In 2013 and 2014, Stanford University and Historypin, with support from the Andrew W. Mellon Foundation, are taking on some big infrastructure development and looking to contribute to answering a big question: how can the knowledge and capacity of the crowd be harnessed for humanities research through digital tools?
We’ve seen some great examples of crowdsourced contributions to research over the last few years, mainly within the sciences. From the pioneering Human Genome Project to the recent Old Weather project, large amounts of people have been invited to take part in providing, transcribing, verifying and structuring large amounts of data, often to great effect.
More and more scientific crowdsourcing projects emerge all the time and more layers of value are being proven. However, this has not easily translated into similar success within humanities research and the case for widespread adoption is still weak. Why is this?
The role of crowdsourcing within academic research comes up against two key challenges:
i) If the crowd is being employed for its capacity to carry out large amounts of machine-like tasks that a machine can’t quite carry out, is it worth it? More specifically, if you have to build, maintain and manage dedicated tools to support a community of contributors, as well as engage and motivate them, would you be better off employing research assistants to churn through it directly?
ii) If the crowd is being employed for its knowledge, how can those contributions become “useful” rather than a big array of unverified, unrefined, objective stuff?
On the first challenge, the jury is most definitely out. The Transcribe Bentham project based at University College London successfully engaged enough participants to complete a number of transcription tasks to a high degree of accuracy. However, the team very commendably did the math on whether this approach was worth it and had to conclude that it would have been quicker and cheaper to employ research assistants.
Mostly, conclusions on efficiency are less clear cut (and, perhaps, analysis less self-effacing), but what is so far clear is that if you ignore all the incidental benefits of these kinds of tasks and outputs (popular engagement with research, social interaction etc.), there isn’t enough evidence to inspire interest from traditional humanities academics.
On the second, the leap from traditional research activities is perhaps smaller. Academics within the humanities often engage the public in their research, gathering first hand accounts, accessing privately held materials and working closely with communities of all kinds in different ways. So, the appetite for making this engagement easier to do at scale feels greater.
Many challenges remain, however, from the limitation of available tools to the scant resources to engage and manage contributions from lots of people to the problems with verifying, standardizing and employing large amounts of data.
It is these challenges that have brought Stanford University’s Center for Spatial and Textual Analysis, Historypin and the Andrew W. Mellon Foundation together.
For Historypin, armed with a core set of social and cultural aims, assessing the value of crowdsourcing through this kind of partnership is very exciting. We believe that heritage materials and memories hold great potential to bring different generations, cultures and communities together to share, interact and learn more. Essentially, we see this type of collaboration as a powerful way of generating and sustaining bridging social capital. However, the motivations for large, diverse audiences to participate in gathering, opening up and contextualising their cultural heritage stem from the perceived value of this activity and its outputs. Proven value within academic research would provide a strong source of intrinsic motivations and help scale and multiply participation and its impact.
Together, we are setting out to develop and test tools that invite contributions to three different types of humanities research within three distinct projects as outlined in earlier posts: Year of the Bay, the Western Railroads Project, and Tagging 500 Novels.
Across all of these projects, the iterative development of tools and testing of different approaches to audience engagement will run alongside the ongoing assessment of the efficacy, efficiency and scalability of this activity. We aim to emerge with both a robust set of tools and a series of insights into the value of crowdsourcing in these areas of research. Through these, we want to make a meaningful contribution to the key challenges that have been identified.
On the first challenge, we propose that the development of free to use and easy to access transcription tools that are refined and fit for purpose will substantially lower the investment and skills needed to harness the crowd’s capacity. The work of the partnership will make some initial inroads into this open transcription toolset.
On the second challenge, which will be our main focus, we propose that the knowledge of the crowd can provide profound value to research if the environment for that collaboration is both simple to use and sophisticated in its role as aggregator and verifier of public contributions. The work of the partnership will invest substantial time and effort in developing, refining and assessing these collaborative environments.
We’ll look forward to keeping you posted on our progress over the next two years.
No comments:
Post a Comment