On Conveying the Probative Value of DNA Evidence : Frequencies, Likelihood Ratios and Error Rates

Part I of this paper discusses the issue of what the frequencies associated with DNA evidence do and do not mean. Part II describes an alternate way of presenting DNA statistics in court based on Bayesian likelihood ratios. This part also addresses issues associated with identifying the hypothesis of interest and characterizing the evidence. Part III offers a likelihood ratio and estimates its magnitude based on proficiency test data that were analyzed by Koehler, Chia, and Lindsey. Part IV identifies and rebuts arguments that have been offered for ignoring the possibility of error when computing DNA likelihood ratios. Part V considers whether likelihood ratios belong in court and concludes that they may not if experts and jurors do not understand them. Part VI reports an empirical study that compares mock jurors' reactions to various numerical representations of DNA evidence. The results of the study indicate that there are good reasons to suspect that likelihood ratio presentations will confuse jurors