Beyond English-only feedback: an exploratory classroom-based study of translanguaging written corrective feedback and L2 writing accuracy

Abstract

Written corrective feedback (WCF) is widely used to promote second language (L2) writing accuracy, but research on translanguaging in feedback has focused predominantly on peer or oral feedback. Little is known about whether planned translanguaging in teacher-provided written corrective feedback can enhance learners' engagement with feedback and subsequent writing accuracy. This exploratory study therefore examined whether strategically integrating Arabic and English into teacher-provided WCF was associated with greater feedback adoption and more favorable immediate and delayed accuracy development than English-only WCF. The study employed an exploratory classroom-based pre-test/post-test/delayed post-test quasi-experimental design involving 70 first-year Saudi EFL university students at CEFR B1 level, assigned through two intact classes to a translanguaging condition ( n = 36) or an English-only condition ( n = 34). Both groups received direct WCF with metalinguistic explanations targeting subject–verb agreement, the English article system, and cohesive markers. In the translanguaging condition, Arabic and English were strategically interwoven, with English used for corrected forms and Arabic used primarily for complex explanations and crosslinguistic reasoning. Feedback adoption was measured through learners' revisions, while writing accuracy was operationalized as the proportion of erroneous obligatory contexts. Accuracy was assessed at pre-test, immediate post-test, and delayed post-test; 55 students completed all three assessments. The translanguaging group adopted a significantly greater proportion of feedback in revision than the English-only group (M = 0.81 vs. 0.68, p = .006), with the difference driven primarily by cohesive-marker feedback. Among the 55 completers, the translanguaging group's overall error rate declined from 0.49 at pre-test to 0.22 at post-test and 0.16 at delayed post-test, whereas the English-only group's rate increased from 0.40 to 0.45 and 0.49, respectively. The Test Time × Group interaction was significant ( p < .001). Baseline-adjusted analyses also showed lower overall error rates for the translanguaging group at both the immediate and delayed post-tests. Target-specific analyses indicated significant advantages for articles and cohesive markers at the immediate post-test and for subject–verb agreement, articles, and cohesive markers at the delayed post-test. The findings provide preliminary evidence that planned translanguaging in teacher-provided WCF may facilitate feedback adoption and support more favorable short- and longer-term accuracy trajectories, particularly for discourse-level features and forms requiring crosslinguistic explanation. However, because the intervention involved only one class per condition and the groups differed at baseline, the findings should be interpreted as exploratory associations rather than definitive causal effects. Multi-cluster or randomized replications are needed to determine whether the observed advantages can be attributed specifically to translanguaging WCF.