Abstract
Although mobile learning offers novel pedagogical opportunities, the development of advanced speaking abilities in German as a Foreign Language remains insufficiently examined, leading to the psycholinguistic fluency-accuracy trade-off being largely overlooked in synchronous environments. This study sought to assess the effectiveness of a unique mobile framework, the Asynchronous-Collaborative Construction Cycle (ACCC Model), in improving B2-level speaking proficiency. This research employed an explanatory sequential mixed-methods design. A cohort of 38 Iranian B2 learners was allocated to either a mobile experimental group (n = 20) utilizing asynchronous WhatsApp or a synchronous face-to-face control group (n = 18). Data were gathered through pretest-posttest Goethe-Certificate B2 oral assessments and a verified learner perception survey. Quantitative data were subjected to ANCOVA and manual Complexity, Accuracy, and Fluency (CAF) analysis, whilst qualitative replies were assessed by inductive content analysis. The experimental group markedly surpassed the control group in overall speaking proficiency (p < .001). Importantly, they attained concurrent, substantial improvements in Accuracy and Fluency, although Syntactic Complexity showed no notable intergroup differences. Qualitative data indicated that asynchronicity contributed to this accomplishment by reducing speaking anxiety and allowing for iterative acoustic self-monitoring. Separating language output from immediate demand effectively addresses the fluency-accuracy trade-off, illustrating that including asynchronous mobile practice is a highly scalable and effective educational method for institutional curricula.