Pushing the Public Policy Pendulum

Elise Weaver uses a systems dynamics model to study how and why public policy thresholds shift over time in response to different influences, such as changing social mores and undesirable events.

In the case of homeland security, Fig.1 shows a balanced system.

In Fig. 2, a call for increased security measures results in more people--and, consequently, more innocent people--being searched.

Fig. 3 shows the role of historical incidence, in this case the collective memory of cases when innocent people were searched, and how this influences policy making.

National security. Homeland defense. These catchphrases saturate the news, and since September 11, they are embedded in the collective American psyche. They beg an important question: What price safety? How many personal freedoms should we relinquish to ensure malcontents don't slip through the cracks?

The debate generates sensational headlines, but there's a more productive way to study the problem using a model developed by professor Elise Weaver, a psychologist in WPI's Social Sciences and Policy Studies Department. Weaver, along with George Richardson of the University at Albany, SUNY, is studying how policy thresholds change over time in response to public concerns. Using a systems dynamics model--and building on Kenneth Hammond's work in this field--Weaver illustrates the cyclical nature of policy making, in fact, any kind of decision making where thresholds (also known as cutoff points) are set.

"Once they understand the cycling, decision makers could spend less time arguing the opposite sides of the issue and more time trying to strike the appropriate balance," says Weaver. "Using this model, we could shift focus from 'right vs. wrong' to how can we make a better tool for reducing the number of false positive and false negative outcomes."

To test her model, Weaver examines the case of a police officer deciding to initiate a search. A Taylor-Russell diagram (Fig. 1) shows the cutoff point. In the case of the police officer, it is the officer's cutoff for the perceived level of evidence that something is awry. The officer will use clues, such as a suspect's nervousness, to help judge whether or not to conduct a search. The diagram shows that given a set threshold, there will always be a certain number of false negatives (guilty people that get away) and false positives (innocent people that get searched). The trick is finding the proverbial happy medium.

But what happens when the policy receives pressure from one side or the other? In this model, what happens when there is a heightened sense of vulnerability, and a public outcry for more searches? Policy makers react, and the police officers are given new guidelines--a mandate to execute more searches. Consequently, as the model shows (Fig. 2), more innocent people get searched. It's not long before advocates for protecting personal freedoms raise their voices. If they are loud enough, policy makers shift the cutoff point again.

"In some situations the model illustrates a wild oscillation," says Weaver. It can be used to study all types of decisions, from SAT score cutoffs for college admissions to drug approval using medical tests and indices. When there is a cutoff point, some number of smart kids won't get admitted to a college while others who are not meant for college will get in; potentially dangerous drugs will gain approval while a number of useful drugs will have approval delayed.

Weaver analyzes three alternative models to show how structural characteristics of a social system affect its behavior over time. The "grudge" model represents the role of memory for past cases contributing to the oscillation (Fig. 3); a second charts the effect of delayed responsiveness on the part of policy makers; a third looks at what happens when people change their opinions in the wake of undesirable events. In doing so, Weaver pioneers the university's efforts in this area; WPI is the only school in the world that teaches system dynamics to undergraduates.

--CC

Pushing the Public Policy Pendulum

Related Links