Individualized Geocoding in Stated Preference Questionnaires: Implications for Survey Design and Welfare Estimation

Evaluating willingness to pay for the temporal distribution of different air quality improvements: Is China's clean air target adequate to ensure welfare maximization?

“…The parameters on each AQL are specified random with a normal distribution and the parameter on the ASC is specified as fixed. The coefficient on sign‐reversed cost is specified using a bounded triangular distribution, ensuring a positive marginal utility of income (Hensher & Greene, ; Johnston, Holland, & Yao, ).…”

Optimized quantity-within-distance models of spatial welfare heterogeneity

“…The theoretical justification for distance-based analysis is drawn from a use-value perspective, largely reflecting a recreation-demand paradigm; greater travel distances to a resource imply lower net values, ceteris paribus, given the greater cost of reaching the site (Bateman et al 2006). Similar justification applies neither to nonuse values, nor to use values for which site access is unimportant (Bateman et al 2006;De Valck et al 2017;Hanley et al 2003;Johnston et al 2016). This may explain why distance-decay analyses of nonuse values have found mixed results, along with differences in distance-decay gradients between users and nonusers (Bateman et al , 2006Hanley et al 2003;Johnston and Ramachandran 2014;Jørgensen et al 2013;Loomis 2000;Rolfe and Windle 2012).…”

“…Design efficiency was then reevaluated using alternative assumptions for preference parameters and utility structure. Additional details are provided by Johnston et al (2016).…”

“…From the underlying DCE data (see below), respondent addresses were first geocoded using an online application programming interface, with minor random perturbations to protect respondent confidentiality (Johnston et al 2016). This provided the location of each respondent's household, which was used as the basis for subsequent calculation of respondent-specific distances and areas.…”

“…Information was conveyed via a combination of text, graphics including Geographic Information System (GIS) maps, and photographs, all of which were subject to pretesting. Questionnaires included a detailed map of the policy area, including landmarks, roads and potentially affected riparian land in the watershed (Figure 2).The policy area was illustrated at three nested magnifications with a distance scale to enable respondents to infer proximities(Johnston et al 2016). The questionnaire also included statements highlighting consequentiality, following Carson and Groves(2007).Choice options were characterized by four ecological attributes, two attributes characterizing development restrictions/enforcement, and one attribute characterizing unavoidable annual household cost.…”