As Japan is facing the prospect of becoming a super-aging society in the near future, prevention of traffic accidents caused by the elderly has been one of the most important issues.　However, as few studies have investigated how elderly drivers behave in real traffic conditions, especially expressways, we have few references for planning countermeasures to reduce elderlies’ accidents.　To address this, we performed an experiment with 26 elderly and 14 non-elderly drivers at the Kashiwara toll barrier （TB）, which is known as the most accident-prone spot in the Kansai region.　In our previous study, we analyzed each subject’s gaze position data and evaluated his scanning behavior while checking around for safety.　As a result, we learned that the elderly performed significantly less scanning behavior for safety checking at the TB than the non-elderly.　Based on these results, in order to learn why elderly drivers behave in a risky manner, in this paper we examine the interview data from each subject concerning how he perceived the risk of the surrounding traffic while driving.　The finding of this paper is that the elderly drivers were unable to appropriately perceive increasing risk when approaching the TB, thus leading to risky behavior.

The presence of a peer passenger psychologically affects a driver’s behavior, which may be considered a risk factor for a vehicle accident.　Although previous studies have focused on the peer-passenger effect among middle-aged drivers, little is known about such effect both in middle-aged and young drivers.　This study explored the factor structure of the peer-passenger effect and determinants of this effect in both middle-aged and young drivers.　A total of 282 drivers completed a self-report questionnaire.　Factor analysis revealed five factors of the peer-passenger effect : dissatisfaction with advice, self-uneasiness, security consciousness, and calmness and silence in car.　Multiple regression analyses were conducted to examine how strongly the five demographic variables of gender, age, a license career, and frequencies of driving and of a presence of peer passenger could influence each factor.　Results suggested that a middle-aged driver became irritated more easily than a young driver, when receiving unnecessary advice from a peer passenger.　Younger drivers concentrated on driving so much that they could not afford to pay attention to peer passenger.　Furthermore, the frequency of the presence of a peer passenger could promote the peer-passenger effect in both young and middle-aged drivers.