The aim of the study reported in this paper was to provide rational aid tools to quantify the performance of nondestructive testing (NDT) tools. This study focused on the quantification of the performance of impact–echo sources (steel balls of varied diameters), applied with a new, con-tactless robot, for duct void detection and thickness measurements in a reinforced concrete wall. Because of uncertainties during the testing, the data were analyzed in a probabilistic context, with the knowledge that on-site inspections were affected by uncertainties. The αδ method was used in this regard, where the probabilities of detection and false alarm rates were used to build receiver operating characteristic curves. The methodology was applied to data measured on the same wall with two steel ball diameters: 0.16 and 0.125 m. The quantity analyzed here was the impact–echo method (resonance) frequency. This methodology could be extended to other parameters of the impact–echo setup as well as to other NDT methods.