ABSTRACT
This chapter investigates the evidence concerning criteria of culpability. It discusses what testable consequences a correct criterion of culpability will lead to, and refers to the first research into the validity of criteria, and show that erroneous coding does exist. The general problem concerning culpability is that there is no accepted criterion that can be applied to any crash. Single crashes and shunting would seem to be accepted as this driver's fault, but thereafter, the agreement seems to stop. It can be hypothesized that the probability of any driver being involved in a crash without having any responsibility is the same because an at-fault driver does not intentionally select whom to strike. A short time period will also insert a small amount of artificial correlation between culpable and non-culpable crashes. Therefore, it can be recommended that samples with a mean number of accidents above two should preferably be used within this type of research.
