Citation

Smith RG, Lovegrove A. Accid. Anal. Prev. 1983; 15(2): 95-104.

Copyright

(Copyright © 1983, Elsevier Publishing)

DOI

unavailable

PMID

unavailable

Abstract

The phenomenon of danger compensation is said to occur where environmental safety improvements installed at a site lead to a decrease in road safety owing to the tendency of drivers to adjust their behaviour elsewhere in the traffic system so as to compensate for the safety measures used. This hypothesis of danger compensation was tested by installing a safety improvement, viz., a STOP sign, at a previously uncontrolled intersection in order to compel drivers to approach it more slowly than they otherwise would have done. Whether this enforced safety led to an increase in risk elsewhere in the traffic system was tested by measuring drivers' approach speeds at the next (test) intersection after the intersection where the STOP sign had been installed. The roadway on which these two intersections were located was in a quiet residential area and both intersections comprised a cross-road with a restricted right sight distance and right-of-way priority. A similar intersection on a parallel road with a separate population of commuters served as a control. Data were collected unobtrusively during morning peak periods. Upon comparing drivers' speeds before the installation of the STOP sign (n = 58) with the speeds immediately after the STOP sign had been installed (n = 53), it was found that the percentage of infrequent commuters exceeding the speed (48 KM/H) at which they could just have stopped their vehicles at the test intersection increased relative to the control from 12 to 25%. This increase was again observed at the four-month follow-up (n = 39). However, the regular commuters showed no such increase at the post-test, and approached the test intersection more cautiously during the follow-up than in the pre-test. An analysis of variance of the infrequent and regular commuters' mean approach speeds during the post-test and follow-up, adjusted for the two groups' pre-test differences, showed that their mean speeds were significantly different, F(1,125) = 7.9, p $0.01. Accordingly, it was concluded that the behaviour of some infrequent commuters was consistent with the danger compensation hypothesis, but that an alternative explanation was needed to account for the response of the regular commuters.