Evaluating Data Quality and Changes in Species Identification in a Citizen Science Bird Monitoring Project

Citizen science has potential to provide multiple benefits to participants and the professional scientific community, and those benefits can be realized if citizen science projects are intentionally designed to achieve research objectives, and if participants have the skills, knowledge, and training to collect high-quality data. Using three years of data from a citizen science bird monitoring project in Salt Lake City, Utah, we assessed bird songs and calls learned by volunteers, and compared species detections, number of birds, and distance measurements between point counts by citizen scientists and professional biologists. We found significant increases in correct species identification for citizen scientists after going through the training program; the average percentage of bird songs and calls identified rose from 42.5% before training to 72.7% after training (p < 0.00001). For two data quality metrics, citizen scientists and professional biologists collected similar quality data: the average number of birds and average detection distances were not significantly different for point counts conducted by citizen scientists and professional biologists in the same locations. However, professional biologists identified an average of 1.48 more species than citizen scientists (p < 0.00001). Our findings emphasize the importance of evaluating training programs and data accuracy for citizen science projects. In instances in which citizen scientists may not be performing at the same level as professional biologists, identifying these patterns ensures that they can be fully explained and accounted for during data analysis.