Economics ETDs

Author

Jee Hwang

Publication Date

9-12-2014

Abstract

In this dissertation, the value of wildlife watching recreation is used as the context for exploring different modeling approaches available for combining data from nonmarket valuation surveys such as the travel cost method (TCM) and contingent valuation method (CVM). Another topic explored in this dissertation is the issue of nonresponse in nonmarket valuation surveys. The results of this dissertation are useful for wildlife and land managers interested in obtaining theoretically consistent values of wildlife watching recreation using a combination of TCM and CVM data. The practical solutions to nonresponse in nonmarket valuation surveys are useful for researchers who wish to implement relatively small-scale surveys due to limited budgets and are concerned about maintaining the sample size. In the first segment of this dissertation (chapter 2), differences in the preference structure of wildlife watchers is studied by estimating a finite mixture (FM) model of wildlife watching recreation using data from a national survey. This model is unique in that the multiple imputation (MI) method was applied to the FM model framework to address missing travel cost data. When compared to a truncated FM model, it was found that using the MI approach led to different consumer surplus estimates, but also greater performance in terms of goodness of fit. Combining TCM and CVM data has many advantages for recreation demand modeling in that it helps overcome some of the shortfalls where only one type of data is used. An area less explored is the use of a utility-consistent modeling framework that unifies the TCM and CVM components to arrive at a single estimate of welfare. However, the scenarios reflected in CVM questions may not always have a TCM counterpart, therefore leading to a discrepancy in the corresponding welfare measures. Chapter 3 presents a utility-consistent joint model of wildlife watching recreation where the scenarios and the welfare measures from the TCM and CVM components are the same. An advantage of this model is that welfare effects can be interpreted in terms of net benefits or willingness to pay. Nonresponse in CVM surveys can lead to a loss in statistical efficiency and bias. In the final segment of this dissertation (chapter 4), a simple recoding procedure is introduced to address missing follow-up responses from a double-bounded dichotomous choice CVM survey. A Monte Carlo simulation was used to examine the performance of recoding compared to list-wise deletion across three scenarios that included randomly arising nonresponse and systematically arising nonresponse. The results of the simulation showed recoding led to lower losses in statistical efficiency.

Degree Name

Economics

Level of Degree

Doctoral

Department Name

Department of Economics

First Committee Member (Chair)

Thacher, Jennifer

Second Committee Member

Horn, Brady

Third Committee Member

Hansen, Wendy

Language

English

Keywords

Wildlife Watching Recreation, Nonmarket Valuation, Travel Cost Method, Contingent Valuation Method, Combining Revealed Preference and Stated Preference Data, Nonresponse

Document Type

Dissertation

Share

COinS