Using images to present stated preference information: an application to the built environment

Stated preference techniques have traditionally been used to measure preferences and trade-offs in areas such as marketing, transport, health and environmental economics. With some exceptions these techniques have not typically been applied to public space within the built environment, nor have they attempted to incorporate computer generated visual imagery or the Internet. This thesis addresses these omissions in a number of ways. The thesis tests whether compute generated images can be used successfully to communicate information to respondents. It also examines the advantages and disadvantages of using different media including the Internet, PCs and paper questionnaires to convey and collect survey data. The built environment was selected as the context for the thesis, as urban public space is an important factor in urban quality of life and the 'liveability' of European cities. Choice experiments and contingent rating are the stated preference techniques used to determine trade-offs between attributes. The choice experiment investigated how changes to the configuration of physical attributes affected preferences for the layout of a town square. Respondents chose their most preferred layout within each choice set. The survey was conducted online and on a stand-alone PC and it was possible to determine significant attributes from the data collected. In terms of the questionnaire itself, fewer technical problems were experienced with the stand-alone PC version although both methods successfully collected meaningful data. The paper based contingent rating study addressed a number of significant issues that arose from the choice experiment; mainly the inclusion of non-physical attributes in the experimental design, such as weather, season and time of day. The aim of this study was to investigate how attributes affected perceptions of safety and attractiveness, and whether respondents would visit a greenspace under the conditions displayed. Time of day had the biggest influence on responses, with the sites themselves also providing to be a major factor. The thesis demonstrates that stated preference questionnaires are suitable for use within the built environment, with many of the attributes having a significant influence on responses. The use of images to provide respondents with complex sets of information has proved successful. It is important the images are realistic, which can best be achieved using a combination of physical and non-physical attributes. In terms of the delivery media, the PC questionnaire was found to be the most successful within this thesis. However, issues such as time, cost, sample population and technology must all be taken into consideration when developing future methodologies.