Community-based Data Validation Practices in Citizen Science

In a study of the dynamics of citizen science data curation, researchers Andrea Wiggins and Yurong He identified several interesting trends impacting the perceived quality of observational data shared on the iNaturalist social network. The paper was nominated for an Honorable Mention and appears in the proceedings of the ACM CSCW 2016 conference.

The study was motivated by the huge volumes of data now being produced in citizen science, which often requires some form of validation prior to its use. Since professional researchers are unable to keep up with quality management, Wiggins and He evaluated how well the citizen science participants are able to curate the data they contribute. The researchers based their findings on observations from an educator workshop and publicly accessible data from iNaturalist, using mixed methods to ground statistical analyses in actual practice. They extended a theoretical framework that focuses on information assessability, linking information provenance with reliability, or ability to evaluate changes to the record, and connecting information stewardship with informativity, or amount of detail in the record.

While the researchers were not surprised to find that data about birds drew more attention than content about species, other findings were more unexpected. For example, observations for which the contributors asked for help with species identification were actually less likely to receive assistance. A more substantial concern was that observations uploaded from mobile phones were also less likely to be reviewed, suggesting that mobile devices may have some drawbacks for supporting this type of participation.

Read the preprint of the article here.