QUANTIFYING BURDENS OF PROOF: A LIKELIHOOD RATIO APPROACH

In previous studies, burden of proof has been quantified using estimates of probabilities associated with the forensic standard "beyond a reasonable doubt." It is argued here that quantification of burden of proof requires consideration of prior opinion and that the relationship between prior opinion and posterior opinion associated with the "beyond a reasonable doubt" standard defines an inferred measure of the strength of evidence necessary to arrive at this standard. The methods used in this study differ from those used in other attempts to quantify these standards. Individuals were asked to indicate on an odds scale their prior opinion about the innocence of defendant and the odds value corresponding to an opinion about defendant's guilt beyond a reasonable doubt. The distance between these two estimates defines an inferred likelihood ratio measure of the strength of evidence required to convict beyond a reasonable doubt, and thus represents an inferred measure of burden of proof. Statistical issues in the analysis and reporting of odds estimates are also discussed.