From Prompting to Proficiency: A Mixed-Methods Analysis of Prompting with ChatGPT Versus Lecturer Interaction in an EFL Classroom

Abstract

The rapid advancement of artificial intelligence (AI) and extensive language models like ChatGPT have had profound implications for English as a Foreign Language (EFL) pedagogical practices. However, while its applications are widely explored, a notable gap persists in understanding the effectiveness of its core "prompting" feature for academic help-seeking compared to traditional lecturer interaction. This study aimed to address this gap by comparing the effectiveness of ChatGPT versus lecturers, exploring the advancement of student prompting strategies, and identifying associated challenges. A sequential explanatory mixed-methods design was employed over eight weeks with 60 Indonesian university students allocated to an experimental (ChatGPT) and a control (lecturer) group. Data collected via pre/post-tests (proficiency, writing, self-efficacy), interviews, and a Focus Group Discussion were analyzed using ANCOVA and thematic analysis. The quantitative findings conclusively demonstrated that the ChatGPT group significantly outperformed the control group in enhancing general English proficiency, writing competency, and self-efficacy (p < .001). Qualitatively, students’ prompting strategies evolved from simple, single-turn queries to sophisticated, multi-turn dialogic interactions, which were a key determinant of deeper learning. The most critical challenge identified was pedagogy, specifically the tendency for cognitive offloading alongside difficulties in vetting the AI’s accuracy and pervasive anxiety regarding academic integrity. These findings suggest that the effective use of Generative AI (GenAI) is not an innate skill but a learned competence, necessitating a pedagogical shift from merely providing technological access to the explicit instruction of AI literacy and strategic prompting to harness its full potential.

Ähnliche Arbeiten