ABSTRACT
This chapter looks at some of the many proxy variables used by various researchers, discusses the logic behind them, and relates what evidence there is for their validity. Two basic types of proxy variables can be identified; incident and behaviour parameters. The first consists of happenings that are experienced by drivers as pre-stages to accidents that are stopped in time, or incidents where road users have come extremely close to each other in time and space. The second type is those behaviours that are considered dangerous, that is might lead to accidents although they usually do not. Driving simulator variables have very little backing as safety proxies for individual differences, and it is strange that such research can actually be funded, given the fairly large sums involved. Traffic conflicts is, in comparison to the other incident variables, a fairly well discussed and somewhat formalized variable, mainly used by traffic technicians for safety evaluations of places, normally junctions.
