Literature Review

A short introduction featuring institutional repository (IR) definitions (Appendix F1) and early IR projects (Appendix F2) sets the stage for a literature review mirroring the topics on which MIRACLE Project questionnaires queried census respondents: (1) the extent of involvement with IRs at academic institutions (Appendix F3); (2) the people involved with IRs (Appendix F4); (3) the budget for an IR (Appendix F5); (4) important investigative activities leading up to IR implementation (Appendix F6); (7) IR-system software, content, and features (Appendix F7); (8) IR practices and policies (Appendix F8); (9) IR benefits and beneficiaries (Appendix F9); and (10) institutions where no IR planning has been done (Appendix F10).

F1        IR Definition and Attributes

Branin (2005) traces the origins of IRs to the knowledge management (KM) movement. He cites the impact of Peter Drucker’s seminal Harvard Business Review paper “The Coming of the New Organization,” which proposes that knowledge is the most important asset of the modern organization. Drucker does not define “knowledge” in this article but instead says that it “manifest[s] itself in various forms … patents, trade secrets, operational routines, expertise inside the heads of employees” (Branin 2005, 237). Branin (2005, 238) sums up by saying the “KM movement in the 1990s established the nomenclature of IRs, lit a fire under organizations to pay attention to knowledge assets, and forced organizations to define knowledge assets broadly.”

The IR may also be a technology-based solution to the problem of escalating serials costs and reform of the scholarly communication process. Abstract models, guidance, and technology standards that have sparked the development of IRs and ensured their interoperability with other online systems are the Open Archival Information System (OAIS) Reference Model, the Open Access Initiative’s Protocol for Metadata Harvesting (OAI-PMH), and the Metadata Encoding and Transmission Standard (METS) (Branin 2005).

Not unlike IRs are open-access e-print and preprint repositories that scientific and scholarly disciplines have promulgated for the dissemination of research within disciplines and subdisciplines. Instead of the discipline-based focus that underlines e-print and preprint repositories, IRs feature the intellectual output of educational and research units in universities, research institutes, and state and federal governments, and thus are more likely to be encyclopedic in subject coverage, representing the full range of academic fields of study, especially IRs at universities, and including a wide range of document genres, e.g., working papers, simulations, specimens, and course syllabi.

Clifford Lynch (2003) defines IRs and enumerates their attributes:

An institutional repository is a set of services that a university offers to the members of its community for the management and dissemination of digital materials created by the institution and its community members. It is most essentially an organizational commitment to the stewardship of these digital materials, including long-term preservation where appropriate, as well as organization and access or distribution … It represents a collaboration amongst librarians, information technologists, archives and records managers, faculty, university administrators, policy makers. An institutional repository will be supported by a set of information technologies, but a key part of the services … is the management of technological changes, and the migration of digital content from one set of technologies to the next as part of the organizational commitment to providing repository services.

F2        The First IR Systems, IR Projects, and IR-related Projects

Until there was a critical mass of institutions with operational IRs to survey, researchers occupied themselves by describing the first IR systems, IR projects, and IR-related projects. Examining open-access activity in the United Kingdom (UK), Pinfield (2003) focused on the FAIR (Focus on Access to Institutional Resources) Programme of the Joint Information Systems Committee (JISC). Shearer (2003a, 2003b) published periodic updates on IR efforts of libraries in the Canadian Association of Research Libraries (CARL). A report on the first 45 IRs and IR-related projects included an examination of publishers’ responses to an online questionnaire about their attitudes toward IRs (Mark Ware Consulting 2004; Ware 2004a).

A key step forward for IRs in the United States was funding from The Andrew W. Mellon Foundation for the joint development of the open-source DSpace IR-software system by the Massachusetts Institute of Technology (MIT) Libraries and Hewlett-Packard Company (Smith et al. 2003). Since 2003, institutions with the requisite technical expertise have been able to download DSpace and customize it as needed for experimentation, pilot testing, or public implementation.

