4.13 Probative value of evidence

The probative value of evidence is the degree to which a piece of evidence can discriminate between competing propositions. When the evidence is scientific, its probative value is determined by the expert and communicated to the court. It is essential that the expert does this using a well-calibrated and validated process (see Section (reliable-probabilities)), so that the determined probative value does not discriminate between propositions with undue bias. The fact finder then decides the truth of the competing propositions based upon this and any other evidence that is relevant to the propositions.

In order to determine how much the evidence discriminates between competing propositions, the expert considers the competing propositions from counsel and determines how likely the observed evidence would have been assuming each proposition to be true. This logic addresses the questions: how likely would the observations have been if the prosecution’s assertions are true? and how likely would the observations have been if the defence’s assertions are true? Both of these questions reflect uncertainty in the evidence conditioned on the propositions of the prosecution and defence. This means that they can also be described as conditional probabilities: probabilities of the evidence conditioned on the propositions and any background information (such as agreed facts).

Highly discriminatory evidence, or high probative value, is achieved when there is a large difference between the probabilities under each proposition. If a piece of evidence cannot discriminate between propositions then it is not probative for those propositions, but it may still be probative of other propositions. Little-to-no probative value occurs when there is little to no difference between the conditional probabilities under each proposition. These probabilities should have their calibration validated to ensure that any discrimination is fair and unbiased.

In the examples in the next section, we will use the idea of a match. The word match has a specific meaning when referring to a comparison between a questioned sample and a reference sample, such as comparing a finger mark found at a crime scene with a fingerprint taken from a person of interest to an investigation. A match generally means that a selected number of features have been compared between the two samples and that a critical number of those features have been determined as indistinguishable. For example, comparative features of fingerprint examination include skin ridges on the finger such as the ending of a ridge or the splitting of a ridge into two ridges (known as bifurcation), as well as scars and other damage to the skin. The features that are compared and the critical number that determines a match differ between evidence types. The interpretation of a match requires care because it does not necessarily mean that the two samples originate from the same source or that an alleged activity occurred.