Evaluation of Sheen (2007) by Sang-Ki (2nd evaluation assignment)

Sheen, Y. (2007). The effect of focused written corrective feedback and language aptitude on ESL learners’ acquisition of articles. TESOL Quarterly, 41, 255-283.


What are the purposes of giving feedback on learner errors? This is surely an important question to be addressed when evaluating studies on error feedback. I understand that the purposes can vary substantially depending on the language mode, that is, whether the feedback is given orally or given in a written format. This is because written feedback could demand less cognitive load for learner memory than oral feedback, since feedback given in a written format will be available for a longer period of time than the other type of feedback. Thus, by giving written feedback, teachers may expect their students to benefit more not only in terms of the depth of its effectiveness, but they also expect that the benefit would reach wider contexts of language use.

Most of the oral feedback studies to date have focused on the correction of form-related errors, with an (exclusive) aim of improving the accuracy of a focused linguistic element (e.g., Doughty & Varela, 1998; Iwashita, 2003; Long, Inagaki, & Ortega, 1998; Mackey & Philp, 1998). Conversely, written feedback studies have usually had varied foci other than the improvement of accuracy in student writings; they have also intended to improve writing content, organization, and the overall quality and fluency of writing products. This difference in purposes may account for why “in contrast to the SLA research that in general has shown that oral CF [corrective feedback] is effective, L2 writing researchers have not been able to convincingly demonstrate that written CF leads to improvement in grammatical accuracy in new pieces of writing” (p. 257).

The researcher’s main question in the current study is whether written feedback, if given in a focused manner, improves the accuracy in writing. In order to employ the focused type of feedback by design, teacher written feedback was given only on the article-related errors (i.e., inappropriate uses of a and the) as target to be acquired by the 91 adult intermediate-level research participants. The two factors presumed by Sheen to affect the efficiency of written feedback were (1) two types of written feedback: direct-only correction (provisions of corrected forms) versus direct metalinguistic correction (provisions of corrected forms and metalinguistic comments) and (2) individual differences in language-analytic ability. The patterns of major findings turned out to be predictable. The focused, metalinguistic feedback was found more beneficial for accuracy improvement, and the positive effects appeared to last over a month. In addition, significant benefits from the higher levels of language-analytic ability were found, and the benefits were most prominent within the direct metalinguistic correction condition. The helpful role of metalinguistic information in grammar learning would be in line with what has generally been found from the accumulated oral feedback studies (e.g., Bitchener, Young, & Cameron, 2005; Carroll & Swain, 1993; Ellis, Loewen, & Erlam, 2006).

I enjoyed reading this article, and I think this article is useful for a couple of reasons at least. First, this study is the first to my knowledge which looks into the effectiveness of written feedback on accuracy improvement in student writings, with methodologically sound designs (e.g., true control group design with a focused target rule, and the like). Second, this study rightfully switches our attention to the mediating roles of individual differences in grammar learning through written feedback (i.e., individual differences in language-analytic ability). An improvement of the study design could be made, however, if the individual difference factor could serve as an independent variable. (The current study examines only the correlations between language-analytic scores and acquisition scores.) That is, having different independent groups with differential language-analytic abilities and giving them various types of written feedback could shed more light on the potential causal role of the individual difference variable across different feedback types.

Again, if I have rating scales, I would vote for the “recommended” option.

(Personally, I prefer this article to my previous choice (Sheen, 2006) as our course reading.)