ACCTG 1 Lecture Notes - Lecture 14: Expert Witness, Bayesian Network, Likelihood Ratios In Diagnostic Testing
Document Summary
Likelihood ratio paradigm or bayesian approach to evidence (page 193) All that is needed to estimate likelihods of evidence occurring under primary and under alternative scenario is to put them in this table and divide the elft column likelihood by the right column likelihood. Challenges of introducing likelihoods for practical usage (courts) It is unclear whether the data should be made up when true and false positives cannot e estimated in a scientifically backed up manner, or when it should just be accepted that it i impossible to generate likelihoods. What to do in cases in which a piece of evidence is provided with a likelihood ratio. Multiply:only warranted if pertinent pieces of evidence areindepenent from each other (which in practice is not always the case) Solution; bayesian network analysis, because associations between various pieces o evidence can be controlled for. Apart from limitations, the use of the likelihood ratios in legal decision making would offer protection against various biases.
