Learning to avoid on-coming motorcycles at intersections
Crashes involving drivers turning right across the path of oncoming motorcycles are an ongoing problem for rider safety. This study demonstrated the acquisition of avoidance learning in a decision task similar to that made by drivers when turning across oncoming traffic at an intersection. A sample of 48 participants was divided randomly into two groups ?a Learning Group and a Control Group. In the first phase of the study, participants responded to 200 digital photographs of oncoming traffic at intersections, presented on a computer monitor, by indicating whether they would turn across the traffic in each situation or would choose to wait. Half the photographs included one or more motorcycles in the oncoming traffic, and a long inter-trial interval was used as a negative consequence if the participant?s decision to GO was deemed to be unsafe by the software. The Learning Group was subjected to a contingency between motorcycles and crash risk such that the computer-determined crash risk for a GO decision was higher when a motorcycle was present than when there were no motorcycles. The Control Group?s crash risk was unrelated to the presence or absence of motorcycles, and the average crash risk across all trials was the same for the two groups. The result was consistent with the acquisition of avoidance learning, with Learning Group participants becoming more cautious as the first phase continued ? but only when the image included a motorcycle. A second phase conducted four weeks later demonstrated that the avoidance learning had not decayed, and showed the effect of prior learning in the avoidance learning paradigm. The implications of this demonstration of avoidance learning for the development of training programs are discussed.