<![CDATA[Ensuring academic quality assurance (QA) in large-scale distance learning (PJJ) presents complex challenges influenced by technological, pedagogical, and social factors. Conventional analytical methods often fail to capture the probabilistic and causal relationships among these variables due to data uncertainty. This study aims to model and analyze the probabilistic and causal interactions that determine QA success in PJJ using an integrated approach that combines Structural Equation Modeling (SEM) and Bayesian Network (BN). Using a quantitative explanatory survey design, data were collected via questionnaires that covered variables such as technology availability, instructor support, student interaction, learning motivation, and administrative governance. Data analysis was performed using R software with the lavaan package for structural model evaluation and the bnlearn package for probabilistic network modeling. The regression analysis results indicate that the availability of technology, the quality of instructor-student interaction, and learning motivation are the primary determinants of QA success ($R^2$ = 0.577). SEM evaluation confirmed an excellent model fit (CFI = 0.999; TLI = 0.999; RMSEA = 0.011), with technology availability providing the largest relative contribution at 33.3%. The developed BN model effectively estimates QA success probabilities, finding that high learning motivation levels increase the likelihood of QA success to 0.70. Conversely, administrative support was not significant, and isolated administrative interventions tend to be ineffective at increasing QA probability ($P$ = 0.45). The integration of BN-SEM offers a comprehensive predictive framework that enables policymakers to conduct scenario simulations for digital education quality management.]]>
