LMS as a determinant of academic success of international students in Ukrainian as a foreign language testing at Ukrainian technical universities

The article provides an empirical evaluation of the impact of learning management system (LMS) utilization on the academic performance of international STEM students during high-stakes summative certification examinations in Ukrainian as a Foreign Language (UFL). The study substantiates that, under distance and blended learning paradigms, the LMS transcends its traditional role as a passive technical repository, evolving into a complex, integrated digital ecosystem that delivers systematic pedagogical mediation, continuous behavioral monitoring, and compensatory digital scaffolding in the absence of an immersive target-language environment. Methodologically, a quasi-experimental research design was adopted at the National Aerospace University "Kharkiv Aviation Institute". The study comprised 146 international students engaged exclusively in distance learning modalities (86 native speakers from the People’s Republic of China and 60 from the Republic of Azerbaijan) who completed the UFL course hosted on the MENTOR LMS platform. A comparative synthesis of the longitudinal dynamics demonstrated a highly significant, positive performance trend across both groups: the mean score for the Chinese students increased from 52.4 to 79.1 points (a relative growth rate of 50.95%), while the Azerbaijani students advanced from 63.2 to 84.5 points (a relative growth rate of 33.70%). The data reveal that, while statistical parity was maintained across the aggregate LMS utilization indices (0.81 and 0.79, respectively), the most pronounced longitudinal acceleration occurred among learners navigating initial language acquisition from lower baseline competencies. It is concluded that the LMS functions as a critical pedagogical mediator, optimizing autonomous learning, fostering metacognitive self-regulation, and mitigating affective barriers. Future research directions are outlined, emphasizing the need to apply multiple linear regression and multivariate statistical analyses to granular, student-level digital footprints.