Аңдатпа
Much of the policy discussion around digital education proceeds on the assumption that technically capable systems, if properly implemented, will translate into improved learning outcomes. The empirical record does not consistently support this expectation—and the inconsistency is not random. This systematic meta-synthesis of 147 SSCI-indexed studies from 26 national contexts (2015–2025) examines why, by shifting the analytical lens from what technologies can do to the governance conditions determining whether they will. Following PRISMA 2020 guidelines, studies were coded across five dimensions (Cohen's κ = 0.81) and analyzed through a two-stage pipeline combining hierarchical clustering with crisp-set qualitative comparative analysis (csQCA). Four configurations emerged above the consistency threshold of 0.75. AI/ML tools combined with decentralized governance and explicit outcome alignment yielded positive outcomes in 73.4% of cases (consistency = 0.84); any technology under centralized governance without alignment, only 31.2% (consistency = 0.82). Outcome alignment was present in 94.1% of positive-outcome studies; formal necessity-analysis consistency and coverage values were not calculated, so this is reported as a descriptive association rather than a necessity finding, and the identification problem between alignment-as-cause and alignment-as-institutional-quality-marker remains unresolved. The Technology–Governance–Outcome (TGO) framework repositions institutional design as the constitutive condition of technology effectiveness, not its organizational backdrop. Two structural tensions run through the findings: centralized and decentralized governance both achieve positive outcomes, but through mechanisms that cannot be reconciled into a single policy prescription; and the governance conditions most predictive of success are systematically least available to the institutions most in need of improvement. The synthesis provides an early quantitative reference point for Central Asian digital governance, surfacing intra-national variation that aggregate statistics obscure.