Skip to content

Using mixture models to examine group difference among jurors: an illustration involving the perceived strength of forensic science evidence

Journal: Law, Probability & Risk
Published: 2021
Primary Author: Naomi Kaplan-Damary
Secondary Authors: William C. Thompson, Rebecca Hofstein Grady, Hal S. Stern

The way in which jurors perceive reports of forensic evidence is of critical importance, especially in cases of forensic identification evidence that require examiners to compare items and assess whether they originate from a common source. The current study discusses methods for studying group differences among mock jurors and illustrates them using a reanalysis of data regarding lay perceptions of forensic science evidence. Conventional approaches that consider subpopulations defined a priori are compared with mixture models that infer group structure from the data, allowing detection of subgroups that cohere in unexpected ways. Mixture models allow researchers to determine whether a population comprises subpopulations that respond to evidence differently and then to consider how those subpopulations might be characterized. The reanalysis reported here shows that mixture models can enhance understanding of lay perceptions of an important type of forensic science evidence (DNA and fingerprint comparisons), providing insight into how the perceived strength of that evidence varies as a function of the language forensic experts use to describe their findings. This novel application of mixture models illustrates how such models can be used, more generally, to explore the importance of juror characteristics in jury decision making.

Related Resources

Autopsy of a Crime Lab: Addressing the Sources of Error in Forensics

Autopsy of a Crime Lab: Addressing the Sources of Error in Forensics

Keynote presentation:  Autopsy of a Crime Lab book and overview of IMPL 1 Project Area
Ensemble SLRs for Forensic Evidence Comparison

Ensemble SLRs for Forensic Evidence Comparison

This CSAFE webinar was held on August 25, 2022. Presenter: Danica Ommen Assistant Professor – Department of Statistics, Iowa State University Presentation Description: To strengthen the statistical foundations of forensic…
Does Image Editing Improve the Quality of Latent Prints? An Analysis of Image‐Enhancement Techniques in One Crime Laboratory

Does Image Editing Improve the Quality of Latent Prints? An Analysis of Image‐Enhancement Techniques in One Crime Laboratory

The following was presented at the 74th Annual Scientific Conference of the American Academy of Forensic Sciences (AAFS), Seattle, Washington, February 21-25, 2022
What types of information can and do latent print examiners review? A survey of practicing examiners

What types of information can and do latent print examiners review? A survey of practicing examiners

The following was presented at the 74th Annual Scientific Conference of the American Academy of Forensic Sciences (AAFS), Seattle, Washington, February 21-25, 2022.