Enhancing collaborative writing with AI-enhanced feedback in graduate-level action research courses

Abstract

Purpose This study investigates the comparative effectiveness of peer-reviewed feedback and AI-enhanced feedback, specifically through Grammarly, in graduate-level online action research courses. Grounded in constructivist and sociocultural learning theories, the study explores how students perceive and apply feedback from either peers or an AI-based tool across dimensions such as feedback characteristics, application and writing quality outcomes. Design/methodology/approach A convergent parallel mixed-methods design was employed with 39 graduate students who selected their preferred feedback mechanism. Data collection included surveys, feedback characteristics, revision tracking and rubric-based assessments. Statistical tests assessed feedback application and writing improvements. Findings Findings showed that AI feedback via Grammarly was more effective in addressing surface-level issues such as grammar and clarity, while peer feedback more often engaged with higher-order concerns like organization and originality. Both feedback types contributed to improved writing quality, with peer feedback yielding greater gains. Grammarly was chosen for this study due to its wide availability in higher education and institutional support at the time of data collection in early 2023, before the widespread classroom integration of generative AI tools like ChatGPT. While the results suggest complementary strengths between human and AI feedback, limitations include the narrow scope of AI used and reliance on feedback frequency as a proxy for feedback quality. Originality/value This study uniquely compares peer and AI feedback in graduate writing, allowing students to choose their preferred method. The findings highlight the strengths of both approaches and emphasize the need for a balanced integration of AI and human feedback in academic settings. This research contributes to AI-assisted education by offering insights into optimizing feedback strategies for enhanced learning outcomes.

Ähnliche Arbeiten