Updated: May 8, 2019
‘Cascade bias’ is a relatively new term on the block that we introduced in the literature in 2017. It has great significance for how we understand the inferences that are made in forensic reconstructions, and raises complex questions about how we present those inferences and the conclusions that we make from them.
This is not a new phenomenon, but only more recently are studies being conducted to establish when cascade bias can occur and to what degree. Research by Claire van den Eeden and colleagues in the Netherlands demonstrated that prior information concerning the likely cause of death did influence the ultimate interpretation of the crime scene by the experienced crime scene investigators who took part in the study.
A study from us led by my colleague Dr Sherry Nakhaeizadeh and published last year, demonstrated that exposure to extraneous contextual information at the crime scene has an impact on the subsequent sex assessment of an excavated skeleton made by the participants back in the laboratory. Male skeletons were buried in different types of clothing (some in neutral clothing, others in more clearly female clothing). Of the participants that excavated the male skeleton dressed in female clothing only 1 participant (of 11) in the subsequent sex assessment said that the skeleton was male. In the group that excavated the male skeleton that was in neutral clothing, 7 (of 10) participants said that the same skeleton was male. We also had a third group who just assessed the skeleton in the lab (with no extraneous clothing information) and all 15 (of 15) participants assessed the male skeleton to be male. In that paper we were able to demonstrate the existence and out-workings of ‘cascade bias’ (first introduced by our paper led by Dr Itiel Dror
Bias has been demonstrated to be an important and broad issue to consider in forensic reconstruction approaches, especially in conditions of uncertainty. And it is an issue that impacts every type of forensic science evidence, as it impacts every aspect of any decision-making. There are many ways of mitigating its impact, but it is perhaps unrealistic to think that it can be illuminated entirely. The human actors in the forensic reconstruction process will always be absolutely crucial. But there are opportunities to consider how we present our inferences and conclusions in a way that adequately and transparently communicates the uncertainty and potential risks inherent to that conclusion. We need to take them.