Analysis of the educational performance of European Union countries and Türkiye through multi-criteria decision-making methods within the framework of the 2030 Barcelona targets

This study aims to comparatively evaluate the educational performance of the 27 European Union member states and Türkiye based on nine key indicators identified in line with the 2030 Barcelona Targets of the European Education Area. In the study, the weights of the criteria were determined using the LOPCOW method based on objective data, while the ranking procedures were carried out through the AROMAN and CRADIS methods. The indicators considered in the evaluation consist of digital skills, the NEET rate, early school leaving, participation in early childhood education, education in STEM fields, participation in lifelong learning, achievement in basic skills, and higher education graduation rates. The findings indicate that indicators such as early childhood education, digital skills, and the NEET rate become particularly prominent in terms of their criterion weights. Türkiye's placement in the lower group in the overall ranking reveals the continuing need for structural policy interventions in digital transformation, inclusive education, and lifelong learning In addition, sensitivity analyses were conducted by changing the criterion weights under different scenarios; as a result of these analyses, it was observed that weight variations in certain indicators affect the country rankings. The sensitivity analysis was applied to test the stability of the established model and to determine to what extent the results are maintained against changes in criterion weights. This situation demonstrates the multidimensional nature of educational performance and the impact of interactions among criteria on ranking outcomes. Moreover, the rankings obtained through the CRADIS and AROMAN methods were compared with those derived from other multi-criteria decision-making methods such as TOPSIS, CODAS, EDAS, MARCOS, and WASPAS, and the level of methodological consistency was examined through correlation analysis. The variation of correlation coefficients between 0.769 and 0.984 indicates that the methods rank countries with largely similar tendencies and that the model produces stable results despite methodological differences. While the strongest relationship was identified between MARCOS and CODAS (ρ = 0.984), the weakest relationship was observed between CODAS and CRADIS (ρ = 0.769); furthermore, the high level of agreement between the rankings obtained from CRADIS and AROMAN (Spearman ρ = 0.947) strengthens the internal consistency of the methodological approach employed in the study. The methodological framework applied in the study enabled an objective, comparative, and multidimensional analysis of education policies and contributed to identifying critical intervention areas particularly for low-performing countries. The findings provide guidance for policymakers in developing strategies aimed at improving digital skills, expanding early childhood education, and reducing NEET rates.