Home/Paper Details

Archive

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

AI-Assisted Academic Writing: A Quasi-Experimental Study Among Grade 10 Students

Open Access
Journal Type:Research Article
Subject Field:Linguistics and Education
Downloads:11
Publish Date:May 11, 2026 11:23 pm
Views:73
Volume:196, Issue: 1, May, 2026
Subject:Education
Pages:489-504

Abstract

The continued difficulties in writing among high school-age Filipino students are also affected by the very large classroom sizes found in many Philippine Public Schools. This quasi-experiment evaluated the potential of AI-based writing tools to aid in the development of a student’s academic writing skills. Over a four-week period, 50 students in two intact classrooms were divided into an Experimental Group that used the AI-based writing tool Grammarly for grammar checking and the AI based brainstorming tool ChatGPT when completing essays, and a Control Group that received standard instructional techniques. Prior to treatment, there existed significant differences in the pretest writing skill abilities of the two intact groups (Experimental Mean = 18.28, Standard Deviation = 1.86; Control Mean = 12.62, Standard Deviation = 4.12; t(48) = 8.06, p < .001, Cohen’s D = 1.60). To address these baseline ability differences, Analysis of Covariance was performed with the prior test as the covariate. Following the use of analysis of covariance to adjust for the baseline ability differences, results indicated that the Experimental Group significantly exceeded the Control Group on the posttest (F (1, 47) = 11.05, p = .002, Partial Eta Squared = .190; Adjusted Mean for Experimental Group = 18.10; Adjusted Mean for Control Group = 16.56). Furthermore, the Experimental Group demonstrated substantial within-group growth (Cohen’s D = 0.91). On the Perceptions Survey, ratings from the Control Group indicated that they believed AI-based writing tools were more useful, easier to use, and more satisfying than those from the Experimental Group – a result contrary to expectations. However, both groups expressed a desire to continue using AI-based writing tools. Ninety-six percent of respondents from the Experimental Group expressed interest in continuing to utilize AI-based writing tools, and ninety-two percent of respondents from the Control Group expressed similar interest. Overall, the results suggest that AI-based writing tools can serve as intelligent scaffolding within Vygotsky’s Zone of Proximal Development and provide a replicable example for the Department of Education’s “Education Center for AI Research” (E-CAIR) initiative.

© 2026 International Journal of Research Publications (IJRP). All rights reserved.