A comprehensive measurement framework for performance assessment of the national innovation systems in EU countriesPopovska, Jasmina (Avtor) Umek, Lan (Mentor) Jaklič, Marko (Komentor) innovation policyinnovation performanceinnovation efficiencyinnovation capacitynational innovation systeminnovation indexData Envelopment AnalysisThis dissertation develops a conceptually coherent and empirically robust framework for assessing the performance of National Innovation Systems (NIS) in the European Union (EU). It addresses a central constraint of existing benchmarks, such as the European Innovation Scoreboard, regarding the conflation of innovation outcomes with inputs, and framework conditions, which limits diagnostic value and policy relevance. Building on innovation systems theory, the study distinguishes between NIS’s performance, efficiency and capacity. Methodologically, the dissertation introduces Multi-Cluster Feature Selection for transparent, data-driven indicator selection; applies a two-stage Data Envelopment Analysis model distinguishing knowledge production from commercialisation; and proposes the Efficiency-Adjusted Result-based Performance Index, which integrates performance and efficiency through a performance-first aggregation rule. The framework was empirically implemented for the EU-27 over 2017-2024. Findings reveal that result-based performance and efficiency are empirically independent dimensions, commercialisation constitutes the binding structural constraint across most EU countries, confirming the European Paradox at the efficiency level, and innovation systems cluster into four distinct performance-efficiency configurations rather than converging around the EU average. Robustness checks confirm ranking stability and construct validity. The framework advances innovation measurement by providing policymakers with stage-specific diagnostics for evidence-based policy design and structurally grounded peer benchmarking. Limitations include the sample-dependent nature of DEA efficiency, normative weighting choices, and a fixed two-year transformation lag. Future research could extend the framework through meta-frontier approaches, dynamic efficiency analysis, sector-differentiated lag structures, and application to regional or sector-specific innovation systems.[J. Popovska ]20262026-06-19 12:05:03Doktorska disertacija183829UDK: 330.34:330.341.1:061.1EU(043.2)VisID: 18114COBISS_ID: 282672131sl