Files
Abstract
The dissertation is concerned with practical modelling issues in stated preference non-market valuation applications when undesirable (unintended) eects are present. Eects that are common to valuation studies are described and classied. Practical approaches and methods to build empirical models in view of unintended eects are reviewed and discussed. Three case studies follow up to illustrate the range of possibilities in empirical treatment of unintended eects. Case Study I implements a panel mixed logit model to account for respondents' heterogeneity in decision-making rules, a sample selection bias, and \warm glow." Case Study II is dedicated to respondents' rationality in a sequential multiple commodity valuation setting. It develops a stochastic model which is consistent with the axioms of reexivity, transitivity, and continuity. Case Study III presents a fuzzy logic system as a
practical solution to the issue of impossibility of exactly quantifying qualitative categories in commodity description and elicitation format options.