Advantages and Drawbacks of Open-Ended, Use-Agnostic Citizen Science Data Collection: A Case Study

Abstract

Citizen science projects that collect natural history observations often do not have an underlying research question in mind. Thus, data generated from such projects can be considered “use-agnostic.” Nevertheless, such projects can yield important insights about species distributions. Many of these projects use a class-based data schema, whereby contributors must supply a species identification. This can limit participation if contributors are not confident in their identifications, and can introduce data quality issues if species identification is incorrect. Some projects, such as iNaturalist, circumvent this with crowdsourced species identifications based on contributed photographs, or by grading confidence in the data based on attributes of the sighting and/or contributor. An alternative to a class-based data schema is an open-ended (instance-based) one, where contributors are free to identify their sighting at whatever taxonomic resolution they are most confident, and/or describe the sighting based on attributes. This can increase participation (data completeness) and have the benefit of adding additional (and sometimes unexpected) information. The regionally-focused citizen science website NLNature.com was designed to experimentally examine how class-based versus instance-based schema affected contributions and data quality. Here, we show that the instance-based schema yielded not only more contributions, but also several of ecological importance. Thus, allowing contributors to supply natural history information at a level familiar to them increases data completeness and facilitates unanticipated contributions.