A stated preference (SP) ranking experiment was designed to estimate the willingness-to-pay (WTP) for reducing the level of noise in a group-based residential location context. Important issues to be discussed are the proper definition of context and the variable metric for the noise attribute. A sample of 300 households living in flats on medium to high rise buildings, near important streets of the city, were asked to rank a maximum of two sets of six options arising from a careful design for four attributes (travel time to work or study, noise level, rent or mortgage value of the apartment, and orientation with respect to the sun); the first three attributes have three levels of variation and the last only two. The experimental design allows for the estimation of main effects and several interesting interactions. An important part of the discussion in the paper is devoted to the experimental design, in an effort to make this rather complex subject more easily understood by the normal practitioner. Multinomial (i.e. after applying the rank explosion method) and mixed logit models (which allow to treat the six or twelve observations proceeding from each individual as correlated) will be estimated for various stratifications of the data (income, location) and from these results subjective values of time (SVT) and WTP for reductions in the level of noise are derived. The SVT will be compared with those obtained in previous experiments, also based on a residential location choice context, which have successfully allowed us to estimate WTP for reductions in environmental pollution and improvements in accessibility. For the covering abstract see ITRD E124693.
Abstract