Highway safety is a much broader field now than it used to be. The focus has expanded from trying to prevent crashes by educating people to change their behavior. This approach was too narrow. And it failed because education alone almost never changes driver behavior. For more than 30 years, highway safety professionals have recognized that what's needed to reduce crash losses is a mix of measures aimed at drivers, vehicles, and the roadway environment. Today's vehicle and road safety programs are based largely on research and engineering. But when it comes to changing the behavior of drivers and others on the road, research findings often are ignored. Many programs are based on wishful thinking instead of science. This Status Report summarizes what the research literature tells us works — and what doesn't — to influence driver behavior for the better.
Before there were safety belts or airbags, before vehicles had crumple zones and padded interiors, before guardrails and breakaway signposts were used on highways and shoulders were cleared of roadside hazards, there were "Please Drive Safely" signs. Trying to prevent crashes by educating motorists was the almost exclusive focus of highway safety efforts for half a century, beginning soon after cars began to proliferate on the roads in the early 1900s. The entire idea of reducing the consequences of crashes wasn't a consideration.
A few advocates for a broader approach wanted to include things like installing and using safety belts to reduce deaths and injuries during crashes. These lone voices were ignored by the safety establishment of the time, but they didn't fade away. They continued to grow, which made the existing road safety establishment uncomfortable. This discomfort was apparent in 1961 remarks to the National Safety Congress by the president of General Motors, who criticized the work of "self-styled experts" whose "suggestion that we abandon hope of teaching drivers to avoid traffic accidents and concentrate on designing cars that will make collisions harmless is a perplexing combination of defeatism and wishful thinking."
Science wins out
A few years later, the "self-styled experts" prevailed. Legislation enacted in 1966 gave the federal government its first major responsibility for highway safety. As a direct result, the focus of safety efforts became much broader.
The new approach sought to reduce crash losses by focusing not only on driver behavior and crash prevention but also on reducing injury risk during crashes and mitigating the consequences after crashes by, for example, decreasing the likelihood of fuel leaks that could lead to postcrash fires. Equally important was an unprecedented emphasis on scientific methods to evaluate highway safety programs.
This systematic, scientific approach has saved thousands of lives and prevented countless injuries since implementation in the 1970s. Today's passenger vehicles are much safer. So are roadways. And there has been progress toward improving the behavior of drivers and other road users.
Mix of approaches needed
Because most motor vehicle crashes involve driver error, some people continue to this day to believe that improving driver behavior should be the overriding priority. Claims continue to be made that "getting rid of drunk drivers" or "improving driver skills" is more important than setting speed limits or equipping cars with airbags. Such claims persist despite evidence gathered over the years that many driver-oriented prescriptions are ineffective. Besides, they're easier said than done. Major efforts around the world to "get rid of drunk drivers," for example, haven't succeeded in wiping out the problem of alcohol-impaired driving.
Crash deaths and injuries occur in events ranging from pedestrian impacts to collisions involving tractor-trailers. No single program or approach can have a major effect on such a range of crash types. We need a broad mix of science-based measures aimed at drivers, vehicles, and roadways. There's no reason to prefer measures aimed at drivers over those aimed at the other two. Preference should go to programs shown to be effective.