A trial training based on "Mirroring method" (Koivisto & Mikkonen, 1997) was held at the department of water service of Aomori City in 2002.　In the training program, the participants were asked to evaluate their own driving behavior regarding stopping and confirming of traffic conditions at intersections which has bad visibility using 100 level rating scales.　Then they were shown a videotape in which the driving behavior of their colleagues and other drivers at the intersection near the department were videotaped.　After observing the videotape, participants were asked to discuss on the following theme: Why didn't so many drivers stop completely?　What kind of traffic conflicts will happen when drivers ignore to stop at the intersection?　Are the evaluation on their own behavior previously made were right or not?　Before and after the training, observation of drivers' behavior at the intersection was carried out.　The analysis was focused on the changes of the self-evaluation and driving behavior at the intersection before and after the training.

The main results were as follows;

1. The self-evaluation of participants regarding stopping and confirming of traffic conditions at intersections declined after the training.

2. Although the percentage of drivers who completely stopped at the intersection after the trammg was not changed compared with that before the training, the passing speed at the intersection taken by drivers who belonged to the department became slower compared with ordinary drivers.