Eficacia de los Sistemas de Retroalimentación Automatizada en la mejora de las habilidades de Escritura Académica en inglés

Resumen

The objective of this study was to analyze the impact of automated feedback tools on improving academic writing in English among university students, considering direct effects and explanatory mechanisms associated with mediating and moderating variables. A quantitative approach was applied with a quasi-experimental pretest-posttest design and control group with a population of 140 fifth-level students from the Language Center of the Technical University of Babahoyo, distributed in two parallel groups (control and experimental). The intervention integrated automated feedback platforms (Write & Improve, Grammarly, Ginger, and Quilbot), and grammatical accuracy, coherence-cohesion, syntactic complexity, and lexical range were evaluated as dependent variables; intensity and type of feedback were evaluated as independent variables; and self-regulation and motivation toward writing were evaluated as mediators. The analyses included descriptive statistics, t-tests for related and independent samples, ANCOVA controlling for the pretest, and mediation models. The results described showed increases between the pretest and posttest, highlighting the intensity of feedback (M = 49.96 to 65.30) and grammatical accuracy (M = 69.76 to 79.39). In the experimental group, pre-post comparisons showed significant improvements in all dimensions (p<0.001), and the posttest comparison between groups confirmed significant advantages for the experimental group (p<0.001), with relevant mean differences in grammatical accuracy (MD = 18.07) and coherence-cohesion (MD = 17.38). The ANCOVA showed very significant adjusted differences by group when controlling for pretest, defending the effectiveness of the intervention. Additionally, mediation indicated that self-regulation partially explained the relationship between feedback and writing indicators (indirect = 0.049; 95% CI [0.013, 0.085]; p=.007). Taken together, these findings support the notion that automated feedback significantly improves the quality of English writing and that its impact is enhanced by self-regulation processes.