he Miss Error Rate for Threat Items in X-ray Screening is Reduced by Enhancing Individual Attention

X-ray screening plays an essential role in ensuring the safety of public transportation and people’s lives. However, the low prevalence effect (LPE) in X-ray screening can lead to a high miss error rate for threat items. Current research has primarily focused on reducing or avoiding the LPE by examining judgment criteria, motor response errors, and premature search termination, but the results have not been very satisfactory. This study aimed to explore effective ways to reduce the miss error rate of threat items in low prevalence conditions from the perspective of "failures of attention". Through four experiments, this study systematically investigated how to enhance individual’s attention to reduce the miss error rate for threat items in X-ray screening. The results showed that under low prevalence conditions, participants produced a miss error rate of about 20% (Experiment 1) in both the attentional boost and no attentional boost conditions. However, when simulated machine learning (with 80% accuracy) was combined with attentional boost, the miss error rate dropped to only 10.44% (Experiment 2). This outcome is attributed to the joint contribution of the attentional boost and the attention allocation tendency caused by instructions. When only attentional boost manipulation was performed (Experiment 3) or only instructions were given (Experiment 4) in the simulated machine learning situation, the miss error rate was about 15%. This study suggests that combining simulated machine learning with the attentional boost could improve the accuracy of X-ray screening in real environments.