A comparative study of DEA-based ranking methods applied in multi-sport events

Purpose To compare DEA-based ranking methods and develop a fair approach to evaluate states’ efficiency and rankings at the National Games, India's national-level multi-sport event, adjusted for population and wealth, while excluding less relevant macroeconomic factors.Design/methodology/approach Using data from the 2023 Games (37th edition), we compare three widely applied Data Envelopment Analysis (DEA) models: variable returns to scale (VRS), zero-sum gains (ZSG), and cross-efficiency. Results are validated using data from the 35th edition of the National Games to ensure robustness.Findings The cross-efficiency method best produces complete rankings without ties. The VRS method identifies peer groups for the seventeen inefficient states out of 29 participating teams.Originality This study contributes by providing a comparative evaluation comparing DEA models and identifying the most suitable method to assess states’ efficiency in promoting athletes at multi-sport events. It also highlights each method’s particularities, suitability, and characteristics.Research Contribution By combining technical rigor with practical relevance, the paper addresses a gap in the literature by comparatively evaluating VRS, ZSG, and cross-efficiency models, underscoring their methodological distinctiveness and applicability.Practical implications It provides actionable insights for state sports bodies, organizers, and policymakers to optimize resources, support talent development, and strengthen future National Games performance.