Michael DeKay - US grants

Risk ranking has become a popular way for citizens to provide advice to environmental decision makers. If such advice is to reflect citizens' informed judgment, methods must be found to provide participants with both an adequate understanding of the underlying science, and with procedural assistance that helps them to address systematically the complexities of the problem. Under this grant, an interdisciplinary team is developing and demonstrating such methods.

Risk ranking requires both risk analysis and value-based judgments. In the Carnegie Mellon method, experts create "risk summary sheets," which describe each hazard in clear simple language, and in multi-attribute terms. Then, representative groups of lay people, akin to citizen juries, rank the risks using two converging methods. The first is a "holistic" ranking procedure -- after examining the risk summary sheets, participants are guided through a structured procedure for making overall judgments of relative risk. The second method is a simplified "multi-attribute utility" (MAU) ranking procedure, in which participants assign relative weights to each of the risk attributes described in the summary sheets, allowing implicit ranks to be inferred. Groups then use their initial holistic ranking, and the results of the MAU procedure, in order deliberate and construct a final ranking. The procedure uses a combination of group and individual methods. The former attempt to exploit the multiple perspectives and mutual clarification possible with group interaction. The latter are designed to ensure that individuals continue to express (and reflect on) their personal views. Although the basic approach has been built upon theoretical and empirical results from behavioral social science, making it work in a realistic setting has required a substantial amount of experimentation. To that end, the CMU group has created an elaborate experimental testbed involving 22 health and safety risks in the fictional Centerville Middle School.

The current grant is supporting refinement and evaluation of the method using the school test bed. It is also supporting the extension of the method beyond risks to human health and safety to include ecological and other environmental risks with the ultimate goal of developing ways to combine health, safety and environmental risks in a single ranking process.

ABSTRACT for GENERAL PUBLIC'

Better Safe than Sorry:
Precautionary Reasoning, Utility Reversals, and the Creation of Dominance

How do people evaluate the possible outcomes of risky situations? Do these evaluations of possible outcomes guide people's decision-making in such circumstances? According to most theories of "rational" decision making, people's choices should be based on what they think might happen, how desirable or undesirable those possible outcomes are, and how likely they think those outcomes are. In other words, people should think about the possible outcomes before making decisions, so that they can make better decisions.

Sometimes, however, people may "jump to conclusions" before considering all of the possible outcomes. This seems particularly likely when one of the possible outcomes is very bad (for example, when people are killed as the result of a dam failure because a warning was not issued soon enough). Also, it appears that jumping to a conclusion about the preferred course of action (issuing a dam-failure warning, in this case) can affect the way that people think about the other possible outcomes (for example, issuing a warning that leads to an unnecessary evaluation because the dam does not fail). When people think about the possible outcomes only after they decide what to do, the decisions that they make may not be as good as possible. In particular, jumping to conclusions can lead people to indicate that one alternative is always better than the other (for example, that warning "dominates" not warning) when this is not really the case. Our research is designed to assess when and how people evaluate possible outcomes after rather than before decisions, and to assess the role of "precautionary reasoning" in such situations.

We propose to study these issues in three computer-based experiments using members of the general public. Experiment I is designed to discover whether putting a person in the role of decision maker (as opposed to a more uninvolved evaluator) increases the chances of jumping to conclusions and evaluating some outcomes after decisions are made. Experiment 2 is designed to explore the mental processes that people use when they jump to conclusions. Specifically, the experiment will assess how participants seek information about possible outcomes, how participants form their initial preferences among decision alternatives, and how these emerging preferences affect participants' evaluations and use of additional information. Experiment 3 will focus on the effect of uncertainty about the possible outcomes and the effect of ambiguity about the likelihood of the event in question.

The results of this research will broaden the literature on decision-making processes and should have important implications for the assessment of possible outcomes in public and private decision making. In addition, these studies will be among the first to address the relationship between the recently popularized "precautionary principle" and more traditional decision-making methods.

