Considering forensic science: juror decision making and unvalidated identification evidence

Abstract

Rapid scientific advances mean that new techniques and areas of research are being used by crime labs to test forensic evidence, but as innovations grow, so does fear that invalid science will make its way to the courtroom. If jurors and judges are not informed of the threats to validity that are associated with identification evidence they are at risk of overestimating the reliability of that evidence. The overarching goal of this thesis was to investigate whether scientifically informed opposing expert testimony or cross-examination will educate jurors about unreliable forensic science, and whether there are individual differences that will affect the perception of forensic evidence. Study one investigated whether opposing expert testimony could educate jurors about anthropometric facial comparison evidence. In addition, participants’ scores on measures of epistemological sophistication and argument skill were used to test for direct effects on verdict, and indirect effects through ratings on a measure of methodological reliability. Path analysis did not show support for relationship for the two individual difference measures. Opposing expert testimony was able to reduce ratings of the reliability of the forensic evidence, indicating that presenting participants with criticisms of unreliable forensic evidence is capable of informing jurors of limitations. Study two aimed to replicate the main findings of the first study while testing a different measure of individual difference: bias towards forensic evidence, as captured by the Forensic Evidence Evaluation Bias Scale (FEEBS, Smith & Bull, 2012, 2014). Opposing expert testimony reduced reliability, and scores on the pro-prosecution subscale of the FEEBS led to higher ratings of reliability, indicating that when the participants were predisposed to see forensic evidence as highly trustworthy and conclusive they were more likely to convict. Qualitative analysis of responses justifying verdict choice showed that opposing expert testimony was informative, but that many participants struggled with understanding scientific methodology and had unreasonable expectations about forensic science. Study three tested whether scientifically-informed cross-examination would lead to reduced reliability. Three types of forensic identification sciences were used: anthropometric facial comparison, fingerprint, and voice identification. Participants read through expert testimony regarding one of the three types of evidence, and then either scientifically-informed cross-examination, or questions that focussed on the qualifications and experience of the expert. Multi-group analyses and individual path analyses were conducted. Only the relationship between examination type and evidence type was different between groups, and scientifically-informed cross-examination did not affect ratings of reliability. Scores on the FEEBS affect neither reliability nor verdict. This suggests that differences in testimony and either the origin, or complexity, of criticisms towards evidence may have a large impact on verdict. This thesis contributes to furthering our understanding of juror decision making regarding unreliable forensic evidence as it has demonstrated that perception of reliability, even if based on substantially biased or incorrect reasoning, will have the largest impact on verdict choice. The findings will be useful to researchers looking into the best ways of educating jurors and judges, as well as those calling for validation studies of forensic sciences.