Another IR-system software package that has benefited from extensive foundation funding is Fedora (Flexible Extensible Digital Object and Repository Architecture). Fedora’s origins can be traced to Defense Advanced Research Projects Agency (DARPA) and National Science Foundation (NSF) funding at Cornell University in 1997. Developed jointly by Cornell University Information Science and the University of Virginia Library with grants from The Andrew W. Mellon Foundation, Fedora is a general-purpose digital repository service. Fedora applications are IRs, library collections management, multimedia authoring systems, archival repositories, and digital repositories (Fedora Project 2005–2006). Institutions that download Fedora’s open-source software will need technical expertise to customize it.

F3        IR Surveys and the Extent of Survey Respondents’ Involvement with IRs

In 2005, researchers began surveying educational institutions regarding their efforts with IRs. Lynch and Lippincott (2005) distributed their survey to the 124 member institutions of the Coalition for Networked Information (CNI) and to 81 liberal arts colleges bearing CNI consortial membership. They queried respondents about the extent of IR implementation at their institutions, database size, content, system software packages, policies, and administrative responsibility for the IR. Response rates were high.2% for CNI members and 43.8% for CNI consortial members.

About 40% of CNI members have “some type” of IR operating, and 88% of those who do not have an operational IR have undertaken IR planning. Only 6% of CNI-consortial members have an operational IR. Lynch and Lippincott (2005) note that “deployment of institutional repositories beyond the doctoral research institutions in the United States is extremely limited … Most of the engagement … is at colleges and universities where students and faculty have strong commitments to locally created materials for teaching and learning or that document student research.” They suggest two factors that might diversify the types of institutions involved with IRs: (1) large-scale student contributions in the form of electronic portfolios (e-portfolios) to IRs; and (2) a groundswell of faculty demand for IR services to facilitate the dissemination of their research and teaching objects.

Conference organizers from CNI, the JISC in the UK, and SURF Foundation in the Netherlands conducted an international survey to determine the current state of IR deployment in the academic sector (van Westrienen and Lynch 2005). Survey results reveal the extent of IR implementation by country, database size and content, system software packages, contributor activity, policies, and factors that stimulate and inhibit IRs. The researchers estimate “a spread from around 5% in a country like Finland, where repositories are just getting started, to essentially 100% deployment in countries like Germany, Norway and the Netherlands, where it is clear that repositories have already achieved some status as common infrastructure across the relevant national higher education sector and, hence, can form the basis for other initiatives that presuppose the near-universal availability of institutional repositories.”

Shearer’s (2004) informal telephone survey of member libraries in the CARL reveals the extent of their involvement with IRs, system software packages, content, recruiting contributors, and working with early adopters. Of the 20 universities she contacted, 25% have operational IRs and 35% are in the process of implementation. The remaining 40% are in the planning stages.

A team of University of Houston researchers surveyed the 123 member libraries of the Association of Research Libraries (ARL) regarding their IR efforts and published findings in an ARL SPEC Kit (Bailey et al. 2006). Completing the survey were 87 (71%) ARL libraries. Survey content was comprehensive; it queried ARL members about the extent of their involvement with IRs, people involved in the IR effort, budgets, policies, recruiting content, system software packages, benefits, and evaluating IRs. Of the 87 respondents, 37 (43%) have implemented IRs, 31 (35%) are planning for IRs, and 19 (22%) have no plans for IR involvement. The survey researchers note that perceptions of staff planning for an IR do not always correspond with experiences of staff who have implemented an IR, particularly concerning time and resources required to implement an IR and the level of difficulty in recruiting content.

F4        The People Involved with IRs

At the heart of Lynch’s (2003) definition of an IR is a “collaboration of librarians, information technologists, archives and records managers, faculty, university administrators, and policy makers.” Pelizzari (2005) advocates the library as “the standard bearer for the IR.” Quint (2002, 8) urges librarians to assert themselves as the leaders in IR implementation. She writes, “This is the best chance librarians will ever have to break the chains that have bound them and their budgets … Who will step up and help to create a better process of scholarly communications? … If academic librarians do not step up to pay that price and right now, they could find themselves blocked out of that future, and perhaps, any future at all. Now or never.” A survey of CNI members and consortial members reports that 80% of institutions with operational an IR have put sole responsibility for the IR into the hands of the library.

