Opening Up Open Access Institutional Repositories Opening Up Open Access Institutional Repositories to Demonstrate Value: Two Universities’ Pilots on to Demonstrate Value: Two Universities’ Pilots on Including Metadata-Only Records Including Metadata-Only Records

INTRODUCTION Institutional repository managers are continuously looking for new ways to demonstrate the value of their repositories. One way to do this is to create a more inclusive repository that provides reliable information about the research output produced by faculty affiliated with the institution. DESCRIPTION OF PROGRAM This article details two pilot projects that evaluated how their repositories could track faculty research output through the inclusion of metadata-only (no full-text) records. The purpose of each pilot project was to determine the feasibility and provide an assessment of the long-term impact on the repository’s mission statement, staffing


INTRODUCTION
Institutional repository (IR) managers, including the authors of this paper, understand the continuing need to demonstrate the value and purpose of repositories within their institutions. IR managers continue to strive to provide value by focusing on services, such as faculty-assisted submissions (FAS), 1 that attempt to increase the number of open access (OA) full-text (preprint, postprint, or publisher's PDF) articles in their repositories ( Figure  1). These services can be time intensive and have little to no impact on providing a more comprehensive record of the institution's scholarly output.
1 FAS provides mediated review and deposit of previously published works to an IR. Faculty send their CVs to library staff who research the copyright permissions for each publication and deposit what they can into their IR. jlsc-pub.org eP2220 | 3 Under the full-text-only model, there are many items that are discovered in publication harvests that IR managers are not able to load into their platforms. A common reason that full-text content cannot go into the IR is that a majority of publisher policies only allow either a pre-or a postprint copy, instead of the publisher PDF, of an article to be archived in an IR. Obtaining pre-or postprint copies of articles can be time consuming for staff when such versions are available at all. Copyright concerns can either be a hurdle or a total barrier to representing institutional works within an IR. Projects like the DAEDALUS, an IR advocacy program based out of the University of Glasgow, succeeded in adding new content to its IR but was "offered significant amounts of content that cannot be added because of restrictive publisher copyright agreements" (Mackie, 2004). To further complicate the issue, some faculty authors on the respective campuses would like to see their content included, but do not feel comfortable adding versions other than the publisher version. Faculty from the authors' universities have expressed that only publisher-formatted content is acceptable for uploads into the IR and have formally opted out of participation with any other formats. Additionally, some faculty have elected to participate in OA on their own through personal websites or discipline-level archives. These faculty have resisted having content stored in multiple places and instead support linking metadata records to the external content to represent works within an IR. There may also be older content that is not available in a digital format, or yet-to-be published items that do not presently have full-text versions or any digital content (data, video, image, etc.) that can be uploaded to the IR, but faculty may want to let scholars and researchers know that these works were produced. Because some content cannot be uploaded due to publisher restrictions, author preferences, or various other limitations, the only remaining ways to represent this content in an IR is through metadataonly or metadata-with-link records ( Figure 2). These types of records represent a potential way to provide access through hyperlinking and documentation of institutional activities.
IR managers are looking to redefine their repository services of collecting, organizing, curating, and preserving the scholarly record of the institution (Lynch 2017). This paper presents pilot projects that were conducted to broaden and rethink the scope and mission of an IR to provide a more comprehensive record of the scholarly output of the institution and enhance the repository's range of services. The authors sought to determine the long-term impact on the repository's mission statement, resources (including staffing), and collection development policies that would result from the inclusion of metadata-only records in the IR and the associated workflows. To determine the workload requirements and impact of this change, pilots were performed at the authors' universities to add or simulate metadataonly records or metadata records with links to publisher and/or catalogued content to the IR. It is important to note that even with this approach librarians would still make efforts to include as much full-text content as possible. The IR managers and supporting teams felt that faculty and their universities would benefit from this adjustment by having a more jlsc-pub.org eP2220 | 5 complete record of their scholarship. All content loaded into the IR would remain open access, but users might find records that are "dead" (no content or links) or with links that lead to locked content, meaning that researchers may hit paywalls when the library does not subscribe to the journal in which the article was published. In addition to its impact on faculty and end users, this new process will have two major implications for the respective IRs. First, a larger portion of scholarly output could be represented in a central location for each institution. Second, an IR might decline in rankings, depending on the rating system used to report. With a focus on highlighting the value and purpose of the IR to the institution, the authors determined that potential negative national rankings were worth the risk to increase the value of the IR for the institution.

