Improving Voice Identification Procedures

In certain crimes, the voice of a perpetrator is heard by a victim or witness, but no recording of the incriminating speech is available. A robber may have been masked, or have attacked from behind, for instance; or contact may only have been over the telephone. If the witness has received sufficient exposure to the perpetrator's voice, earwitness evidence may be collected through a voice parade. A voice parade is conducted using a similar format to a visual identity parade, but using voices rather than faces: the witness is asked whether he or she can pick out the voice of the speaker heard at the crime scene from a line-up of recorded speech samples which includes the suspect's voice along with a number of 'foil' voices.

In England and Wales, the current guidelines on how a voice parade should be conducted (published in 2003) were developed as an extension of the police procedure for visual identification parades. However, since the original voice parade guidelines were developed, psychological research has emphasised that although face and voice processing exhibit many parallels, there are also marked differences, and further research is needed to ensure that the details of the voice identification procedure are set up in a way that optimises earwitness performance. The practical and resource requirements of the current procedure are time-consuming and expensive, and the willingness of police forces to engage with the procedure is very variable such that in practice very few parades are undertaken.

The IVIP project has four main strands, with the overall aim of improving understanding of earwitness behaviour and improving the interaction of the criminal justice system with the use of earwitness evidence. The first strand will examine aspects of voice identification which are yet to be fully tested within the current procedure (e.g. length of samples, number of foil voices, witness instructions, parade type), with a view to modifying the procedure to optimise earwitness performance. The second strand will investigate from a phonetic perspective why it is that certain speakers are more distinctive-sounding than others and whether speakers judged to be more distinctive are also more memorable. The third strand will entail a study of the degree to which social perceptions, judgements, attitudes and stereotypes related to voice(s) can motivate witness decision-making during voice parades. The final strand of the project will assess and evaluate the extent of police and legal practitioners' awareness and experience of voice parades, beliefs about earwitness memory, attitudes to conducting voice parades and how earwitness evidence is received in court.

The IVIP project will achieve significant impact in the criminal justice system, benefitting police and legal practitioners, the judiciary, forensic practitioners and, ultimately, the general public. In addition, its findings will make an important broader contribution to scholarship in phonetics, sociolinguistics, psychology and criminology. Crimes involving the witnessing of a voice are widespread, yet at present this evidence is rarely being adequately capitalised on in the United Kingdom and other countries. Improvements in the understanding of earwitness behaviour and in techniques for collecting earwitness evidence efficiently are essential in addressing these kinds of crimes. This project will offer advances in the implementation of voice parades and improvements in the legal interaction with earwitness evidence which will lead to improved outcomes in the justice system.

The IVIP project will achieve impact by benefitting a wide range of groups of people: the general public, agencies within the criminal justice system, forensic practitioners, researchers and students.

The general public will be the ultimate beneficiary of the research through its impact in the criminal justice system. There are many cases in which earwitness identification evidence could offer an additional source of evidence to assist in the prosecution of criminal cases, yet it is not currently being exploited. The public will benefit from the increased access to justice that will come about through an improved theoretical understanding of earwitness behaviour and the improved techniques for the collection of earwitness evidence. The research will optimise the chances of accurate earwitness identification occurring and reduce the risk of false convictions.

Further, the general public will benefit from the research in broader ways, through the project team's dissemination of our findings in general interest contexts, such as knowledge festivals, schools talks, community events, publications for a general audience, media appearances and the IVIP website. We aim to stimulate interest in the understanding of linguistic, phonetic, psychological, forensic and legal topics.

Agencies within the criminal justice system will immediately and directly benefit from this research, that is those engaged in the investigation, prosecution and prevention of crime, including the police, the Crown Prosecution Service, the Security Services, and their international counterparts. The project will examine current understanding of earwitness identification evidence and possible misperceptions by lawyers and police officers and address ways to improve the interaction of these key decision makers with this type of evidence. IVIP impact activities will inform this group about the developments in the understanding of earwitness behaviour and implications for voice parade methodology arising from the project.

The research will be of great benefit to forensic phonetic practitioners as direct users of its findings, either within the agencies above or in private consultancies. The results of IVIP will provide forensic analysts with a stronger, more detailed, and more reliable knowledge base from which to advise about and implement voice parade construction.

Beyond the direct application in analysis of earwitness identification evidence, the improved understanding of voice distinctiveness and individual speaker variation arising from this project will benefit other areas of forensic speech casework, in particular, forensic speaker comparison, in which comparison of suspect and perpetrator speech recordings is required. Again the general public, decision makers in the criminal justice system, and forensic practitioners will benefit from the improved accuracy and efficiency of this type of analysis brought about by the findings of IVIP.

Finally, the research will be of great benefit to researchers and students of a wide range of relevant disciplines. Linguists, phoneticians, and forensic scientists will benefit more an improved understanding of earwitness behaviour, individual speaker differences, and social perceptions of voices. Psychologists will benefit from IVIP through its highlighting of links between advances in eyewitness research and voice identification, and the importance of investigating system variables. The research will further impact the work of criminologists, especially policing researchers, and academic lawyers, in whose sphere the field of earwitness identification has a relatively low profile. In addition, the project will benefit students at each of the three universities represented by the IVIP team, through enhancing research-led teaching. Through identifying future thesis research topics and opportunities, it will also enthuse the next generation of researchers.