Abstract
Some of the psychological research into driving is examined, and especially Quenault's typology of drivers. Attention is drawn to the need for general theory of driving behaviour based on "homogeneous" groups of subjects drawn from the general population, rather than from groups that are differentiated on the basis of chance-occurrences such as "accidents". A modified typology is proposed on theoretical grounds to facilitate the objective testing of large numbers of relatively non- specific behavioural hypotheses.