The ARL SPEC Kit survey results reveal that the library plays a critical role in initiating, planning, and implementing IRs (Bailey et al. 2006). “All respondents, implementers and planners alike, indicate that the library has been a driving force in the creation of or planning for an IR” (Bailey 2006, 14). Following a successful IR implementation, the library department that manages digital library initiatives usually takes charge of IR operations and collaborates with other library departments such as technical services, archives, and/or cataloging. The ARL SPEC Kit study reports IR staffing unusually high numbers (Bailey et al. 2006, 15):

If the mean FTE [full-time equivalent] values for each of the four units are added together, the average number of staff working on an implementer’s IR is 28.1. The breakdown by staff category is 7.4 librarians, 7.3 other professional staff, 9.5 support staff, and 3.9 students. The average number of staff working on a planner’s IR is 61.2. The breakdown by staff category is 8.8 librarians, 20.8 other professional, 22.2 support staff, and 9.4 students.

Gibbons (2004, 17) advocates a partnership of librarians, instructional technologies training staff, and computer services staff for the IR effort. Ware (2004a) asserts that IRs start with a partnership between the library, the institution’s instructional technology unit, and a vice president’s office (e.g., bursar or provost for academic studies).

IR staff at institutions where IRs have been implemented have been eager to tell their story (e.g., Phillips, Carr, and Teal 2005; Jones, Andrew, and MacColl 2006, 159–189; Chan, Kwok, and Yip 2005; Rogers 2003). The IR effort usually involves central administrative, academic, and service units. For example, Hewlett-Packard Labs and MIT software engineers, MIT librarians, faculty, and administrators, and early adopters were involved in the development and deployment of DSpace (Baudoin and Branschofsky 2003). Case studies generally present a range of experiences such as IR-system software development, system implementation and deployment, policy creation, collection development strategies, and preservation initiatives.

Librarians have been enthusiastic about carving out a role for themselves vis-à-vis the IR. Allard, Mack, and Feltner-Reichert (2005) identify librarians in these roles: learning how the IR works; leading the IR implementation effort; developing policy (especially with regard to defining collections); leading anything that pertains to metadata, reviewing submissions to the IR; and training contributors. Additional roles include evaluating IR systems, being an advocate for the IR, recruiting content, and serving as advisory contributors on intellectual property issues (Chan, Kwok, and Yip 2005).

Archivists are more wary of their relationships with the IR. Crow (2002a) observes this phenomenon.

Depending on the university, an institutional repository may complement or compete with the role served by the university archives. University archives often serve two purposes: (1) to manage administrative records to satisfy legally mandated retention requirements, and (2) to preserve materials pertaining to the institution’s history and to the activities and achievements of its officers, faculty, staff, students, and alumni. Compared to institutional repositories, which aim to preserve the entire intellectual output of the institution, university archivists exercise broad discretion in determining which papers and other digital objects to collect and store. Still, the potential overlap of roles of the two repository types merits consideration at institutions that support both.

Crow’s custodial and passive view flies in the face of recent efforts of archives to become more involved in the research and teaching missions of colleges and universities. He also downplays the difficulties in documenting all the intellectual output of an institution and the benefits of a selection process in the IR collection development plan. Arguing for selection of faculty output, Bicknese (2003–2004) posits that archivists bring to the table substantial expertise in collecting and appraising faculty papers as well as other university records and recommends that IRs use this expertise.

F5        The Budget for an IR

The single largest line item in the budget for an IR is staffing (Gibbons 2004, 54). Technical staff, at the very least, a systems administrator and a programmer, are needed to profile, program, maintain and customize IR-software systems. Jones, Andrew, and MacColl (2006) advocate an IR staff composed of systems developer, a liaison officer, a project manager, and a metadata editor. Librarians should be responsible for recruiting content from faculty, students, and other members of an academic institution’s learning community. Costs for software, equipment, and backup equipment are minor compared with staffing costs.

