EBird

Citizen/community science samples are self-selected, limiting their value for predicting population behavior. We field a general-population survey to elicit different levels of knowledge about (or engagement with) the eBird project and transfer a fitted sample-selection function, combining that function with eBird member attributes to correct a model of spatial consideration sets.

Members of Cornell University's eBird project report bird sitings at geo-coded destinations. With origin information, we can infer the value to this group of biodiversity in bird species by observing how much more in travel costs they are willing to incur to gain opportunities to see more diverse populations of birds.