Apart from helping to establish links between
behavioural characteristics and risk of involvement
in different types of accident, understanding those
links should also prove useful for development of
countermeasures. It is one thing to be able to inform
drivers that they run an increased risk of accident,
but common sense would suggest that it is much
more helpful to know specifically which types of
accident various types of driver might be susceptible
to, in order to target training, education and enforcement.
A system for classifying accidents based on
brief written reports by drivers has recently been
devised (West 1993). The system involves assigning
the accident to one of several “scripts” based on
the physical movement of the vehicle(s) involved,
and assigning an “active” versus “passive” role to
the reporting driver based on the part played by his
or her vehicle. While there is no obvious way in
which the drivers’ descriptions could be checked
by objective recording of the accidents, there are
several features of the classification system that suggest
that it is reasonably valid. First of all, the categories
involved are sufficiently distinct that it would
require complete fabrication of accident details by
drivers to move an accident from its true category
to another (e.g. from a shunt to a right-of-way violation).
If drivers were unwilling to provide at least an
approximate account of their accidents, one would
imagine that they would simply leave the description
blank or, more likely, not return the questionnaire.
If drivers were tending to bias their accounts in order
to minimise their own role in the accident, one would
expect to end up with far more “passive” than “active”
accidents. (The distinction between passive
and active is based primarily on whether the reporting
driver would be likely to be held responsible in
law.) In fact, this is not what was observed. Third,
to the extent that the classification system contains
random errors of classification, there would be a
failure to observe relationships between demographic
characteristics and involvement in specific
accident types. In fact, it has been shown that young
males are more at risk of certain types of accidents,
defined using this classification system than others
(West 1993).