Tweaking our transcription numbers (again).
As many of you know, Notes From Nature collects more than one transcription for each image (or subject, in Zooniverse speak). This is not uncommon among Zoonivere projects. This means that we get more than one set of data for each specimen image, making sure that each transcription is done by a different user. In fact, the algorithm does not permit a user to transcribe the same image more than once. This is why it is possible for a collection to be incomplete, but some users are no longer served images to transcribe.
The purpose behind this is to assist with quality assessment of the data. It is not that our volunteers do not provide quality transcriptions! In fact, in many cases Notes From Nature volunteers provide higher quality transcriptions than trained technicians. However, there are various situations where the information on a label can be up to interpretation. This is especially true for handwritten labels and information that goes into our free text fields (e.g. locality). Having more than one transcription helps us come up with the best possible transcription for a given image, using software that analyzes the whole set of transcription.
This is not the first time we have changed the number of transcriptions. When we first started Notes From Nature in 2013 we were collecting 10 transcriptions per image, which we quickly realized was many more than was needed. We reduced the number of transcriptions to 4 per image. We recently reanalyzed this issue again and found that 3 transcriptions per image would be sufficient. This means that we have moved from 4 to 3 transcriptions per image as of this week. So far this has taken place in the CalBug, Herbarium and Macrofungi interfaces.
Some of our volunteers have already noted that our progress bars showed a bit bump this week and this is the reason why. When the algorithm was changed any images with 3 transcriptions were immediately retired from the queue. As always, we greatly appreciate your hard work and are continually looking for ways to make this project more efficient and enjoyable!
Cheers for the update!