Out of the Database, Into the Classroom

Final Report from the RLG Instructional Technology Advisory Group

February 2005

Authors: Günter Waibel (RLG) and Arnold Arcolio (RLG)

RLG Cultural Materials provides licensed access to digitized manuscripts, photos, art, historical documents, and memorabilia from around the world to support teaching and learning. Its terms of use explicitly encourage faculty and students to take full advantage of the over 500,000 digital surrogates represented for "educational purposes, including research, presentations, and student assignments." While many faculty are increasingly eager to use digital images in their curricula, they often encounter a variety of difficulties and frustrations in making the switch from the slide carousel to digital projection. RLG formed an Instructional Technology Advisory Group in the summer of 2003 to test our assumptions about how digital images are discovered, acquired and used—and about preferences for the future.

RLG's assumptions at the point of convening the advisory group can be summarized as follows: RLG Cultural Materials, like any database providing digitized materials to support education, should interface with as many existing campus tools as possible in order to fully capitalize on the potential for enriching instruction with quality digital images. Faculty need content from different sources. Campuses often license multiple digital image databases. Increasingly, campuses also digitize materials from their own special collections or from their slide libraries to provide additional sources of digital content. To support their lectures, faculty should be able to bring together digital images from all sources available (including their own personal collections) into their preferred presentation tool. Rather than incorporate functionality for classroom presentation into its services, RLG should focus on supporting existing instructional technology tools already in use, as well as those being developed by others.

Advisory group

To test this hypothesis and make sure the methods we support for transferring digital images and metadata evolve in step with instructional technology tools and strategies, RLG convened the Instructional Technology Advisory Group. The diverse roster of participants included representatives from eight RLG member campuses plus James Madison University, and brought together visual resource specialists, librarians, and experts from instructional technology units:

• Marcelo Clerici-Arias (Stanford University)
• Jane Cody (University of Southern California)
• Elizabeth Dupuis (University of California, Berkeley)
• Laurie Harrison (University of Toronto)
• Byron C. Mayes (Temple University)
• Jon Mott (Brigham Young University)
• Henry Pisciotta (Pennsylvania State University)
• Oya Y. Rieger (Cornell University)
• Christina Updike (James Madison University)

The group was charged with testing RLG assumptions, and making sure that our path forward meets the needs of the staff and faculty interacting with image databases. In particular, the group was asked to make recommendations for data export approaches that allow images and metadata to travel from a database to the software used for presentation in classrooms, as well as to support other instructional needs. While the advisory group used RLG Cultural Materials as its exemplar, we aimed for findings which would be generic enough to apply to any database providing image content for education.

Since the ultimate hope was to make teaching from digital images more convenient, the group decided to conduct interviews with faculty to provide more background on how they currently use digital images in their classrooms, what obstacles they encounter, and how they envision teaching with digital surrogates in the future. Program officers Merrilee Proffitt and Günter Waibel, the RLG liaisons to the advisory group, along with Arnold Arcolio, RLG information architect, interviewed humanities faculty at three campuses:

• University of Southern California (October 2003): Four faculty members from the Classics, Art History, English and History departments
• University of California, Berkeley (October 2003): Three faculty members from the Anthropology, Art, Technology and Culture, Art History departments
• Stanford University (February 2004) Two faculty members from the History and Center for Teaching & Learning departments

The advisory group helped to shape the discussion guide that was used in the interviews. The guide covered image gathering (where the images come from), image management (how images are made available for repeated use), and teaching (how images are presented in the classroom). We also asked faculty to imagine for us how all these processes might be different in the future.

Faculty interview results

While the specific circumstances and interests of the faculty members we talked to varied, a surprisingly straightforward picture of the current state and hopes for the future emerged. The four paragraphs below summarize what we learned.

Where do the images come from?
Most participants reported that they create and gather their own digital images for the classroom—they take pictures with digital cameras while traveling or they scan images from books. Concerns about rights and fair use inhibit most from making these images available to students outside the classroom, or to colleagues, as widely as they would like. Additionally, they find images on the Web. Almost every faculty member interviewed regarded Google Image Search as a quick, reliable way of retrieving images for teaching. While the common deficiencies in terms of file size and color fidelity are apparent to them, ease of use and the search engine's ability to deliver a suitable image for almost any request outweigh those shortcomings. Those interviewed were largely unaware of licensed resources available to them. Among those that had tried to use licensed resources, a number of faculty found that the content does not always speak to their immediate interest. Most of them portray their own needs as highly eclectic and specific, and some feel that licensed resources cater too much to a "greatest hits" mentality. Most faculty mentioned that the "hassle" involved in logging in from off campus made it unlikely that they would use licensed resources from home. By and large, the library plays only a small role in supplying the faculty with digital image content. While we heard about library attempts to make faculty more aware of licensed resources, these communications seem to largely bypass their audience.

