Correct every error or let students discover their mistakes? The effects of corrective feedback strategies on EFL learners’ grammatical accuracy and retention

Abstract

The study examined the effect of direct corrective feedback (DCF) and metalinguistic corrective feedback (MCF) on grammatical accuracy and retention of university level English as a Foreign Language (EFL) learners. A quasi-experimental intervention lasting eight weeks was carried out among forty undergraduate students in a public university in Australia. Subjects were grouped into three types of instructions: DCF group (n = 13), MCF group (n = 13) and conventional instruction group (n = 14). Each group was provided instruction focused on similar grammatical forms such as past tense, present perfect, articles, subject-verb agreement, prepositions, conditionals and modal verbs; the main variation was the type of feedback provided during the writing task. Pre-test, post-test, and delayed post-test of grammar test and writing task were conducted among the subjects. Quantitative data analysis made use of ANCOVA and linear mixed effects models. ANCOVA revealed a significant difference of instructional conditions on post-test grammatical accuracy in terms of the effect of pre-test (2, 36) = 3.95, p = .028, partial η² = .180. The DCF group had the highest post-test mean score (M = 30.92, SD = 2.10), followed by the MCF group (M = 30.62, SD = 6.12), whereas the conventional group had the lowest mean score (M = 27.36, SD = 3.57).