This paper establishes a new method for eliciting willingness to pay (WTP) for non-market goods in contingent valuation (CV) studies with an open-ended elicitation format: the Range-WTP method. In contrast to the traditional approach for eliciting a single value (Point-WTP), Range-WTP explicitly allows for preference uncertainty in responses. Using data from two novel large-scale surveys on the perception of solar radiation management (SRM), a little-known technique for counteracting climate change, we compare the performance of both methods in the field. In doing so, we use the criterion of theoretical validity and measure the degree to which WTP values are consistent with theoretical expectations. In addition, we analyse the test-retest reliability and stability of our results over time. Our evidence suggests that the Range-WTP method clearly outperforms the Point-WTP method.