LITERATURE REVIEW
The evolution and the ambiguity between definitions and scope of an IR stretch as far back as some of the first position papers. Johnson (2002) and Crow (2002) both focused on two defining rationales of institutional repositories, labeled the "New Scholarly Publishing Paradigm" and "Institutional Visibility and Prestige." The crux of both rationales, however, hinges on equating products of scholarly publishing with a university's collective intellectual capital (Johnson 2002) and the IR forming a mechanism to disrupt the scholarly publishing model. Clifford Lynch, director of the Coalition for Networked Information, proposed something different when he stated that "a mature and fully realized institutional repository will contain the intellectual works of faculty and students-both research and teaching materials and also documentation of the activities of the institution itself in the form of records of events and performance and of the ongoing intellectual life of the institution" (Lynch 2003b, p. 328). Lynch expanded on his views from 2003 in a 2016 interview, stating, "The point of IRs, in my view, isn't to disrupt the existing scholarly publishing system, but to allow it to be expanded and diversified by providing access and stewardship for material that mainly falls outside of the traditional scholarly publishing system as it exists today-both material created by faculty and material created by the institution, or departments or other groups within it" (Poynder 2006, p. 13).
The literature on the development and growth of the IR also represents the difficulty IR managers are facing when trying to get faculty to deposit their publications into an IR. Zhang, Boock, and Wirth (2015) revealed that "it has been widely reported in the literature that faculty don't self-archive in IRs without library mediation" (p. 3). Uncovering this difficulty revealed the struggles repository managers have with fulfilling the commitment to capture and showcase the research output of the university. Zhang et al. (2015) explored whether a university-wide OA policy would increase faculty deposits of their publications into an IR. This study showed the reluctance of faculty to deposit articles into the repository, even with an institutional-wide mandate. In fact, there was a slight decrease in the number of authors who deposited their articles in the IR in the year after the OA policy was passed. The study discovered that while faculty had "positive attitudes towards OA publishing, they are not so positive towards OA mandates" (Zhang et al., 2015, p. 1) and that "passing an OA policy alone is not a guarantee of increased faculty engagement in OA initiatives" (p. 9). An OA mandate, the literature reveals, also did not mean that faculty retained the permitted (pre-or postprint) versions of their publications. Hazzard and Towery (2017) discovered that "faculty retain nearly none of their pre-print versions of their published articles, and so we are unable to archive those titles in the repository. Most

| eP2220
Journal of Librarianship and Scholarly Communication faculty were unable to produce versions of their work (pre-prints) other than the publisher's PDF, which many publishers restrict from upload into a repository" (p. 9).
Bull and Schultz (2018) focused on a process to semiautomate the collection of new works of research produced by faculty to help IR ingest sustainability and to hedge against what they considered stagnation and inability to capture an institution's scholarly content. They presented an example of a workflow designed to aid IRs and associated services to transition "from a pilot phase or a post-pilot stagnation phase to the use of a fully operational metadata archiving service" (Bull and Schultz, 2008, p. 2). They envisioned that the service would be similar to a Current Research Information System (CRIS), except that full-text items would be added when possible. Like Zhang et al., they discovered that while "more of scholarly record is captured with this new workflow, problems remain, including a lack of full-text addition" (p. 16).
With challenges and dwindling resources, there is reason for IR managers to look to the formal melding of IRs and CRISes, which have appeared as add-ons to IR platforms for years (SelectedWorks and the Expert Gallery Suite being notable for Bepress's Digital Commons platform), as a potential path for the future. DSpace-CRIS, created by the partnership of University of Hong Kong and the Italian Interuniversity Consortium Cineca, represents a unique offering where an open-source IR has been altered to serve the dual purpose of IR and CRIS (Palmer, Bollini, Mornati, & Mennielli, 2014). Similar in nature, CRIS systems were developed to store and manage data about research conducted at an institution, while IRs traditionally pair scholarly and research records with corresponding digital objects. Writing on CRIS systems and repositories in the United Kingdom, Nicholas Joint (2008) states that the "highly specialized nature of research publications lists are ill-suited to take over the role of repositories, but several UK based institutions are happily using their single open access repository as a research publications system for research evaluation with complete success" (p. 573).
The inclusion of metadata-only records within an IR to better showcase institutionally created outputs represents just a stepping stone in the broader evolution of IR platforms. The push to include metadata-only records in an IR has been undertaken by a few institutions, in various geographic locations, and largely independent of any unifying strategy. This includes the pilots explored within this article, which were planned and administered individually. Speaking at the 11th International Conference on Open Repositories in 2016, representatives from King Abdullah University of Science and Technology presented on the development of a publications tracking process and its integration with their repository (Baessa, Grenz, & Wang, 2016). Part of that integration involved modifying the IR to deposit metadata-only records. "The success of these jlsc-pub.org eP2220 | 9 tracking and harvesting services in making our repository comprehensive and up-to-date has allowed us to rely on our repository as the key source of publications information for additional integrations that update ORCID records with publication information, populate a PlumX metrics dashboard and, most recently, support the implementation of a current research information system" (Baessa et al., 2016, p. 1).

