Solving Real-World Problems

David J. Weiss

We all hope that our work will have practical value, but most researchers’ efforts have little impact on human welfare. In order for science to solve an important real-world problem to be solved, three distinct steps are necessary. The ordering of the steps is crucial.

For an example of a solved problem, I will consider the conquest of polio. As is the case for many real-world problems, contributions from more than one discipline are needed. It is probably unrealistic to expect a single researcher to have sufficiently broad expertise to accomplish all three steps.

The first step is to identify the underlying mechanism, the root cause of the problem. People had to understand that the disease was not a curse, or a punishment from God, but was caused by a virus.

Once the cause was identified, someone had to figure out how to prevent the disease. This is the second step, finding a solution. Here the model of a vaccine had already been established, so the trick was to isolate the virus and proceed following established guidelines.

The third step was to get people to accept the vaccine and distribute it to the people who need it. The implementation step involves public health folks, who need to generate confidence in the vaccine and make it available in a practical way. Also relevant are behavioral issues, such as prompting parents to get their kids vaccinated.

Now consider a societal problem that seems to be largely behavioral, namely violence in children. What is the underlying mechanism? I’m not sure, but modeling is a candidate. So one might do studies that examine whether children who are exposed to violence are more likely to engage in violent behavior than children who are not. This would be a difficult research topic, because (at least in the U.S.) exposure to violence is the norm. One might try to contrast the impact of violence seen on the screen with violence observed directly. Or perhaps exposure to violence committed by a family member has more impact than exposure to violence committed by strangers.

What is the solution? Suppose we determined that seeing violence committed by a parent was a strong predictor. We might propose that people be examined (by a psychologist!) before being allowed to have children. Alternatively, we might propose that children be automatically removed from a home in which a parent has committed a violent act. How could we examine whether such proposals would be effective? Even if we thought these ideas were worth testing, there would be legal and ethical roadblocks.

Suppose that someone were clever enough to devise a way to test a solution, and showed that it worked. The implementation issue would still be a powerful obstacle. In the case of violence, we know that access to a gun is a primary element in at least the severity, and probably in the likelihood, of committing a violent act. The simplest way to demonstrate that fact is via the natural experiment of comparing homicide rates among adolescents across countries; the U.S. is way in front of other industrialized societies in homicide rates and gun availability. Yet, elimination of guns in the U. S. has proved impossible.

Many interesting problems combine elements from medicine and psychology. The implication of tobacco in a host of medical problems is one of medicine’s great triumphs of the twentieth century. The mechanisms are understood well enough to know that the solution is to end smoking. But implementation is slow, for several interesting reasons. Will psychologists or economists be the ones who ultimately end the reign of tobacco?

As diseases associated with lifestyle become better understood, how to bring about implementation will present major opportunities for behavioral researchers. The grand experiment conducted by the United States, Prohibition, has been acknowledged as a well-intentioned failure. As researchers achieve more solutions, insights regarding implementation will be increasingly crucial.