When evaluating evidence for use in a court of law, evidential data are presented together with two competing propositions which may account for this data: one from the prosecution and one from the defence. The aim of the court is to evaluate which of the two propositions is more likely. In other words, is the probability of the prosecution hypothesis given the evidence bigger than the probability of the defence hypothesis given the evidence?
We want to know whether the ratio of these probabilities is greater than one (overall support for the prosecution hypothesis), or less than one (support for the defence).
This ratio is known as the Likelihood Ratio and it is meant to represent the odds before seeing any evidence to the odds taking the evidence into account.
What is the theory and models behind this approach? Dr. Richard Sleeman explains further in the attached paper from March 2015 issue of Mass Matters Magazine.
MM-March2015-p14-v5-final-(1).pdf
Image Credit: © Aleksandar Radovanovic | Dreamstime.com