Pilot Project Background
A common goal for both pilots was to create a more inclusive repository that would provide reliable information about the research output produced by faculty affiliated with the institution. The purpose of each pilot project was to determine the feasibility and provide an assessment of the long-term impact on the repository's mission statement, staffing, and collection development policies.
In 2016, Portland State University (PSU) administration approached library leadership to discuss a plan to study the feasibility of using the IR to track the scholarly output of the university. Prior to this, university administration had investigated faculty activity reporting software but had been unable to fully implement it on campus, seeing the IR as a potential solution. During the same year, the director of the Kansas State University (K-State) IR was pushing for increased inclusion of the scholarly output of the university within the IR. One way to accomplish this goal was to add records of all faculty outputs to the IR regardless of the format or ability to provide a full-text file. Although there was much hesitation and concern about including metadata-only records by library employees, the director and IR team were permitted to explore the issue through a pilot.
Both pilot projects sought to examine shifting the primary purpose of the university's IR away from being full-text, open access IR-neither university had an open access mandate, so library staff relied on faculty to participate voluntarily. The authors envisioned projects that would celebrate the research and creative works produced, build new relationships with faculty, and provide transparency of the scholarship created at the universities. At the onset of these pilots, both universities' IRs were open-access repositories with only full-text content, excluding minimal items with limited embargoes and campus restrictions; metadata-only records had not previously been included.
Public, ~24,000 students, high research activity IR Purpose PSU provides students, faculty, and staff with opportunities to openly share their work, both formally and informally. The IR offered the opportunity to share the knowledge that is created with disciplinary colleagues and with our global community at large. This value is present not only in the benefit to others who use that knowledge, but also in the recognition that scholars gain through the dissemination and impact of their work. To capture the institution's scholarly activities of all faculty output from the first quarter of 2015. PSU also included an approach to systematically research individual faculty through personal websites and Google Scholar citations. K-State, however, limited focus on harvests from third-party databases, Web of Science and Scopus, for their pilot.

PSU Pilot Process
The pilot at PSU was a short-term (two-month) project designed to capture "all" of the 2015 scholarly activities of the College of Engineering and Departments of Computer Science and History. PSU chose the time frame of February to March 2016 because it ensured that publications published at the end of 2015 would be included in the indexes. The College of Engineering and Departments of Computer Science and History were chosen for their diverse range of scholarly activity, with the expectation that they would provide a well-rounded project and enable assessment.
Citations were discovered and added to the IR through the following process: 1. Harvested Web of Science, Directory of Open Access Journals (DOAJ), Elsevier Journals, Google Scholar, arXiv, IEEE Xplore, Medline, PLOS One, PubMed, and systematically researched individual faculty websites and curriculum vitae.
2. Verified that the publications were not already in the IR.
3. Cleaned up problematic citations that were generated. Many of the citations harvested did not include dates or author affiliations. This process also included deduping citations harvested from the multiple sources.
4. Conducted name authority control utilizing our locally controlled authority structure.
5. Checked copyright status to determine what version could be uploaded; emailed authors asking for the correct version (preprint or postprint). The project team had mixed results with collecting the correct version from the authors and questioned the long-term sustainability of continuously reaching out to faculty with low response rates.
6. Added metadata using batch upload process. Part of this process included manually adding metadata to the spreadsheets for batch uploading.

