Stated preference scenarios often provide information on intermediate biophysical processes but omit information on the resulting final services that provide utility. This may cause respondents to speculate about the effects of intermediate outcomes on their welfare, leading to biased welfare estimates. This work clarifies distinctions between intermediate and final ecosystem services within stated preference valuation and develops a structural model by which to infer respondents’ speculations when a final ecosystem service is omitted. The model also derives implications for welfare estimates. Methods and results are illustrated using an application of choice experiments to fish restoration in Rhode Island's Pawtuxet watershed.