How are images made available for repeated use?
Most faculty save the images they gather or create for future use. Some simply keep the PowerPoint slides of their lectures for repurposing at a later date. Others save image files to a folder structure organized by courses or topics of interest. Several participants are beginning to question the continued utility of the media they had chosen; images on Zip drives and other removable storage media closing in on obsolescence, for example, are no longer easy to access. Most were concerned when these issues were raised. Most said they don't really have time to spend developing a means for image organization. A few use databases such as Filemaker, Microsoft Access, or low-end multimedia management systems such as iViewMedia to keep track of their images. Almost all readily admit that their particular system for storing images lacks even the most basic metadata necessary for effectively retrieving the content in the future. We encountered a few databases shared among colleagues or departments; these generally put more emphasis on descriptive metadata. One fledgling project integrates digital images from various licensed and local sources into a single teaching database, generating great anticipation among the faculty.

How are images presented in the classroom?
Most faculty use the readily available PowerPoint for presenting digital images in the classroom, yet they dislike it as an instructional tool. Their chief complaints are that PowerPoint does not allow them to zoom or to compare / contrast different images side by side. Some faculty expressed concern that the set nature of a PowerPoint presentation discourages the lively interchange and dialogue they would like to create around the images. Our small sample of institutions did not yield any faculty who use more sophisticated presentation tools such as Luna Insight or MDID. A few reported that they use the courseware package their campus offers (such as Blackboard) to make digital images available to their students for preparation and study.

In an ideal world, how would you find, manage, and use digital images?
In their dream of the future, faculty envision access to high-quality, rights-cleared, persistently available images with the same retrieval success rate as Google Image Search. They are also quite certain that they will continue to use digital images they have had a hand in producing. One of our interviewees revealed a latent feeling towards licensed resources by saying that he simply refuses to log into resource after resource without finding what he needs. The idea of searching across all licensed resources and the Web at the same time found many proponents.

Testing the interview results with our expert advisors

The findings from the interviews were shared with the advisory group, who compared them with their own experiences. Some shared them with faculty or instructional technology professionals at their own campus to see if their views differed. A lot of what they reported reinforced what had been heard in the interviews. We also actively participated in or studied other work in the area, including:

• The Visual Image User Study at Penn State. Henry Pisciotta, who served on the steering committee of this study as well as the RLG working group, often deepened our understanding of what we had heard from faculty by citing data from the Penn State study.
• The Digital Resources Study at the Higher Education in the Digital Age program, Center for Studies in Higher Education, UC Berkeley. RLG's Arnold Arcolio, Merrilee Proffitt and Günter Waibel took advantage of an invitation extended by Diane Harley, the project's principal investigator, to take part in a number of project meetings. We funneled preliminary results from our findings into the project, and benefited from the discussions as well as reading drafts of the report.
• The California Digital Library's Image Service Demonstrator project developed discussion guides for testing their Luna Insight client with faculty at the same time we developed our discussion guides, and we exchanged questionnaires for our mutual benefit.

Although our sample was small and the approach non-scientific, the confirmation of much of what we had learned in related projects gives us confidence that the results from our study will be quite useful.

The outcomes were both sobering and encouraging. PowerPoint, though limited in its features, is widely used. Metadata, though of demonstrable value, is rarely captured. Of course we had hoped that licensed resources would have found more favorable reception with faculty. On the other hand, the amount of passion faculty display for teaching from digital resources confirms that they are potentially quite interested in aggregated resources like those RLG offers. And the more we understand what they value, the better positioned we are to deliver the goods to them in a convenient way.

