<![CDATA[Road safety remains a major public health concern, especially in low and middle-income countries, such as Indonesia; it is closely linked to unsafe driving behavior, along with poor awareness of safety and erroneous perception of risk. Although various training programs are already established, there is limited empirical evidence to identify which configuration produces the greatest safety improvement. This study explicitly aims at analyzing the effects of training characteristics, including the number of sessions, total duration, and training delivery format, on transportation safety outcomes of university students. The research had an experimental pre-post design, with 32 participants. Safety was measured along different dimensions, namely safety attitude, positive safety behaviour, negative safety behaviour, safety violations, and safety knowledge. Safety improvement was defined as the difference in total safety scores post-training and pre-training. Multiple linear regression revealed that total duration, the number of sessions, and training delivery significantly affect the improvement in safety at the 0.05 level. To find a suitable solution, a Genetic Algorithm was used to optimize the solution. The results show that 8 sessions of 1 hour are better than 1 session lasting 8 hours. This is in line with more traditional learning theory that repeated presentations over a given period are more effective for learning and behavioral response. Also, the gap between the 8x1 sessions offered a period over which adjustments could be made. Overall, the results highlight the value of training delivery formats on transportation safety improvements.]]>
