Modelling the effects of crime type and evidence on judgments about guilt. (November 2018)

• Record Type: Journal Article
• Authors: Pearson, John; Law, Jonathan; Skene, Jesse; Beskind, Donald; Vidmar, Neil; Ball, David; Malekpour, Artemis; Carter, R.; Skene, J.
• Is Part Of: Nature human behaviour. Volume 2:Number 11(2018)
• Page Start: 856
• Page End: 866
• Publication Date: 2018-11
• Subjects: Human behavior -- Periodicals; Psychology -- Periodicals; Sociology -- Periodicals; 300
• DOI: 10.1038/s41562-018-0451-z
• Languages: English
• ISSNs: 2397-3374

Abstract

Abstract Concerns over wrongful convictions have spurred an increased focus on understanding criminal justice decision-making. This study describes an experimental approach that complements conventional mock-juror experiments and case studies by providing a rapid, high-throughput screen for identifying preconceptions and biases that can influence how jurors and lawyers evaluate evidence in criminal cases. The approach combines an experimental decision task derived from marketing research with statistical modelling to explore how subjects evaluate the strength of the case against a defendant. The results show that, in the absence of explicit information about potential error rates or objective reliability, subjects tend to overweight widely used types of forensic evidence, but give much less weight than expected to a defendant's criminal history. Notably, for mock jurors, the type of crime also biases their confidence in guilt independent of the evidence. This bias is positively correlated with the seriousness of the crime. For practising prosecutors and other lawyers, the crime-type bias is much smaller, yet still correlates with the seriousness of the crime. Using a high-throughput experimental design and statistical modelling, the authors show how jurors and lawyers weigh different types of crime and evidence when assessing the guilt of someone accused of a crime.