The notion of speaker individuality and the reporting of conclusions in forensic voice comparison

Abstract

This contribution addresses some principal issues in forensic voice comparison, reflecting on some of the topics which have dominated the discussion among experts in the past two decades. The issue of speaker individuality is linked to the way in which conclusions in forensic voice comparison cases are expressed. The recent discussion about expressing conclusions in terms of likelihood ratios in forensic voice comparison is critically reviewed here. It is argued that likelihood ratios are not as unequivocal as they are said to be, neither are they popular with the triers of fact. Results of a survey among members of the judiciary on this topic are presented. It demonstrates once again that verbal probabilities are preferred even though, strictly speaking, they are logically flawed.