Determination of human tolerance to impact-induced damage or injury is needed to assess and improve safety in military, automotive, and sport environments. Impact biomechanics experiments using post mortem human surrogates (PMHS) are routinely used to this objective. Risk curves representing the damage of the tested components of the PMHS are developed using the metrics gathered from the experimental process. To determine the metric that best explains the underlying response to the observed damage, statistical analysis is required of all the output response metrics (such as peak force to injury) along with the examination of potential covariates. This is conducted by parametric survival analysis. The objective of this study is to present a robust statistical methodology that can be effectively used to achieve these goals by choosing the best metric explaining injury and provide a ranking of the metrics. Previously published data from foot-ankle-lower leg experiments were used with two possible forms of censoring: right and left censoring or right and exact censoring, representing the no injury and injury data points in a different manner. The statistical process and scoring scheme were based on the predictive ability assessed by the Brier Score Metric (BSM) which was used to rank the metrics. Response metrics were force, time to peak, and rate. The analysis showed that BSM is effective in incorporating different covariates: age, posture, stature, device used to deliver the impact load, and the personal protective equipment (PPE), i.e., military boot. The BSM-based analysis indicated that the peak force was the highest ranked metric for the exact censoring scheme and the age was a significant covariate, and that peak force was also the highest ranked metric for the left censored scheme and the PPE covariate was statistically significant. IRCs are presented for the best metric.