Date: January 28, 2001

To: Kevin Finneran, Editor in Chief

Issues in Science and Technology

From: Dr. Leon James (leon@hawaii.edu)

RE: Forum commentary on A. Smiley's article

I agree with Alison Smiley's conclusion that improvements in road safety intended by technological innovations interact with driver behavior in relation to the new capabilities, so that the expected reduction in crashes may not in fact occur. The research she cites clearly indicates that drivers quickly adapt to the new features by increasing risky behaviors like driving faster, driving under poor conditions, or being less attentive and expecting the new technology to take care of them. She also points to drivers' lack of understanding of the limitations of the new assistive systems.

Gerald Wilde has discussed this adaptation phenomenon in his book "Target Risk" which he defines as "the level of risk a person chooses to accept in order to maximize the overall expected benefit from an activity." His concept of "risk homeostasis" provides insights into human risk-taking behavior. People set a "risk target" and adjust their behavior accordingly. Following the introduction of safer equipment or roads, drivers will adjust their actions to the same level of risk as before, which is why the "three E’s" -- enforcement, engineering, and education -- do not necessarily improve road safety and crash statistics. The research he cites on human risk-taking indicates that there are large individual differences in risk homeostasis based upon personality, attitude and lifestyle.

It appears that drivers are less interested in reducing risk than they are in optimizing it. All drivers have a preferred level of risk which they maintain as a target. When the level of risk they perceive in a situation goes down, they will adapt by raising their risky behavior so that the preferred target level remains constant over time. Technological improvements that drivers perceive as lowering the risk is thus followed by a change in behavior that is less cautious and raises the risk to the level prior to the improvement. The data discussed in Smiley's paper conform to this homeostatic explanation.

Her solution calling for better driver understanding of the limitations of the new driver assistive technology may not work for the same reason. While driver training improves skill, it also increases confidence, which in turn lowers the perception of risk, and increases unsafe behavior. What is needed in addition to training is the introduction of increased benefits to safer behavior. When this motive is introduced into the driving equation it acts in opposition to homeostasis and many drivers will respond by inhibiting risky behaviors. As Smiley points out, "it is not enough to develop a better device; one has to know how humans interact with it."

Drivers need training in two areas--understanding how the new technology works, especially its limitations, as Smiley points out, but also understanding the risk compensation effect upon their decisions. The latter understanding, reinforced with positive incentives for safe behavior, will make it more likely that society can benefit from the introduction of the new driver assistive technology.