Cadaver data from experiments and data from field studies collected for the purposes of risk analysis are almost always censored. If the form of the underlying distribution is known, then the best method of analysis is to estimate the parameters using a maximum likelihood approach, but if it is not known, the best method is a nonparametric approach. The Consistent Threshold Estimate, introduced in this paper, is a method to estimate the underlying distribution that is both non-parametric and a maximum likelihood estimate. In this paper, we will use it to estimate threshold HIC values for skull fracture or tissue damage, but it can be used for any application that has censored data. In addition to mathematically defining the Consistent Threshold estimate, a simple method to compute it for doubly censored data is given and it is compared to other estimators by means of Monte Carlo Tests. The Consistent Threshold estimate is then applied to experimental head impact to cadaver data.