In Wilde's risk homeostasis theory, it is assumed that people have a t
arget risk which guides their behaviour. The prime purpose of this exp
erimental on-road study was to establish the effect of the need for st
imulation or sensation seeking on this target risk. in addition, the r
elationship between sensation seeking and accident history was explore
d. Based on their score on Zuckerman's Sensation Seeking Scale, 21 mal
e sensation avoiders and 21 male sensation seekers were invited to par
ticipate in a controlled car-driving experiment. To control for the in
fluence of driver experience on both target risk and accident involvem
ent, equal numbers of inexperienced, intermediately experienced and ex
perienced participants were selected. While following another car at a
distance chosen by the participant, target risk was measured by time-
headway - an index of risk taking, and verbal ratings - indices of cog
nitively perceived risk. Accident involvement was measured by reported
past accidents. As expected on the basis of risk homeostasis theory,
sensation avoiders preferred a greater following distance than sensati
on seekers. Unexpectedly, however, they did not perceive this longer f
ollowing distance as less risky. Therefore, it was concluded that sens
ation seekers did not take more risk deliberately. The results of a Po
isson regression analysis gave preliminary evidence for an influence o
f sensation seeking on the relationship between driver experience and
accident involvement. For drivers with an experience of less than appr
oximately 80,000 kilometres it was observed that sensation avoiders ha
d been less liable to accidents than sensation seekers, whereas for dr
ivers who had driven more than 80,000 kilometres, the accident involve
ment of sensation avoiders was greater than that of sensation seekers.
The observation that sensation seekers become involved in accidents e
arlier in their driving career than sensation avoiders is discussed by
referring to the observation that sensation seekers drove in a more r
isky fashion (e.g. shorter time-headway) without perceiving their beha
viour to be more risky. Copyright (C) 1996 Elsevier Science Ltd