(1)    I selected three FCAs (Focus Correction Areas) per term for all students in a given class. The FCAs were based on the main curricular objectives of the term; for example, with my year 7 beginner class, in Term 1 I focused on adjectival agreement; articles and number of nouns.

(2)    With more gifted students, one or two more FCAs were added;

(3)    Students were free to use L.I.F.T. (Learner Initiated Feedback Technique) a student-driven procedure whereby the students write in the margin of their written piece questions on any doubts they may have about a language item they have used (e.g. ‘Sir, is the use of the article I have underlined correct?); the teacher then replies with a metalinguistic explanation.

(4)    In order to enhance the students’ ability to edit, identify and correct errors in the FCAs, I have staged short and snappy ‘Spot the error’ tasks and other editing-enhancement activities (e.g. ‘Error Auctions’, explained here).

(5)    As the term proceeded, I have increasingly resorted to indirect feedback with error coding (i.e. self-correction). Although, as noted above, I am usually against indirect feedback with coding, in my context, the coding system was (a) much easier to manage for both teachers and students, as it referred only to three error categories, (b) tied in with the topics under study, (c) used in synergy with editing instruction (see point 6 below).

(6)    To further scaffold the process, I have used editing checklists like the one in picture (2) below; the students were expected to go meticulously through their written pieces before handing them in noting down every occurrence of any items referring to the FCAs and ticking it to signal they checked it. Note that this process was totally managed by the students who handed in the checklist with the assignment.

Picture 2 – Editing checklist

(7)    The tasks have been graded based on Effective Communication (i.e. the extent to which they communicated the requested information) and Accuracy across the FCAs, 70 % of the marks being allocated for the former and 30 % for the latter (10% per FCA).

(8) With some classes I have used error tally sheets in which students logged the number of errors in each category from week to week to enable us to track their error-making trends overtime in each of the FCAs.

It must be noted that in order to ensure that the student output contained as many instances of the target FCAs as possible, the written tasks included alongside unstructured assignments (e.g. ‘Write a 100-words piece about your family’) highly structured tasks designed to elicit the use of specific structures. For instance, to get my year 7 to demonstrate effective production of adjectival agreement, number of nouns and articles:

Write a 100-word piece about your family. Include the following information:

(1)    name, age, birthday;

(2)    a description of yourself both in terms of personality and appearance (including hair and eyes);

(3)    a description of each family member’s appearance;

(4)    a description of their personality;

(5)    a description of one or more pets you have;

I have also used short translation tasks (e.g. narrow translations) for the very same purpose.

Other types of tasks adopted

As I will discuss in a future post, my students do not simply engage in highly-structured tasks. I typically assign six other types of writing tasks:

(1)    Brainstorming writing – (usually in pairs) students brainstorm ideas on a given topic and note them down in full sentences in the target language. Accuracy is not a concern, as the purpose of this activity is to train them in idea generation.

(2)    Unstructured communicative writing – accuracy not a concern. Grading based on effective communication only.

(3)    Fluency writing – the students are asked to write about a topic in a very short amount of time

(4)    Editing tasks – (see above)

(5)  Transformative writing (e.g sentence combining; paraphrasing; summarising)

(6) Creative writing

Did it work?

The results were very encouraging. A survey showed that my year 9 students found Focused Correction more helpful than previous forms of error correction they had experienced, because they felt it was more manageable and the narrower range of mistakes to deal with meant they knew what to look for when they were editing and/or asked to self-correct.

Although there were fewer mistakes in all the FCAs targeted throughout the year at post-test than pre-test, to attribute the observed increases in accuracy solely to the corrective methodology adopted is problematic. The reduction in error-making is likely to be the result of the synergy between teaching and corrective feedback, and in particular the fact that the two supported each other more closely than in the error treatments traditionally adopted.

 

Concluding remarks

Error Correction is a time-consuming necessity. Over the years, I have experimented with various techniques. When it comes to beginner-to-novice students, Focused Correction combined with explicit editing instruction has been the most effective with my novice-to-intermediate for the following reasons:

(a) it concentrates on only two or three major target areas (so that the students’ attentional resources can be used more efficiently);

(b) such areas are perceived as relevant to their current learning (so that they feel more motivated to address them);

(c) it provides frequent feedback on the same target areas week in week out (so that it enhances their understanding and keeps them constantly focused on the same error types week in week out);

(d) students have the declarative knowledge for self-correcting the errors with minimal prompting by the teacher (so that students are self-reliant in the process );

(e) the students’ levels of accuracy in the target areas has a substantive bearing on the grading system adopted (so that students are more motivated to be more accurate).

Give it a try !