Aligning ChatGPT with e-portfolio assessment as EFL learning model: Its effect on student’s speaking performance and feedback literacy

Abstract

The rapid development of technology, particularly Generative Artificial Intelligence (Gen AI) and e-portfolio assessment, has transformed English language learning practices. Many studies about these two technologies have shown positive impacts separately. However, the implementation of their combined effect on language learning outcomes remains unexplored. This study investigates the impact of aligning ChatGPT with e-portfolio assessment (CEA) on students’ speaking performance and feedback literacy in the English for Interpersonal Communication course at a vocational college. Following an explanatory sequential mixed-methods design, this study involved 60 first-semester students divided into experimental and control groups. The experimental group experienced CEA integration. However, the control group experienced a conventional e-portfolio without exposure to ChatGPT. Data were collected through speaking tests, feedback literacy questionnaires, and semi-structured interviews. The result of MANOVA analysis revealed significant simultaneous effects of CEA on speaking performance and feedback literacy (Wilks’ λ = 0.476, p < 0.001, partial η² = 0.524). The experimental group outperformed their control group counterparts in speaking performance (M = 82.733) and feedback literacy (M = 79.666). The thematic analysis of the interview data revealed seven key themes related to the student’s feedback literacy experience in the CEA implementation. These findings indicated that the integration of ChatGPT with e-portfolio assessment created a dynamic learning ecosystem that enhanced language skills and feedback literacy. The result of this study offered a promising framework for technology-enhanced language learning in vocational education.