The goal of this dissertation is to improve on existing nonmarket valuation techniques by incorporating three sources of information rarely used in the literature; (a) prior information on the distribution of willingness-to-pay (WTP), (b) individual perceptions regarding environmental quality, and (c) contingent behavior data based on hypothetical environmental quality improvements. Consideration of each of these information sources constitutes an essay in the dissertation.;The first essay focuses on incorporating prior information on the distribution of WTP when designing dichotomous choice referendum (DCR) format survey. A key disadvantage of DCR format is that each survey respondent provides limited information about this WTP distribution. Improving the bid design provides a means of improving the efficiency of WTP estimates. Classical and Bayesian approach to bid design are reviewed including simulation based designs and the curve fitting methods. The cost of ignoring parameter uncertainties are investigated and single- and two-stage designs are implemented. Lessons for practitioners are discussed.;The second essay focuses on individual perceptions regarding water quality. Although researchers argue that individual choice decisions are made based on their perceptions, there is relatively little research using perceptions due to the cost of collecting individual perceptions data. Trip behavior and individual water quality assessments are collected from the 2003 Iowa Lakes Survey. Using these two data sets and physical water quality measures provided by ISU limnology lab, the impacts of water quality perceptions and physical measures on recreation demand are investigated. Further, the linkage between physical measures and water quality perceptions are investigated. Welfare measures ignoring and considering individual perceptions are then compared.;The third essay focuses on measuring the impact of hypothetical water quality improvements using three types of trips data (actual trips and anticipated trips under current and hypothetical water quality) collected from the 2004 Iowa Lakes Survey. While pervious contingent behavior research typically lack of sufficient variation in quality attributes, the advantage of this analysis is there is ample variation in the water quality attributes. Individual responses to current and hypothetical water quality are estimated. Welfare measures under these different responses to water quality are then compared.