All forensic opinions can be characterized as probabilities, and all forensic methods
upon which the probabilities are based can be described as tests of relevant evidence.
Probability can also be described as a way to quantify uncertainty (as [1-probability]
= uncertainty) and thus a means of assessing the validity of a forensic opinion. In the
present paper a method of quantifying the uncertainty in forensic test results called the
Error Odds is described. The Error Odds is a Bayesian metric that allows for calculation
of the degree of uncertainty in a test result by using ordinary clinical terms and concepts.
As an example of how the method can be easily applied to a common test result, the
Error Odds calculation is used to quantify the uncertainty in the use of unexplained
pediatric fracture as a test for child abuse. It is suggested that an Error Odds of 10:1
is the minimum threshold for the consideration of a single test result as evidence of guilt in a criminal proceeding