Hypothetical bias mitigation in representative and convenience samples

Abstract

This is a case study comparing outcomes for a probability-based representative sample versus a non-probability convenience sample for the valuation of beach condition information among Gulf of Mexico residents. We test the efficacy of several techniques used to adjust for hypothetical bias and sample weighting to reduce hypothetical willingness to pay (WTP). Weighting makes the WTP between the two samples similar, but model equivalence with respect to the significance of explanatory variables is rejected. The results support the use of certainty follow-ups, which consistently reduced WTP, regardless of the sampling approach or weighting.