The influence of sampling design on spatial data quality in a geographic citizen science project

Geographic citizen science has much potential to assist in wildlife research and conservation, but the quality of observation data is a key concern. We examined the effects of sampling design on the quality of spatial data collected for a koala citizen science project in Australia. Data were collected from three samples-volunteers (n = 454), an Internet panel (n = 103), and landowners (n = 35)-to assess spatial data quality, a dimension of citizen science projects rarely considered. The locational accuracy of koala observations among the samples was similar when benchmarked against authoritative data (i.e., an expert-derived koala distribution model), but there were differences in the quantity of data generated. Fewer koala location data were generated per participant by the Internet panel sample than the volunteer or landowner samples. Spatial preferences for land uses affecting koala conservation were also mapped, with landowners more likely to map locations for residential and tourism development and volunteers less likely. These spatial preferences have the potential to influence the social acceptability of future koala conservation proposals. With careful sampling design, both citizen observations and land use preferences can be included within the same project to augment scientific assessments and identify conservation opportunities and constraints.