Conclusions

The faculty interviews and advisory group input have clarified two crucial issues: First, image databases need to leverage the breadth and simplicity of online search engines such as Google Image Search to achieve higher use. RLG's Trove.net™ is a scaled-down version of Cultural Materials content on the open web, which search engines spider and index. The original intent of Trove.net was to expose information about digital images to increase awareness of contributors’ collections and to generate licensing revenue for them; however, important in this context, it has the desired side effect of exposing the content to faculty, who (as we have learned) often are not aware of what resources their library licenses. Faculty trolling Google now find RLG Cultural Materials content among their results. If their campus subscribes to Cultural Materials, they can then retrieve the high-resolution images for teaching. If not, the image is available on a pay-per-use basis. Trove.net meets faculty where they choose to congregate, and makes high-quality, rights-cleared images available in their retrieval mechanism of choice.

Second, image databases supporting education cannot merely weave together aggregated content—they also have to supply the functionality for users to unravel the content for reaggregation into their own tools and resources. Faculty, departments, and in some cases entire campuses create their own teaching databases geared to their specific needs by "cherry-picking" from licensed resources and combining those subsets with their own image assets. For campus reaggregation, capturing descriptive metadata along with the digital image is crucial. In this particular context, a powerful and intuitive search interface may not be as important as ready access to the raw content sitting in the database, to be deployed in the method most advantageous for the overall instructional technology strategy. From a campus perspective, this can influence licensed resource selection decisions.

Scenarios for image integration

Together with the advisory group, we turned what we had learned during our interviews into specific recommendations for behaviors or scenarios a resource such as RLG Cultural Materials should support. One variant in the different scenarios is scale—how many images need to be transferred at a time? Another is the ultimate objective for downloading the images—is the goal to satisfy an immediate need such as an upcoming lecture, or to support a long-term strategy such as a more comprehensive reaggregation of selected images into a local image management system?

Immediate re-use
In this most basic scenario, faculty look for a single image or a small number of images that they want to use in their presentation for an upcoming lecture. Metadata is of little interest to faculty in this scenario. Mechanisms to satisfy this need are, for example, a right-click download or e-mail.

Selective reaggregation
When the objective shifts from gathering a small number of images to a more sizeable number (dozens), acting on each image individually becomes increasingly burdensome. If a faculty member gathers images for an entire semester or a staff member cherry-picks images for a departmental resource, they need a mechanism that allows them to select a group of images, review and edit their selection, and download the entire batch once they are satisfied. Since these images may wind up in a database supporting instruction, they need to be searchable, therefore the stakes for transporting descriptive metadata associated with the images rise.

Wholesale reaggregation
In this last scenario, the number of images and descriptions is potentially very large. Instructional technology staff, charged with integrating licensed resources into a campuswide database, load substantial numbers of images and descriptions into their system of choice. This data transfer may happen as a process of extraction and reintegration, in which case all the data and images end up living in the locally supported system.

The Advisory Group also discussed the promise of metasearching to integrate searches across a number of licensed resources. When we informally polled the advisors, they were evenly split into two camps: half of them saw their campus achieving interoperability of their image resources through extraction and reintegration; the other half had high hopes for metasearching to bring images together. While metasearching seems to be the more efficient approach, our discussions also surfaced its shortcomings for faculty use. A metasearch allows the discovery of resources from many different databases, but it does not in and of itself provide a mechanism for integrating the discovered images into a presentation tool. Metasearching would have to be supplemented with additional technology to allow the selection of desired images and local loading into an instructional technology tool.

While our discussions mostly focused on supporting image use in the classroom, the Advisory Group also considered aspects of integrating images into courseware packages supporting learning outside of the classroom, such as class preparation or study. To display images in these online systems, images should not have to make the arduous journey out of the licensed resource and into the courseware package. Rather, licensed resources could provide permanent identifiers, which would allow images to be referenced from within the courseware.

The road ahead

The insights of the Instructional Technology Advisory Group are shaping RLG's plans to add features to RLG Cultural Materials and other image resources and are helping us to deliver digital objects in a form that's useful and intuitive. We are entering into discussions with many major providers of database-driven instructional technology tools, such as MDID (Madison Digital Image Database software), ARTstor, and Luna Insight, to investigate specific data formats for export and reimport and/or APIs (application program interfaces) that RLG image resources can target for other forms of integration. RLG is also investigating ways to make its resources more easily metasearchable. As RLG Cultural Materials continues to grow through a steady influx of contributed collections, and Trove.net exposes those images to unsuspecting faculty members discovering images through Google, we will work to create the ideal future of digital images in the classroom we've heard faculty members describe.

Appendix A

Appendix B