Enhancing Information in Stated Preference Choice Experiment Surveys: Quantifying the Role of Spatial Factors and Illustrating Resource Proximity

Within the stated preference literature, assumptions are often made regarding how respondents use spatial information when they answer a choice experiment survey. For example, it is often taken for granted that a policy area map, included in many surveys, provides sufficient spatial information to support fully informed preference elicitation and that respondents understand all relevant spatial factors that affect their anticipated welfare from hypothetical policy changes. These widely made assumptions are largely untested. I test these hypotheses by creating individually tailored surveys that include the respondent's location and compare the results to those from an otherwise identical, traditional choice experiment survey---identical in all respects except for the amount of spatial information provided. A split sample of respondents received either the status quo survey ("generic survey") or the survey containing additional spatial information ("geocoded survey"). Results suggest that the generic survey does not provide sufficient spatial information to support fully informed preference elicitation and this information shortfall specifically affects respondent understanding of their individual spatial characteristics (e.g., the location of the their homes relative to hypothetical policy affects). I also develop and test a novel (and arguably more relevant) means to characterize spatial heterogeneity in stated preference willingness to pay. The model evaluates heterogeneity associated with the amount of an affected resource surrounding the respondent. This analysis presents the intuitive result that more riparian land around a respondent at distance bands of various length. This "quantity-within-distance-x" measure is used as a substitute for the "distance-to-nearest-point" measure commonly used in models of spatial welfare heterogeneity. This analysis presents the intuitive result that increasing the area of riparian land around a respondent increases marginal willingness to pay (WTP) for riparian land restoration, all else equal. It also shows WTP spatial patterns that are not found in a distance decay analysis---the most common stated preference spatial analysis. This dissertation questions fundamental and underlying assumptions commonly (and often implicitly) made in the stated preference literature regarding the role of spatial welfare elicitation. It shows that these assumptions are either invalid or more nuanced than typically assumed. Additionally, it shows that a novel, and computationally difficult, spatial analysis provides policy relevant insights invisible to common methods of analysis.