Incentivized experiments and stated preference methods are commonly viewed as substitutes in preference elicitation. We leverage the distinct strengths of each approach by combining a fully incentivized risk experiment with a stated preference survey to model utility for intrinsically risky attributes. A door-to-door survey of 981 participants in a drought-prone region elicits preferences for alternative sources of municipal water, conditional on water price and quality. Participants' observed and imputed risk attitudes are incorporated into preference estimation. Our results are consistent with the hypothesis that participants are concerned about water supply risk, but not about new technology risk.