ABSTRACT
If forensic measurements were one-time events like measuring a density, you would already have the knowledge necessary to develop and use uncertainty estimates. To expand into more realistic situations, new tools and concepts are needed. Measurements made in forensic laboratories are rarely one-time events such as our density example; we need robust uncertainty estimations for measurements that may be made tens or hundreds of times a year by one or more analysts using one or more pieces of equipment and instrumentation. In this situation, it is neither feasible nor appropriate to generate uncertainty estimations on the fly with every measurement made. Instead, we need to develop uncertainty estimations that can be applied every time a measurement is made. The approach is different, but the concepts are the same.