Something is fungible if one part or quantity may be replaced by another equal part or quantity in the satisfaction of an obligation. Oil, grain, and money are usually considered to be fungible. In the context of repeated monetary gambles, the money won in one gamble can replace the money lost in another, dollar for dollar. In other decision contexts, the outcomes of repeated trials may not be fungible. Consider, for example, a medical practice guideline that recommends surgery for a particular ailment. In one instance, successful surgery may extend the patient's life by some number of years. In another instance, unsuccessful surgery may result in the patient's immediate death. The life lost by the second patient cannot, in any real sense, be replaced by an equal quantity of life from the first patient.

Perceived fungibility is important because it may affect how people make decisions. When considering decisions about monetary gambles, people often make different decisions when they can play the gamble many times instead of just once, because they realize that the money won in one gamble can replace the money lost in another. For example, many people would decline a gamble that provides a 50% chance of a $200 gain and a 50% chance of a $100 loss (a single-play decision), but would readily accept a series of 10 or 100 plays of the same gamble (a repeated-play decision). However, when outcomes are perceived to be non-fungible, as in the surgery example above, it is expected that people will not make different decisions when there are multiple patients instead of just one, because good outcomes do not really offset bad outcomes.

The initial experiments in this project will assess the extent to which people perceive outcomes to be fungible in many different decision situations. In subsequent experiments, situations in which outcomes are perceived to be fungible or non-fungible will be used to determine how perceived fungibility affects single-play and repeated-play decisions. In addition to other contributions to the understanding of risky decision making, the planned research will help to establish limits on the appropriateness of aggregating outcomes across multiple decisions, with potentially broad implications for decision making in fields as diverse as marketing, research and development, medicine, law, land-use planning, and free-trade policy.

ABSTRACT


Proposal Number: CTS-0628232
Institution: Carnegie Mellon University
P.I.: H. Scott Matthews, Cliff I. Davidson, Michael L. DeKay

Title: MUSES: Sustainable Consumption, Globalization and Information: Impacts and Opportunities

While there is considerable interest in alternative fuels among policy makers,
automobile manufacturers and fuel providers, substantial economic and technical challenges remain. The adoption of any alternative fuel requires changes throughout the supply chain. Even a liquid fuel such as ethanol that is compatible with current infrastructure can induce dramatic changes, so proper accounting of infrastructure requirements is critical to a complete analysis of material flows and environmental impacts and for a transition to sustainable infrastructure. While much existingresearch address processes and technology for producing alternative fuels, most do not account for the necessary changes in infrastructure. The project proposes to focus our work upon the infrastructure requirements and best technologies for a future of sustainable alternative fuels and to explore different infrastructure options by creating scenarios for each alternative fuel/light duty vehicle option of interest.

The project uses the concept of industrial ecology as our system-wide approach to analyzing the body of issues related to material flows. The project will assess the entire life cycle of various infrastructure components, alternative fuels and vehicle options using a hybrid approach combining process and product specific models coupled with a national economic-input-output life-cycle assessment model developed at Carnegie Mellon. The degree to which any alternative fuel and/or propulsion system is adopted for light-duty vehicles in the United States will depend critically on how well the resulting attributes satisfy preferences in the marketplace. The project will use Life Cycle based analyses in conjunction with consumer choice models of heterogeneous demand with engineering simulations of attainable vehicle attributes to predict market acceptance and likely firm response for alternative fuels and vehicles under several regulatory incentives, technology progress, and infrastructural scenarios. Results from this work will suggest the most promising alternatives based on multifaceted criteria considering environmental, social, and economic issues.

Broader Impacts of the Proposed Research and Integrated Education Program: The project plans a series of educational tasks that are integrated with our research activities. Diverse and talented graduate and undergraduate students will be recruited. It will integrate research results into a secondary school apprentice program, into university courses and into educational modules for use elsewhere. It will use the Internet as a mechanism to disseminate educational materials. More broadly, sustainable energy and infrastructure are critical needs for the United States and for the world. The project states that the world cannot continue to assume that inexpensive petroleum will be available for our transportation needs. Society needs to start on a trajectory that will lead to a sustainable future, both for costs and for overall environmental impact. This project is intended to contribute significantly to that movement with both professional and popular analyses and recommendations.

The project is jointly supported by MUSES funds through the CTS, ECS, and DMR divisions of NSF.