Accurate in presence and inefficient in absence: how the presence of criminal evidence affects legal judgements

Confirmation bias, whereby individuals hold unwarranted confidence in a focal hypothesis, might emerge from a combination of errors at the testing and evaluation stages of hypothesis development. We focused on one such combination: Positive testing, whereby questions are asked about features that are expected to be present and the feature positive effect, whereby the presence of features outweighs the absence of features when evaluating the available evidence. Participants were presented with two criminal scenarios and asked to evaluate the utility of four possible queries and the answers to the one they deemed as the most useful. We found evidence for positive testing and a secondary sensitivity to the Bayesian expected utility of questions. We also found evidence for the feature-positive effect. However, the present evidence elicited more accurate judgements, whereas the insufficient weighing of the absent evidence led to confirmation bias. We discussed theoretical and practical implications for legal judgements.