| Crynodeb: | Statistics and statistical evidence have been and are an important feature of litigation. Although
common, how to appropriately manage and utilise statistical evidence in fact-finding has
bedevilled courts across the common law world and led to rampant scholarly and judicial
debates. If evidence ought to lead to a greater likelihood that the truth will be uncovered, it
should be used in fact-finding. So far, what has been missing from the debates about statistical
evidence is how the legal process should come to a view that statistical evidence is ‘good’ for
fact-finding. The goal of this thesis is to articulate a framework by which decision-makers can
use evidence law to admit, assess the usefulness of, and weigh statistical evidence. This thesis
also seeks to address how the objections to statistical evidence may be categorised within the
evidentiary processes of relevance, admissibility, weight and proof. First, the thesis describes
objections to ‘General Factual Causation’, being the ability of the statistical evidence to be
adduced to prove the existence of associations between variables in a state of nature. Second, it
addresses the problem of how to infer the existence of a phenomenon in an individual from
statistical evidence purporting to show an association between variables in a reference class. This
question more than any other has dominated the discussion of statistical evidence, but it often
misses how evidence law would and could respond to this issue. Third, this thesis canvasses
difficulties arising from the proof of facts by statistical evidence and ‘objective probabilities’
derived from reference classes, including the problem of ‘naked statistical evidence’ and how or
whether to use mathematical techniques to calculate the probabilities of individual events.
|
|---|