Ballpark figures for annual costs of the IR effort are $285,000 at MIT (Barton and Walker 2002), $265,000 at The Ohio State University (OSU) (Branin 2005, 247), and $200,000 at the University of Rochester (Gibbons 2004, 56). The ARL Spec Kit reports these costs: (1) for institutions planning an IR, a median of $75,000 and a range from $12,000 to $160,000; (2) for institutions implementing an IR, a median of $45,000 and startup costs ranging from $8,000 to $1,800,000; and (3) for institutions that have implemented an IR, a median of $41,750 and operations costs ranging from $8,600 to $500,000 (Bailey et al. 2006). “The primary method of funding both start-up and ongoing costs is reallocation from existing budgets” (Bailey et al. 2006, 16).

At MIT, Barton and Walker (2002) break down costs into three major line items: (1) staff salaries and benefits, (2) operating expenses, and (3) system equipment escrow. Although MIT is unique because of its development of the DSpace IR, its funding model may inform other institutions contemplating local development. The MIT model anticipates financial support and in-kind assistance from four sources: (1) the MIT Libraries’ operating budget, (2) collaborative development of DSpace that results from related projects that deploy DSpace, (3) enhancement of DSpace by DSpace Federation members at whose institutions DSpace is deployed, and (4) charging for premium services.

“One of the likely largest costs over the long term will be … preservation … also by far the least known and indeed least knowable … [s]o a commitment to an IR amounts to an implicit commitment to an unknown amount of work at some point in the future” (Mark Ware Consulting 2004, 24). Strategies for recovering costs include depositing funds into escrow accounts in anticipation of the future costs of preservation, hosting digital content submitted by partner institutions, and charging for ancillary services such as file conversion, digitizing physical artifacts, metadata creation, and exceeding file-storage quotas (Gibbons 2004, 56; Barton and Walker 2002). Branin (2005, 247) asks rhetorically, “What will it cost an institution or society not to provide stewardship of its important digital assets?”

F6        Important IR Investigative Activities

Institutions beginning an IR project have two comprehensive sources to consult. From the LEADIRS Workbook (Barton and Waters 2004–2005), they will learn about planning an IR, choosing IR-system software, developing policies, anticipating costs, and recruiting content. The workbook features worksheets that institutions can use to guide them through IR planning and implementation, and it links to online examples of planning and implementation documents, e.g., service definitions, promotional brochures, needs assessment surveys, operational IRs, and policy statements. The second resource, The Institutional Repository, provides “a comprehensive outline of the main issues to consider when setting-up and developing an institutional repository—from making the case within the institution and choosing suitable software to formulating workflows, policy, and advocacy plans” (Jones, Andrew, and MacColl 2006, xvi).

Investigative activities that IR committees conducted prior to implementing OSU’s Knowledge Bank IR include a data-warehouse planning project to identify administrative data needed for decision support, an environmental scan of operational and in-development IRs, discussions with representative faculty groups, an inventory of on-campus digital initiatives, a review of relevant technology standards, and a pilot-project compilation of faculty publications (Rogers 2003; Baudoin and Branschofsky 2003; Walters 2006).

F7        IR-system Software, Content, and Features

Gibbons’ (2004) article in Library Technology Reports is a comprehensive examination of IR systems and system features. Institutions approaching pilot-test and implementation phases of the IR effort would benefit from a periodic update of her article, especially now that several for-profit vendors have entered or are about to enter the marketplace with new IR systems. Barton and Waters’ (2004–2005) LEADIRS Workbook also lists available IR software systems and uniform resource locators (URLs) that link to systems in operation. Smith et al. (2003) and Tansley et al. (2003) restrict themselves to the design and functionality of the open-source DSpace IR.

