Today we’re excited to share a guest post from Prof. Amanda Zellmer (Occidental College) highlighting a recent journal publication about Zooniverse usage and camera trap projects.
We are excited to announce a new paper that uses the Zooniverse to study the accuracy of camera trap photo identifications, and analyzing whether accuracy varies depending on the classifier’s level of experience with biology. The paper by Roshni Katrak-Adefowora, Jessica Blickley, and Amanda Zellmer was published in the journal of Citizen Science: Theory and Practice earlier this month and is titled “Just-in-Time Training Improves Accuracy of Citizen Scientist Wildlife Identifications from Camera Trap Photos”.
The study found that citizen scientists with no formal background in biology who received just-in-time training had a significantly higher accuracy of photo identification than those that received no training. In addition, these citizen scientists who received training were able to identify wildlife images with as much accuracy as citizen scientists with a professional background in biology. These results suggest that citizen scientists can contribute accurate data to photo identification studies even when only provided limited training, making such projects more accessible to citizen scientists eager to get involved.
jitt_result
Figure 2 from Katrak-Adefowora et al. 2020 showing that volunteers with no biology background (left pair of boxes) had lower scores without just-in-time-training (JITT; includes a tutorial, interface tools) than participants with higher levels of biology experience. However, through the use of JITT, participants with no biology background were able to achieve a similar level of accuracy as professional biologists (right pair of boxes).
For all the details on this study, check out the paper at this link: https://theoryandpractice.citizenscienceassociation.org/article/10.5334/cstp.219/