Improvement of hazard prediction accuracy after training: Moderation effect of driving self-efficacy and road safety attitudes of learner drivers

•Hazard prediction skill improvement training program was effective for learner drivers who are men.•The training effect was not obtained in the women group.•Driving self-efficacy and road safety attitudes did not moderate the training effect.•On-line hazard prediction training should deal with specific technical challenges. This paper explores effectiveness testing of a newly developed online hazard prediction training program for learner drivers. It also aims to evaluate the role of participants’ driving self-efficacy and road safety attitudes on program effectiveness. Eighty learner drivers who participated in standard driving training took part in the study. Forty-two of them were assigned to the experimental group and thirty-eight – to the control group. The experimental group was exposed to 2 × 45 min. intervention sessions delivered by a trained driver instructor in an online format. The scores of driving self-efficacy and road safety attitudes were assessed by self-reported measures at pre-test. The effectiveness of the program was tested by the improvement of Hazard prediction test accuracy from pre-test to post-test. Results reveal that hazard prediction skills improved only slightly in both experimental and control groups. The training effect occurs only for learner drivers who are men (effect size – d = 0.73), whereas no significant effect is found for women participants. The difference between pre-test and post-test variable does not correlate with either driving self-efficacy or road safety attitudes, therefore, the hypothesis of moderation effect is rejected. The new online program for hazard prediction skills improvement shows promising results for learner drivers, especially among men. The authors of the study advise a replication of the gender effect on the program effectiveness, and a gender-tailored alteration for the program.