Understanding and modelling current and future choices : (collected papers)

University of Technology, Sydney. Faculty of Business.NO FULL TEXT AVAILABLE. This thesis contains 3rd party copyright material. The hardcopy may be available for consultation at the UTS Library. ----- Everyday people make choices. Some are trivial (eg, which candy to buy), while others are consequential (eg, what automobile to buy). The science of modelling choices was pioneered by L.L. Thurstone (Psychological Review, 1927), and his work on choices in pairs was extended by D. McFadden, who won the Nobel Prize in Economics (2000) for this body of work. During the 1970s I studied decision-making and choice behaviour using the paradigm of Information Integration Theory pioneered by Norman H. Anderson (UCSD, psychology). I came to realise that studying simple decisions and observing these decisions using rating scales had serious limitations. Consequently, I took advantage of developing work in discrete multivariate statistics on the analysis of contingency tables, McFadden's relatively new work on probabilistic discrete choice models, work on the design of statistical experiments and Information Integration Theory to develop a way to design discrete multivariate statistical experiments that integrated these research streams, culminating in Louviere and Woodworth (JI of Marketing Research, 1983). This new approach had several distinct advantages. For the first time, one could study choices instead of indirectly observing ratings of choice options. One could model these choices with newly emerging statistical choice models, and one could validate the models by predicting choices in real markets. Much of my work since that time has focused on new developments and refinements in this field. This body of work has led to worldwide adoption of discrete choice experiments as the way to study choices in many contexts, including non-market traded goods frequently of interest in environmental and resource economics and health economics. In the late 1980s I pioneered a new measurement approach now known as "Best-Worst Scaling" that has been widely adopted by academics and practitioners in many fields. My work since that time has focused on refining this new measurement and modelling approach, including developing new ways to model the choices of single individuals. The capability of modelling single individuals provides the opportunity to solve a wide array of problems within organisations, including business-to-business problems. One now can understand and explain how individual decision-makers are making choices. More recently, my work also has focused on potential experimental artifacts associated with various new ways of designing choice experiments and studying the stability of choice experiment outcomes over time. [PRODUCTION NOTE: Collected papers may be consulted at UTS Library