Estimating the benefits to Florida households from avoiding another Gulf oil spill using the contingent valuation method: Internal validity tests with probability‐based and opt‐in samples

Abstract

Abstract This paper evaluates contingent valuation method data quality by examining differences in results between probability‐based and opt‐in internet samples using data from a survey estimating passive use losses associated with the BP/Deepwater Horizon oil spill. Both samples pass the internal validity tests of negative price (bid) coefficient and positive sign on income but only the probability‐based sample passed the scope test and shows evidence of use values. We find that the willingness to pay estimates from the opt‐in sample may be over‐estimated compared to the probability‐based sample, though not statistically different. Results support the utility of less expensive opt‐in samples.