Effects of confidence level and hazard type on the visual search patterns and hazard response times of young drivers.

Objective: While numerous studies have reported that overconfidence affects young drivers' crash risk, direct comparisons of hazard perception differences among overconfident, underconfident young drivers and their peers with relatively accurate self-rated confidence remain limited. This study addressed this gap by exploring the effects of hazard type and confidence level on the hazard perception of young drivers.

Methods: A total of 72 young drivers aged 18-25 years agreed to participate in this study. A 2 (hazard type: environmental prediction hazards/EP, behavioral prediction hazards/BP) × 4 (driver group: overconfident, very confident, moderately confident, underconfident) mixed experimental design was adopted. Twelve video clips with BP hazards and 12 with EP hazards were presented to the four groups of drivers. Response time and eye movement were recorded.

Results: Underconfident drivers had longer response times than very confident and moderately confident drivers did, regardless of hazard types. Overconfident drivers took longer to fixate the AOIs that contained EP hazards and responded slower to EP hazards than moderately confident drivers did. Although overconfident drivers responded slower to BP hazards compared to very confident and moderately confident drivers did, all three groups took similar times to fixate the AOIs that contained BP hazards. Additionally, compared to very confident drivers, overconfident drivers had a higher no-response rate and fewer fixations on the hazards.

Conclusions: These findings indicate that moderately confident drivers outperformed both overconfident and underconfident drivers in response times, highlighting how confidence level influences hazard perception and response efficiency depending on hazard type. The results provide valuable insights for hazard perception training programs tailored to young drivers, emphasizing the need to address both overconfidence and underconfidence in driver education.