Because those who responded to the CNI survey count IR content in different ways, Lynch and Lippincott (2005) do not make an effort to estimate the size of operational IRs. The types of materials CNI members submit to their IRs run the gamut, representing the research, teaching, service, and publicity activities of colleges and universities, and ranging from text-based items to multimedia artifacts—images, audio, video, software, blogs, and e-portfolios. The majority (58%) of CNI respondents have implemented DSpace; other popular systems are bepress, ContentDM, Virginia Tech’s ETD, DigiTool, and locally developed systems. Most CARL institutions are using DSpace (Shearer 2004). CARL IRs feature a wide variety of artifact types, but the rate of deposit is slow and sporadic (Shearer 2004). DSpace is also the most commonly used IR-system software by ARL-member libraries that are planning IRs or operating IRs (Bailey et al. 2006). Some ARL libraries that have implemented an IR are using DSpace in conjunction with other IR-system software, such as ContentDM or ETD-db. The few ARL implementers who are not DSpace institutions have chosen ContentDM, ProQuest Digital Commons, and Archimède.

Content in the IRs of the 13 nations participating in the international survey sponsored by CNI, JISC, and SURF is divided between journal articles and theses; the sole exception is Australia, where 88% of content is primary data (van Westrienen and Lynch 2005). Institutions in 11 of the 13 countries are using DSpace, and institutions in 7 of the 13 countries are using GNU ePrints software. Institutions in Germany favor OPUS while those in Australia favor Virginia Tech’s ETD. Discussing the size of Open Archives Initiative (OAI) repositories, ePrints archives, and IRs, Mark Ware Consulting (2004, 33) concludes that “the majority [of IRs] are clearly in a very early stage of development but even more of the longer-established sites have a relatively small number of documents compared to the research outputs of their institutions.”

F8        IR Practices and Policies

Devising IR policies is an important, necessary, and complex activity during IR implementation. A survey of CNI members and consortial members reports that 60% of institutions with an operational IR put sole responsibility for IR policy making in the hands of the library. Laundry lists of policy issues are given by Mark Ware Consulting (2004), Barton and Waters (2004–2005), and Shearer (2005). In preparation for the Library and Information Technology Association (LITA) Regional Institute “Establishing an Institutional Repository,” Gibbons (2005) compiled a list of links to the IR policies of nearly 20 universities worldwide. Another source of online IR policy data is a section of the ePrints Web site called ROARMAP (Registry of Open Access Repository Material Archiving Policies) that invites institutions to record their commitment to providing open access to peer-reviewed research and to share their departmental or institutional policies regarding open-access provision (ePrints, n.d).

F8.1     Digital Content for the IR

Gibbons (2004, 21–26), Lynch (2003), and Crow (2002a) describe potential digital content for the IR that results from faculty who are active researchers and teachers (e.g., e-prints, working papers, technical reports, conference presentations, data sets, and learning objects) and students who do research and want to document their academic accomplishments (e.g., theses, dissertations, data sets, and e-portfolios). IRs can be exacting about acceptable types of materials. For example, MIT’s DSpace allows only publisher-accepted materials, and the University of California’s eScholarship accepts only materials that authors have not submitted to publishers (Shearer 2002).

F8.2     Making Contributions to the IR

Contributing content involves metadata capture, file management, and license handling (Jones, Andrew, and MacColl 2006). Policies can specify acceptable metadata formats, required and optional metadata tags, and people who are authorized to submit, review, and update metadata. File management pertains to checking file formats, data integrity of files, and the completeness of documents.

Copyright infringement is one of the important challenges that IR staff face. When submitting manuscripts to publishers, most academics sign copyright-transfer agreement forms that assign copyright to publishers. Surveying 542 academic authors, Gadd, Oppenheim, and Probets (2003) report that 49% reluctantly assign copyright to publishers and 41% do so freely. Surveying 1,200 authors who publish in medicine and the life sciences, Hoorn and van der Graaf (2006) report that 71% prefer to keep copyright, 2% prefer to transfer copyright to publishers, 23% are neutral about the choice between the two, and 4% do not know. A little less than half of their survey’s respondents want to keep copyright, allow unlimited reuse of published works for scholarly and educational purposes, and put limitations on reuse for commercial purposes.

So many authors assign copyright to publishers that, for the time being, IR staff may have to be wary of IR contributors infringing on copyright and advise them about copyright issues. If authors assign the copyright for their works to publishers, IR staff should advise authors to scrutinize their copyright agreements with publishers to make sure they retain the right to publish the work or the last prepublication version electronically (Barton and Waters 2004–2005).

