The majority of the transcriptions we prepared in order to make our digital edition were carried out by members of the Estoria de Espanna Digital project team. Some transcriptions, however, were done by crowdsourced volunteer transcribers. Around fifty people signed up to be crowdsourcers for the project, around seven of whom went on to become ‘active’, that is, did any transcribing for us that we could use in the preparation of the edition. This figure is consistent with what we could have predicted, based on research into crowdsourcing by Rose Holley. Of our seven crowdsourcers, and could be considered to be what Holley terms a ‘super volunteer’, i.e. someone who consistently achieves much more work than other volunteers and develops a longterm link to the project, sometimes staying with it for years.
Why did we use crowdsourcers?
Our aim in using transcriptions done by volunteers was never to remove the need for the in-house team of transcribers, and when we take into account the working hours involved in setting up the infrastructure, recruiting, training, moderating the work of crowdsourcers and feeding back to them it would actually have been quicker for us to do all of the transcriptions ourselves. Our aim was not to produce the transcriptions as quickly as possible, since at the level we were crowdsourcing this would just never have been the case – other, larger crowdsourcing projects such as Transcribe Bentham do primarily rely on the transcriptions of volunteers, but this was neither an option nor an objective for us. Instead, we used crowdsourcing as one of the ways in which we hoped to widen the impact of our project, to allow those not working in universities to take part in scholarly research and to have direct contact with and benefit from the research itself.
Who were our crowdsourcers?
Our crowdsourcers were a varied bunch: some were students, some undergraduate, some graduate, of topics related to our project, and some were working outside academia but had an interest in medieval studies, palaeography, transcription or other related themes. Our crowdsourcers came from many different countries, including Spain, Argentina and Australia. What they did have in common, however, was that they each had some kind of motivational factor to want to work for the project. For some this was a perceived direct benefit to their ongoing studies and for others this was the fulfilment of a ‘cognitive surplus’, to use Clay Shirky‘s term, which means they had a desire to learn, or to put their skills and experience to what they deemed to be a good use.
What did crowdsourcers do at the project?
Some crowdsourcers primarily did some of the ‘drudgery’ of creating a digital edition by doing tasks such as ensuring that the lines of text in the transcription matched the lines of text in the manuscript image, and inserting the correct line break tags. Although this may be perceived as ‘donkey work’ it was by no means unimportant: having tasks such as this completed enabled the in-house team to complete the rest of the transcription of that folio more quickly. Once they had mastered this skills, some transcribers chose to progress to more complex tasks such as inputting abbreviation tags and other XML encoding, according to our Transcription Guidelines. After careful training, moderation and feedback on our part, and a great deal of time, effort and hopefully enjoyment on his part, our one ‘super volunteer’ became an extremely accurate transcriber: by the time the transcription section of the preparation of the edition was completed, his folios were as accurate as those produced by the in-house transcription team.
How did we train the volunteer transcribers?
Crowdsourcers were initially trained through an online course produced by the PhD students on the project and using the online learning platform Canvas. Once they had completed their initial training, crowdsourcers were trained in an on-going style, through moderation of their transcriptions and diagnostic feedback given on a one-to-one basis designed to continue to motivate volunteers, to feed their cognitive surplus as they improved at the task, and to raise the accuracy of their transcriptions, making them more useful to use at the project.
Where can I read more about crowdsourcing at the Estoria de Espanna Digital project?
You can read more in Polly’s article on this topic, which is available here.