"Did you find what you were looking for?" The Effect of Context on the Interpretation of Forensic Evidence

Cognitive shortcuts are an essential part of everyday decision-making. They allow us to draw upon contextual information such as our prior experiences, expectations, and knowledge, for quick and efficient decision-making. Although these shortcuts are undeniably useful, they can lead us to make biased and unwarranted decisions. For the most part, this bias will not lead to adverse consequences. In legal and clinical contexts, however, the unwarranted use of contextual information can lead to errors that have far reaching, dangerous, and – in the case of the death penalty – even deadly implications. In this two-part thesis, I explore the potential for contextual information to influence decision-making in two distinct forms of forensic evidence: trace evidence (fingerprints, bitemarks) and projective evidence (children’s drawings). Part One In Experiment 1A (Chapter 3), I hypothesised that emotional context (images and details about a crime) can increase the rate at which people would conclude that two fingerprints match. Undergraduate participants were presented with 96 fingerprint pairs, and were asked if the two prints in each pair were the ‘same’ or ‘different’. Each pair was either a clear match, clear mismatch, or did not provide sufficient information to make a decision (i.e., was ambiguous). Half of the participants were exposed to increasing levels of emotional context (low to high); the remaining participants viewed no contextual information. Participants exposed to context increased their rate of ‘same’ decisions as the emotional context increased, but only on ambiguous pairs. This effect was reduced, however, when participants were able to say that they were ‘unsure’ (Experiment 1B). In Experiment 2 (Chapter 4), I used a similar paradigm to examine whether an emotional context would still influence decision-making when the evidence itself presupposes an emotional context. For this reason, I examined contextual bias in the analysis of bitemarks. Furthermore, by employing a sample of dental and non-dental students, I was able to investigate whether any effects observed differed as a function of expertise. Although there was an effect of contextual information, the findings directly contradicted those observed in Experiment 1. That is, the additional emotional context actually reduced the rate of ‘same’ decisions on ambiguous pairs, regardless of expertise. In Chapter 5, I discuss some of the possible reasons behind these conflicting findings. Part Two In Experiment 3 (Chapter 6), I explored how background information about a child would influence the subsequent interpretation of that child’s drawing. Samples of students, parents, and teachers were presented with three children’s drawings, accompanied by positive, negative, or neutral background information in the form of a brief vignette. From a given checklist of features, participants then selected which were present in the drawing. For half of the participants, the checklists included labels identifying which of the features purportedly suggested ‘good’ and ‘poor’ psychological health. In the labels condition, regardless of sample, negative contextual information made participants select more ‘poor’ features, and fewer ‘good’ features, relative to positive or neutral background information. When participants were not exposed to labels, these effects were reduced. Taken together, the data reported in this thesis suggest that biasing contextual information is a very real concern in the interpretation of both trace and projective evidence. In Chapter 7, I conclude by providing some options for minimising the risk of cognitive bias in these areas