Institutional Repositories

Policies and Practices of Institutional Repositories
Monday, June 27, 8:30 am - 12:00 pm

Most of the LITA events have been extraordinarily well blogged on the LITA Blog. Unfortunately, it looks like I'm the first to cover this session. I was rather hoping that I could do a quick summary and let others do the more substantive work. This was a session well worth the effort to blog, so here goes.

We had five speakers from five different universities--which will make this a very long post.

University of Rochester
At the University of Rochester, Susan Gibbons discovered that one should never let a policy get in the way of getting good content. A policy was written, but what actually happened was a "faculty-sponsored" model--if a faculty member wants something in the repository, it goes in. This includes student papers that faculty members are willing to sponsor for inclusion.

The University of Rochester uses DSpace, an open-source solution from MIT. The out-of-the-box interface failed their usability testing. Fortunately, the user interface of DSpace proved to be easy to modify.

Learning from the experience at MIT (MIT's Institutional Repository had international press coverage, but they had to hire a marketing consultant to get their own faculty to use it), the University of Rochester hired an anthropologist to study the faculty--each discipline as a separate "tribe." With the information from the study, Rochester was able to insert the Institutional Repository into the workflow of the faculty members.

Each faculty member has his or her own page in the repository, since the anthropologist discovered that "faculty are like peacocks." Faculty also felt that an instutional repository, with its rather unfortunate label, was too impersonal. By giving each faculty member the power to put up pictures, personalize their own page, and organize the information on their page the way they saw fit, the repository became less "institutional" and more personal.

Another big issue (and this was a theme in several presentations) was that faculty members don't know if they can put their material in the institutional repository. They don't read the publisher contracts. Susan Gibbons found that professors will defer to the library on copyright issues. So it's up to the library to determine what can go into the repository. When an article has successfully been submitted into the repository, the library staff asks the author for his or her CV so they can check the Sherpa site and the publisher contracts to see if other publications can be deposited.

Susan Gibbons recently co-authored an article on Institutional Repositories for D-Lib Magazine.

Georgia Institute of Technology
Catherine Jannick discussed SMARTech, Georgia Tech's Institutional Repository. This implementation also uses DSpace. They have yet to change the user interface, but intend to do so.

Populating SMARTech began with the idea that the faculty would submit their own work. When this didn't happen, the library harvested digital output from campus units and other repositories and, like Rochester, began using faculty CVs as an index to potential content for the repository. The metaphor that they currently work under is that the faculty has always given their books to the library, which the library catalogs, classifies, and makes available. Now, the faculty members "give" articles to the library and the library does the work of getting them into the Institutional Repository.

"SMARTech is not the only repository at Georgia Tech." For this, the appropriate metaphor is of branch libraries and learning centers. Each collects material that is appropriate to the purpose, with duplicates as needed. The main library sees its collection in more long term ways and selects material that is worthy of preserving. SMARTech is the equivalent of the main library, collecting repository material that should be saved for the future.

Cornell University
Cornell University also has several Institutional Repositories, one is in DSpace.

Cornell has made an effort to teach faculty about scholarly communication issues through department talks and symposia. They direct faculty to "green publishers" as identified by Sherpa that allow authors to publish pre-prints and/or post-prints. They point faculty to the SPARC page about retaining authorial rights by amending the contract--a process the faculty members are surprised to discover that publishers accept with no argument.

DPubs is a repository for technical reports. The Computer Science department administers DPubs although the library provides the servers.

DSpace is a repository that has gone beyond technical reports to include student projects, books, periodicals, films, and electronic theses and dissertations. At Cornell, they have had few problems with self-submission of these resources. DSpace is where Cornell is exploring Open Access publishing ventures.

arXiv.org is a disciplinary repository run by Cornell University Libraries for physics, computer science and mathematics. They receive over 200 submissions a day. They have found that the faculty are more enthusiastic about submitting to a disciplinary repository than an institutional repository--but they are finding a useful argument in "multiple repositories increase exposure." They encourage faculty to store material in disciplinary repositories, institutional repositories, and on their own websites.

University of Pennsylvania
Sandra Kerbel was schedule to present for the University of Pennsylvania, but instead we had Jean whose last name I didn't get. She began with a quote from Benjamin Franklin:

Who is wise? He that learns from everyone.

The Institutional Repository at Penn is not a DSPace implementation. Instead, they used Digital Commons from ProQuest.

The School of Engineering provided the original funding for the IR. Because of this, the library used the goals of the school to develop a strategy. The school wanted high-impact and visibility.

The library took an activist approach. Usually the faculty member was not contacted until the library had the article in hand and had checked that the copyright allowed for submission to a repository. Then, they simply informed the faculty member that the paper was going into the repository. Since the department heads had already informed the faculty members of the merits of depositing material, the library receives no objections from faculty.

In some instances, the library has been able to negotiate with publishers to include material even when the initial contract would not allow it. When this does not worked, the library informs the faculty member that his or her material will not be going into the repository and why. This has proven to be a "teachable moment" and faculty behavior has changed--they choose publishing outlets that will allow re-publishing in an IR, they retain rights that they used to let go, and they request rights that are not present in the boiler-plate contract.

Anecdotally, a professor at the University of Pennsylvania claims that he gets more downloads from the Penn repository than from either his own website or from a disciplinary repository. The faculty recieve a monthly report with the number of downloads of their articles. This encourages them to continue to submit material to the repository.

Using the marketing truism that it takes seven exposures to raise awareness, the library took multiple approaches to letting people know about the IR. They made brochures, took advantage of contacts with administration and faculty, and used coupons for the library café.

The library spends 18 minutes per paper checking if it is eligible for inclusion in the IR and developing appropriate metadata. Someday they may be able to implement a more self-service approach, but for now they feel that this service is what makes the IR work at the University of Pennsylvania.

Florida State University
Robert McDonald spoke about the Institutional Repository at Florida State. They began with breakfast focus groups which seemed to help bring key people on board, but he warned that they don't tend to stay on topic. "FSU opted for a hosted solution BePress." They felt that a fast start-up would allow them to build on the momentum from the focus groups and other early PR work.

A large part of FSU's motivation was to do research and development in digital preservation issues. This included developing metadata and metadata crosswalks.

They started by collecting student content from two undergraduate programs as well as electronic theses and dissertations. As of the fall of 2005, they had all FSU dissertations from 1951 to the present in the repository and all masters theses from 2003 to the present. FSU has each student sign an agreement about copyright, based on a Creative Commons licensing model.

Using the student content, FSU has been able to test several systems and build an infrastructure. Now they want more faculty content. They are looking at ways to "tie-in with Student and Faculty Portfolios."

Questions
The questions at the end were about access (Note to self, try OAIster), submission processes, and work flow. DSPace has (or had at one time) a five page/step process for submission. Cornell consolidated this to two pages--one for the license to give the repository permission to hold the work and one for the metadata. I forget which one answered the work flow question but they said that student workers do the digitizing when the work is not submitted in digital form and that the email account, which is a unique identifier, is the key to the authority file for authors' names.     #

June 2004 July 2004 August 2004 September 2004 October 2004 November 2004 December 2004 January 2005 February 2005 March 2005 April 2005 May 2005 June 2005 July 2005 August 2005 September 2005 October 2005 November 2005 December 2005 January 2006 February 2006 March 2006 April 2006 May 2006 June 2006 July 2006 August 2006 September 2006 October 2006 November 2006 December 2006 January 2007 February 2007 March 2007 April 2007 May 2007