Robust Dominance Levels: code and data for reporting composite-index rankings under construction uncertainty, with an application to the 2026 Environmental Performance Index

This project implements and tests the Robust Dominance Levels (RDL) algorithm to handle construction uncertainty in composite‑index rankings, using the 2026 Environmental Performance Index (EPI) as a case study. In environmental engineering, composite indices often depend on subjective choices: which indicators to include, how to normalize them, how to aggregate them, and how to assign weights. The RDL framework reduces these problems in three steps. First, it screens indicators by measuring their effective influence and how well they separate pairs of units. Second, it simulates the multidimensional space where the index is built and calculates the probability that one unit outranks another for every possible pair. Third, it keeps only

those order statements whose probability exceeds a calibrated threshold, removes any directed cycles, and creates an acyclic dominance graph that defines stable, layered dominance levels. The work includes building the uncertainty models, running sensitivity analysis on the 2026 EPI data set (177 countries, 47 indicators), and checking how robust rank changes are over time with the panel extension. The final framework provides a mathematically sound way to assess environmental risk and evaluate policy performance when epistemic uncertainty is high.

Message: Hi FE, I need help with “Robust Dominance Levels: code and data for reporting composite-index rankings under construction uncertainty, with an application to the 2026 Environmental Performance Index” in “Environmental Engineering”