City Research Online

Abstract

Previous research suggests that earwitness identification is flawed due to suggestive lineup techniques, poor witness memory, and challenges presented during and after the initial voice exposure. Earwitness evidence presented during court testimony is given substantial weight by jurors (Semmler, Brewer, & Douglass, 2012). The reliability of earwitness evidence is an understudied issue compared to eyewitness identification and warrants further exploration. To address the disparity in research, this thesis explored: (1) how well witnesses remember voices, (2) does speaker identification accuracy vary with the gender of the speaker, (3) does speaker identification accuracy vary when the witness is presented with a new voice or new phrase, (4) does speaker familiarity or confidence ratings predict speaker identification accuracy, and (5) how well witnesses recall details of a crime.

The overall aim of this thesis was to investigate the boundary condition for accurate recognition of voices and recall of verbal content. This was addressed in six experiments. The six experiments focused specifically on speaker identification accuracy. In Experiment 1, we evaluated how well witnesses remembered words and voices. In Experiments 3 and 4, we assessed whether exposure duration (Exp. 3) and source confusion (Exp. 4) would impact the encoding of written and auditory statements. In Experiment 5, we did not find that participants’ familiarity or confidence ratings predicted speaker identification accuracy. In Experiment 2, we analyzed how well participants recognized voices associated with a criminal incident and found that overall, speaker identification accuracy was poor. The information that we gathered from our research has shown that memory for speaker identification is poor even when tested within controlled laboratory conditions.

Finally, to further contribute to reducing earwitness identification inaccuracies, we created a mobile application for recording and reporting important information. In Experiment 6, we reasoned that capturing crime-related information in real-time or immediately after an event would help to reduce memory errors that tend to increase with the passage of time (Yarmey and Matthys, 1992; Öhman, Eriksson, & Granhag, 2013). Such a tool will hopefully increase public safety and reduce eyewitness errors by serving as a technological corroborator.

Publication Type:	Thesis (Doctoral)
Subjects:	B Philosophy. Psychology. Religion > BF Psychology
Departments:	Doctoral Theses School of Health & Psychological Sciences > Psychology School of Health & Psychological Sciences > School of Health & Psychological Sciences Doctoral Theses

Export

Downloads