A Binary Ordinal Probit Model for CVM Data Based on On-Site Samples
2007 (English)Report (Other academic)
A commonly used sampling design in economic valuation studies such as theCVM, is on-site sampling. If this sampling design is used, the sampling inclu-sion probabilities may be correlated with respondents’ valuations, invalidatingwelfare measures derived from estimates of the probit model. This paper pro-poses a model based on the bivariate ordinal probit, a model that can be usedto analyze binary choice CV data gathered by on-site sampling. This paperpresents the model, derives the log-likelihood, and illustrates the MLE using asmall simulation study. The model is presented, the log-likelihood is derived andthe properties of the MLE’s are illustrated with a small simulation study. Thesimulation results show the proposed estimator to be an interesting alternative,but the estimator should be evaluated in more population models.
Place, publisher, year, edition, pages
On-site, ML-estimation, Bivariate ordered probit, Sample inclusion.
Research subject Statistics
IdentifiersURN: urn:nbn:se:su:diva-87055OAI: oai:DiVA.org:su-87055DiVA: diva2:600719