F8.3 Preservation

Long-term preservation of scholarly content is an essential role of IRs (Lynch 2003). Not a one-time event, long-term preservation consists of specific functions such as ingesting digital objects in which metadata is created, storing such objects and associated metadata, monitoring technology obsolescence, and evaluating the usage of digital objects (Fyffe et al. 2004).

At the moment, IRs have not become the equivalent of trusted digital repositories “whose mission is to provide reliable, long-term access to managed digital resources to its designated community, now, and in the future” (RLG 2002). Trusted digital repositories are responsible for long-term maintenance of digital objects with organizational, technical, and financial sustainability (Jantz and Giarlo 2005). The RLG-OCLC conception of trusted digital repositories requires a certification process, but the degree to which libraries would pursue this type of certification is not known. Certification would impose a very different type of accountability—an external accountability—on libraries, something that they have not previously experienced (RLG 2002).

Many IRs employ shared standards such as the Open Archival Information System (OAIS) reference model and the Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH). The OAIS model provides a comprehensive framework of all the functions required for digital preservation including ingest, storage, retrieval, and long-term preservation of digital objects. OAI-PMH is a mechanism for harvesting extensible markup language (XML)-based metadata from repositories, and it therefore makes possible interoperable search and retrieval among repositories (Branin 2005). Implementation of digital preservation in IRs, however, is still in its infancy.

Preservation concerns of libraries and institutions of higher education are also shared by IR contributors and users. Swan and Brown (2004a, 2004b) note that 42% of open-access (OA) journal authors are worried about preservation of OA journal contents. Rowlands et al. (2004), however, state that their survey’s respondents disagreed with the idea that OA journals are ephemeral. Although the concern about preservation “should not be a good reason for eschewing open access (indeed, all online only) journals” (Swan and Brown 2004b, 67), preservation of open-access contents, as well as the larger issue of institutional repository contents, is an important issue that needs to be investigated in greater depth (Kim 2006).

Although no consensus on best practices exists, digital preservation has these key functional goals: (1) data can be maintained over time without being lost, damaged, or altered; (2) data can be found and served to users; (3) data can be interpreted and understood by users; and (4) goals 1, 2, and 3 can be fulfilled in a long-term plan (Wheatley 2004).

From a policy standpoint of preservation, identifying file formats for which IRs provide long-term preservation is necessary. When making decisions about preserving file formats, Jones, Andrew, and MacColl (2006, 80) suggest IR staff answer these questions: (1) Is the file format an open standard/format? (2) Is the file format widely used? (3) Is the file format and associated technology likely to be preserved? (4) Are the contents of the file human readable? and (5) Is the file format itself human readable? These authors assert that preserving formats to which IR staff answer positively to questions 4 and 5 is best.

The ARL-sponsored survey reports that 74% of ARL libraries’ operational IRs accept any digital file type. “Relatively few (26%) are committed to functional preservation of every file type” (Bailey et al. 2006).

F8.4     Recruiting IR Content

Recruiting content for IRs has been difficult. Except in the Netherlands and Germany, the number of academics contributing to or knowledgeable of IRs is very low (van Westrienen and Lynch 2005). A year after the IR at the University of Toronto Scarborough became operational, contributors had submitted hardly 1,000 items to it (Chan 2004, 287). Ware’s (2004b) survey of 45 early IRs reports an average number of documents of 1,256; he qualifies the largest IR at the University of Virginia (21,000 items) because two-thirds of its contents come from a digital photograph collection. “Difficult” sums up the response of about two-thirds (63%) of ARL libraries regarding the recruitment of digital content for the IR (Bailey et al. 2006).

