Understanding Performance and Cognitive Efficiency when Training for X-Ray Security Screening

Abstract

We describe an experiment designed to understand the X-ray security screener task via investigation of how training environment and content influence perceptual learning. We examined both perceptual discrimination and the presence/absence of clutter during training and how this impacted performance. Overall, the data show that performance was generally better when there were clutter items in the training images. We also examined the diagnosticity of a measure of cognitive efficiency, a combinatory metric that simultaneously considers test performance and workload. In terms of cognitive efficiency, participants who trained in the difficult discrimination with clutter present experienced lower workload during the test relative to their actual performance. The discussion centers on how improved analytical techniques are better able to diagnose the relative effectiveness of training interventions.