Southeastern Geographer Vol. XXXVIII, No. 2, November 1998, pp. 112-124 ACCEPTING A LOCALLY UNWANTED LAND USE: ILLUMINATING THE DECISION PROCESS WITH A NEURAL NETWORK1 Jerry T. Mitchell The siting of facilities deemed hazardous to the environment has become increasingly more diffi cult. As such, these facilities have become locally unwanted land uses (LULUs). Proponents or opponents of the siting of LULUs may find their position pivots on certain key pieces of informa tion; as this knowledge of the LULU varies, so may their disposition toward accepting or refusing the new facility. The extent of risk from these facilities is not often fully understood and this uncer tainty has led to an inability to gain consensus about solutions. The public often finds itself having to make decisions based upon little information about the risks involved with a new facility. This research is concerned with how the perceived benefits and burdens of a hazardous waste facility are weighted when making a siting decision. Specifically, a neural network is used to understand how people weigh the risks and benefits of a potentially hazardous facility. Although various questions remain given the population tested, the results show that pollution and geographic factors are sig nificant in determining the acceptance of the proposed land use. The siting of facilities deemed to be hazardous to the environment or human health has become increasingly more difficult. For many reasons, these facilities are locally unwanted land uses, or LULU’s (Popper, 1983; Bourke, 1994). The situation is often defined in terms of the popular NIMBY syndrome, an acronym standing for “not in my backyard,” where zoning and other legal means are used to resist the location of unwanted land uses (Lindell and Earle, 1983; Greenberg and Anderson, 1984; Meyer, 1995; Platt, 1996). While one group appeals for the acceptance of the land use (usually a developer or the government), local opposition groups have often been successful in delaying or denying its siting (Himmelberger et al., 1991). Strong proponents or opponents of the siting of LULU’s may find their position piv ots on certain key pieces of information. As this knowledge of the LULU varies, so may their disposition toward accepting or refusing the new use, facility, or activity. This paper is concerned with how the varying perceived benefits and burdens of a hazardous facility are weighted when making a siting decision. Specifically, a backpropagation neural network is used to model the cognitive processes employed by people as they deliberate over the acceptability of a fictitious hazardous waste treat ment facility using six informational statements regarding its risks and benefits. ASSESSING HAZARD RISK. Why are people concerned about some risks and not others? Why do communities fight so hard to keep some land uses out while ignoring others? The fear associated with industrial facilities is often strong despite Dr. Mitchell is a post-doctoral researcher in the Hazards Research Lab at the Department o fGeography, University o fSouth Carolina, Columbia, South Carolina 29208. VOL. X XX VIII, No. 2 113 evidence that shows high levels of safety, while other hazards, such as automobile travel, are grossly underestimated (Kishchuk, 1987). The answer to these two ques tions lies in how residents define the acceptable level of risk and benefits they are willing to bear. The degree of acceptability may be determined, in part, by the ability, or lack thereof, of the residents to control the hazard or find a safer alternative (Stallen and Thomas, 1988). Clarke (1988) suggested that people make risk assessments from biased information, vary their assessments depending on social position such as cit izen or expert, and that they are more tolerant of risks where exposure is thought to be voluntary. Additionally, prior research has shown that people define risk, bene fits, and acceptability in a complex, multidimensional manner (Gardner and Gould, 1989). This study demonstrates that factors such as the perception of positive bene fits, negative risks, opinions of economic health, and location form the complex structure in which people define their risk level, and notion of acceptability, from the siting of a hazardous waste treatment facility. Furthermore, the use of a neural network as a potential tool for policy making is explored by determining...