Examples of strategies that IR staff have enlisted to boost IR submission rates are hiring work-study students to do proxy archiving for faculty, giving presentations at faculty meetings, demonstrating operational IRs, working with early adopters and relying on them to tell their colleagues and students about the IR, long-term archiving of locally edited e-journals, and encouraging local Web site developers to relocate digital content from their sites to the IR (Crow 2002b, 23; Gibbons 2004, 57; Chan 2004; Shearer 2004, 2005; Mackie 2004; Barton and Waters 2004–2005; Jenkins, Breakstone, and Hixson 2005; Graham, Skaggs, and Stevens 2005; Chan, Kwok, and Yip 2005; Bell, Foster, and Gibbons 2005; Bailey et al. 2006). Because the evidence in favor of higher citation rates for open-access material is mounting (Odlyzko 2000; Lawrence 2001; Swan and Brown 2004a; Antelman 2004; Eysenbach 2006a, 2006b), scholars may be inclined to bypass traditional publication channels and publish instead in open-access repositories in the not-so-distant future.

On the basis of the results of an ethnographic study, Foster and Gibbons (2005) take a radically different approach to recruiting IR content from faculty at the University of Rochester. They reconfigure DSpace so that faculty contributions take place in a personal Web page or curriculum vitae (CV) that features vital information about contributors such as name, title, and contact information, and archives in the IR their research and teaching output (e.g., course syllabi, conference presentations, working papers, preprints).

F9        IR Benefits and Beneficiaries

Some IR advocates envision IRs as the critical component in reforming the existing system of scholarly communication. Researchers would have 24–7 access to scholarship, full-text search and retrieval capacity, and the ability to link backward and forward in time using footnotes and citations (Harnad 2001b). Academic institutions would reap benefits such as the following (Crow 2002a, 2002b; Chan 2004):

•          a new scholarly publishing paradigm that wrests control from publishers and puts it back in the hands of the academy

•          increased visibility, prestige, and public value

•          maximal access to the results of publicly funded research

•          increased numbers and diversity of scholarly materials that are collected and preserved

In the ARL Spec Kit Survey (Bailey et al. 2006), the top three reasons ARL libraries give for implementing IRs are institution centered—to increase global visibility of, to preserve, and to provide free access to the institution’s scholarship.

Many barriers prevent authors from fully embracing self-archiving. Examples are authors’ lack of knowledge about self-archiving, questions about whether self-archiving will erode the quality control that results from peer review and will make authors’ works more susceptible to plagiarism, the time required to submit documents to IRs, technical difficulties that authors encounter while making contributions, the possibility of infringing on signed copyright agreements, authors’ concerns about the equivalence of their institution’s IR and a journal publisher, distrust of their institution’s commitment to long-term maintenance, and a reluctance to tamper with the current working system of scholarly communication (Harnad 2001a; Chan 2004; Pinfield 2004, 2005; Swan and Brown 2004a, 2004b; Swan 2005b; Ober 2005; Jenkins, Breakstone, and Hixson 2005; van Westrienen and Lynch 2005).

Pelizzari (2005) voices his concern: “The biggest obstacle may be inertia … amongst academics … [the problems of] intellectual property rights, quality control, workload (their own), undermining the ‘tried and tested’ publishing status quo on which academic reputations and promotions lie.” About 70% to 80% of authors admit that they would comply with mandates from employers or funders regarding submission to an open-access archive (Swan 2005b; Swan and Brown 2004b). Harnad et al. (2004) are vocal advocates for mandates or incentives from employers and funders. Carr and Harnad (2005) report the results of studying two months of submissions into a mature repository, estimate the amount of time spent entering metadata would be as little as 40 minutes per year for a highly active researcher, and conclude that self-archiving is not a time-consuming task for authors or their designees.

In the ARL SPEC Kit survey (Bailey et al. 2006), most IR implementers perceive content recruitment to be difficult; half of the planners are neutral and the other half is divided between “easy” and “difficult.” This finding reveals a change in perception regarding content recruitment between IR planning and IR implementation. IR implementers employ especially diverse and aggressive content-recruitment strategies, digitizing print materials and depositing them in the IR for authors and holding symposia to raise awareness of IRs among prospective authors.

F10      Institutions Where No IR Planning Has Been Done

To date, most of the information we have about IRs targets IR implementation especially at large research institutions. Little is known about institutions that are in the IR preplanning or planning stages or about the experiences of institutions that have a teaching focus. CNI, CNI-JISC-SURF, CARL, and ARL surveys report on the numbers and percentages of such institutions (see Appendix F3), but they report little else about them.