This quasi-experiment was designed to investigate the effects of the t
ype of driver-side mirror on lane-change accidents. The analysis was b
ased on 407 accidents reported from 1987 to 1992 to Finnish insurance
companies, for vehicles with passenger-side convex miners and one of t
hree types of driver-side mirrors (flat, convex or multiradius). The r
esults showed that there was no difference between the multiradius and
convex mirrors in the frequencies of lane-change accidents to the lef
t. Compared to the flat mirror, the mean effect of the multiradius and
convex mirror was a 22% decrease. However, the statistical strength o
f the data was low; 95% confidence interval ranged from a 51% decrease
to a 25% increase. This result was not related to driver characterist
ics or driving conditions. In conclusion, the multiradius and convex d
river-side mirrors, in comparison to the flat mirror, are more likely
to reduce than increase lane-change accidents. A theoretical implicati
on of this study is that minimizing the blind spot is likely to be mor
e important than providing an undistorted image of objects. From a pra
ctical point of view, the present findings support the use of multirad
ius and convex driver-side mirrors.