Abstract
This study explores student perceptions of online versus conventional English teaching effectiveness methodologies on oral proficiency development among Chinese non-native university learners, addressing the critical need for evidence based pedagogical approaches and instructional strategies in contemporary language education. A mixed-methods pilot investigation was conducted with 30 first-year Chinese university students using four validated measurement instruments assessing oral proficiency, online learning experiences, conventional learning experiences, and student engagement. Data analysis employed independent samples t-tests, effect size calculations using Cohen's d, and correlation analysis to examine differential teaching effectiveness patterns. Online learning demonstrated superior effectiveness in developing pronunciation clarity (d = 0.78, large effect) and vocabulary appropriateness (d = 0.74, medium-large effect), while conventional classroom instruction showed significant advantages in spontaneous response generation (d = −0.67, medium effect) and face-to-face communication competence (d = −0.89, large effect). Engagement-proficiency correlations varied substantially across instructional modalities, indicating different psychological and pedagogical mechanisms. The differential effectiveness patterns reveal complementary rather than competitive relationships between online teaching platforms and traditional instruction, supporting integrated pedagogical frameworks that leverage technological advantages for technical skill development alongside interpersonal benefits for authentic communicative competence. Findings inform curriculum design strategies incorporating interactive digital tools, gamified learning modules, and self-assessment functions within comprehensive pedagogical frameworks, while providing guidance for individualized learning modality selection based on targeted learning outcomes, specific skill development objectives, and student characteristics.

This publication has 34 references indexed in Scilit: