Comparing e-learning and blended learning for threat detection in airport security X-ray screening

Abstract

This study investigated the effectiveness of e-learning and blended learning for airport security officers (screeners) aimed at increasing their detection of novel prohibited items in X-ray images of passenger bags. 80 screeners from a large European airport conducted a simulated X-ray baggage screening task (SXBST). Based on hit- and false alarm rates, d-prime, age, gender and job experience, four equivalent screener groups were created. The first group (EL1) conducted six different e-learning modules. The second group (EL2) repeated the e-learning modules once with a predefined time interval of one week in-between. The third group (BL) participated in an asynchrony blended learning course where all participants first conducted the six e-learning modules followed by classroom training. The control group (CG) did neither conduct e-learning nor participate in the blended learning course. All four groups conducted the SXBST again after these interventions. Statistical analyses of the hit rates for novel prohibited items showed a significant main effect of group (EL1, EL2, BL, CG). Post-hoc tests showed that the increase of the hit rate for EL1 vs. CG, EL2 vs. CG and BL vs. CG was significant. Therefore, all interventions helped to increase the detection of novel prohibited items in X-ray images of passenger bags. Additionally, the hit rate for BL was significantly higher compared to EL1, which underlies the benefit of blended learning as a combination of classroom and online-training. The repetition of the e-learning modules (EL2) led to a marginally significant increase of the hit rate compared to EL1. The difference between EL2 and BL was not significant. This suggests that when training duration is controlled, the advantage of blended learning compared to e-learning vanishes. In addition, training did not impact false alarm rate as the main effect of group was not significant. Analysis of response times showed no speed-accuracy tradeoff but screeners needed significantly more time to analyze target-absent images than target-present images and this effect was independent of group. Overall, these findings suggest that e-learning as well as blended learning are effective to improve screeners' detection of novel prohibited items in X-ray images of passenger bags. Implications for X-ray image interpretation training for screeners are discussed.