Top of page

Beyond the Scanned Image: Assessing Scholarly Uses of Digital Collections

Share this post:

Read the slides for a recent talk on this topic here.

The following is a guest post from Nicole Saylor, the head of the American Folklife Center‘s archives at the Library of Congress. Prior to her arrival at the Library, she was a member of the survey team while working as the head of Digital Research & Publishing at the University of Iowa Libraries.

It’s easy to see that digital collections are proliferating on the web. Just look at the growing corpora from Hathi Trust, Digital Public Library of America, ArtStor and Europeana, among many others. Providing online access to scholarship and cultural artifacts gathered in coherent aggregations in a variety of formats is increasingly driving the missions of many cultural heritage institutions. Yet, what is less apparent, is to what degree these digital collections are meeting the needs of current scholars.

A recent study of humanities faculty at twelve research institutions, led by Harriett Green, English and Digital Humanities librarian at University of Illinois at Urbana-Champaign, aimed to find out more about uses of digital collections among humanities scholars. A primary goal of the study was to help inform the areas of digital collection work in which libraries have expertise, such as metadata, information retrieval and other access issues.

The survey, conducted during the 2011-2012 academic year, included a web questionnaire and 17 in-person interviews. It was conducted on behalf of Project Bamboo, a now-completed national initiative to address the question, “How can we advance arts and humanities research through the development of shared technology services?” Green and Angela Courtney, Head of Arts and Humanities and Head of Reference at Indiana University-Bloomington Libraries, presented the survey findings at Digital Humanities 2013 this summer in Lincoln, Neb.

More than 60 percent of those surveyed said that digital collections comprise at least half of the sources they use in their research. The uses of digital collections ranged from the more traditional (researching historic newspapers, government reports, legal cases, etc.) to exploring high definition images of papyrus as the basis for textual reconstructions.

Findings centered largely on the need for sustained access and discovery of digital collections, and the desire for scholars to mix and reuse digital materials. One respondent replied, “The easier objects are to repurpose, remix and reuse the better.” Green categorizes the major themes of the findings into two categories: curation and interoperability.

To make digital collections more useful in research, respondents generally said they would like more completeness of content and a better way to search digital collections for the content they need. Another prominent request was for improved tools to annotate and edit digital collection objects broadly. One respondent said he/she wants “the ability to control your collection, set up your own library and so on and go deeper and deeper, adding tags, etc. Where it’s less of a skill and more of an expectation.”

“Most immediately, this study provided information to the Project Bamboo team on things to consider how to shape digital collections for scholarly needs,” said Green. “But on a larger scale, we hoped these findings will be useful to libraries who are interested in who is using their digital collections and how they’re being used.”

Green and Courtney are working on a full paper that compares their findings to the extensive qualitative data gathered by Project Bamboo research team member Quinn Dombrowski about scholarly practices during the Project Bamboo workshops held in 2008-2009. They hope the paper will be published within the next year.

“The goal of our investigations is to offer concrete analyses of how scholars are integrating digital content into their research workflow and how their research practices are evolving with the growth in digital content,” Green said. “Our Digital Humanities 2013 presentation received a very favorable response, and we hope our forthcoming publications will be useful to libraries and cultural institutions seeking to increase the impact of their digital collections.”

Comments (3)

  1. Thanks for the introduction to this useful survey! And thanks for the link to the slide show with its nice summary of the main findings. As I read the blog and slides, I could not help but muse about photographic collections. One always worries that humanities scholars fail to use pictures as “data,” instead turning to photo collections only when it comes time to illustrate essays or books. (I did note, however, many references to audio and video in the slides, although this was not an occasion to explain how these resources had been used.) The reason photographs came to mind was not simply my long-term wish to see more use-as-data but also to be sure that we call attention to the way that many digitized photograph collections have moved from inaccessible to accessible. In part, this is due to the frequent presentation of collections of _negatives_. At the Library of Congress, hundreds of thousands of our online images have been scanned from negatives for which there were no positive prints. (Lots more from other institutions too!) Thus digitization has made them available in more drastic way than, say, the scanning of books or manuscripts. For the most part, prior to digitization, these photo-negative collections were off limits (“not allowed to touch”) and/or simply in a hard-to-comprehend state (“yikes, all the light parts of the picture are dark!”). Digitization has moved this resource out of a locked room and put it before the scholar’s eyes. Now we all await a humanities specialists doing some nifty data-mining with this big visual resource!

  2. Insightful comments, Carl, as always. With regard to one interesting way in which we can look at digital image repositories as what you term “use-as-data” is Peter Leonard’s recent work on The Chicagoan’s cover design trends over time. See http://youtu.be/FmQ08xqx-Qw beginning at the 16:14 mark. There his main data points were luminance, hue, and saturation values from the covers’ digitized images. Coming from the digital capture side of things as I do at UConn, I particularly like to see this research as it creates yet another use case for the general importance of accurate baseline capture across material types and in turn the applied value of following such still imaging standards as FADGI, and Metamorphoze as closely as we operationally can. In Peter’s case, inaccurate capture could to some degree skew the data set that he is conducting his mathematical tests upon.

  3. I read this post and the commentary with great interest, as scholarly use of digital collections is my central area of research at UNC-Chapel Hill. I look forward to reading Harriet and Quinn’s paper. My own research on how historians search for, access, and use primary source materials, with particular emphasis on digitized collections, will be in the next issue of American Archivist, out in a few weeks. And of course, Schonfeld and Rutner’s Ithaka’s S & R Report is an invaluable resource to gauge trends and changes in historical practice.

    I would love to hear from any historians who use digitized photographs in their research pursuits, as this forms the cornerstone of my dissertation research. Please feel free to contact me to talk more: achass AT email.unc.edu

Add a Comment

Your email address will not be published. Required fields are marked *