While computational models of impact events have the potential to accelerate the design cycle, one's confidence in a material model should be related to the extent of validation work that has been performed for that model. Quantities of interest used for validation are often either scalar volume-averaged quantities (such as the average density, or the force applied to a boundary) or field quantities (such as the strain field obtained from digital image correlation, or density maps computed from X-ray computed tomography (XCT)). Volume averaged quantities are easy to compare quantitatively since they are either a single value or a simple time series. However, these averaged quantities do not capture differences in the failure process within a material and can be blunt instruments for validation efforts. Field quantities provide spatial information, but are difficult to reduce to a scalar that quantifies the goodness of a particular model with respect to another model. This work describes an approach to using XCT data to quantify how well a particular simulation agrees with simulation data while accounting for the statistical nature of failure in brittle materials.