Demand for environmental policies to improve health: Evaluating community-level policy scenarios

Exchange Rules and the Incentive Compatibility of Choice Experiments

“…Total willingness to pay is computed assuming a mean age of 22 years, mean household income of $55,000 and setting each categorical demographic characteristic equal to the base category (i.e., setting each dummy variable describing a characteristic equal to zero) b The 5th and 95th percentiles are reported in brackets and are computed following the methods described in Hole (2007) and Bosworth et al (2009) These results are somewhat surprising. We suspected that in the absence of a provision rule respondents might assume a plurality vote, group provision rule; and this did not occur.…”

“…4) interacted with the ASC. Total willingness to pay is computed assuming a mean age of 22 years, mean household income of $55,000 and setting each categorical demographic characteristic equal to the base category (i.e., setting each dummy variable describing a characteristic equal to zero) b Reported in each cell is the P-value for the convolutions test (Poe et al 2005) of the null hypothesis that total (or marginal) willingness to pay c Percent bias is based on the model including heterogeneity and is computed as |(WTP PPM − WTP PPM-B )/WTP PPM-B | where WTP is total or marginal willingness to pay d The 5th and 95th percentiles are reported in brackets and are computed following the methods described in Hole (2007) and Bosworth et al (2009) e Percent bias is not computed for t-shirt types which did not have a statistically significant coefficient in both treatments than the estimated MWTP from a real-payment treatment. However, while the private good results are not statistically different, the point of estimates of MWTP from the hypothetical survey are larger in magnitude.…”

Attribute Framing in Choice Experiments: How Do Attribute Level Descriptions Affect Value Estimates?

“…Framing effects become problematic if they are unintended by the researcher and if they give rise to biased answers (Fischer et al 1987;Ajzen et al 1996;Bosworth et al 2009;Luisetti et al 2011). In CEs, framing effects have been observed from varying the choice set complexity; such as changing the number of choice sets per survey, alternatives per set, attributes per alternative, or levels per attribute (Swait and Adamowicz 2001;Breffle and Rowe 2002;Caussade et al 2005;Hensher 2006a).…”

Estimating Sign-Dependent Societal Preferences for Quality of Life

“…As stressed by Bosworth et al (2009), it is important to have qualitative information alongside quantitative estimates of preferences. Therefore, we also ask the respondents for explanations for choosing a particular option.…”