| eP2220
Journal of Librarianship and Scholarly Communication

K-State Pilot Process
The pilot at K-State was also a short-term (three-month) project to determine the feasibility of being more inclusive with the research and creative outputs of the university that we include in the IR. The goal was to capture scholarly activities of any K-State faculty from January to March 2016 and demonstrate the increase in representation of this output. K-State chose this time frame to mimic their current practices of harvesting in quarterly intervals.
As it was uncertain whether this pilot would be put into practice, all pilot work was completed within a demo (nonproduction) instance of the IR. Each employee within the pilot process tracked the time used, negative/positive impact on current workflow, and recommendations for any changes. Like PSU for this project, the IR scope was expanded to include metadata-only records and links to publications that were not freely available due to copyright restrictions. However, the workflow used in the subsequent refinement and quality control of the records and metadata is, with few exceptions, the same as when records and items are uploaded into K-State IR in normal production. The fourth step, review, was conducted only to assess the value and success of the pilot project.
1. Harvested records from Web of Science and SCOPUS, assisted by each system's indexed affiliation search function, and removed all duplications. Used permissions verified using a script that queried the Sherpa/Romeo API (Flynn, Oyler, & Miles, 2013) to identify permissible publisher PDFs. All other items had their publisher URL added to the metadata records. Cleaned and enriched records with any missing information. Name authority control was conducted using a university-wide data set that provided all faculty names, affiliated departments, ranks, and hire dates.
2. Files prepared for ingestion were sent for final metadata quality control.
3. Transferred files for batch uploading into demo instance of D-Space.

Completed review of content by public services and other interested parties and gathered feedback.
Faculty status at K-State was determined with a match against a university wide data set that provided all faculty names, ranks, and hire dates. Faculty who do not match up to this data set are removed from our harvest. This may mean that some items are pulled that actually match our universities' affiliation through some other level or a new affiliation.
Harvesting a set publication date range allows us to pull in recent content only. Combining this workflow allows us to pull in content recently written by our current faculty. K-State will load content that faculty members published before coming to our institution; jlsc-pub.org eP2220 | 13 therefore, we do not screen out publications written before their hire date. Also, since we are pulling the recent quarterly publications, there is a slim chance that older content, not written during current employment, would be found in our harvest.

Pilot Process: Both Universities
In the majority of the cases the journals were the copyright owners, and therefore their policies were consulted for harvest workflow determination. Because IR employees initially harvested only publisher PDFs allowed by journal policy, neither institution consulted faculty, meaning we had no immediate struggles with identifying the correct version with authors and only uploaded the "perfect, polished" publications.

Outcomes
Both pilots succeeded in their primary goal of demonstrating that the IR could be used as the central location to showcase the scholarly record of the university. These efforts, although planned and coordinated separately, resulted in a shared 60% increase in representation of faculty-created content outputs (see Tables 2 & 4). Through metadata-only, metadata with links, and full-text records, faculty are able to display their entire body of works in a centralized location, leading to increased discoverability and marketability of their research and works. Both pilots also showed an increased demand on resources; however, the authors justify the allocation of these resources by • Addressing the silo problem across campus by increasing possibilities for interdisciplinary connections among individual authors • Centralizing discoverability and increasing visibility of university content as a whole Because the pilots were slightly different, we will review additional individual outcomes below.  • No longer requesting pre-and postprints; removing a large portion of hours used to communicate with authors to request publications and explaining versions • Using the process and assets created by Flynn et al. (2013) to partially automate permissions checking using the SHERPA/RoMEO API to identify published items where the publisher's version was allowed to be deposited. Performing manual permission checking to ensure compliance with journal copyright policies These adjustments have resulted in roughly 87% less resources required than in the old workflow. The biggest time savers implemented were scripts that automated a portion of the proposed workflow along with the elimination of a majority of correspondence with authors that was performed in the old workflow. K-State's old workflow included many of the same steps as PSU's pilot workflow, so it has been included here for time comparison. As shown in Figure 3, K-State's proposed workflow would lead to only a 2.5 hour-per-week increase over the current workflow.

Survey Responses
To gauge the success of the project, the pilot team at K-State decided to seek feedback on inclusion of metadata records in the IR from three groups of users: (1) faculty outside the library; (2) full-time employees of K-State Libraries; and (3) undergraduate and graduate students (see Appendix A). All members of the pilot team were certified in human research. The research project and associated survey were submitted for institutional review board (IRB) review; the research was declared exempt from review. For Group 1, rather than seeking feedback from all faculty at K-State, we chose to seek feedback only from those faculty who were discovered during the harvest phase and had records in Web of Science or in Scopus for items published between January and March 2015. We decided to seek feedback from every full-time employee of K-State Libraries (Group 2) and as many undergraduate and graduate students as possible (Group 3). Although we solicited student responses, we had no respondents from this category.
The majority of faculty in Group 1 had positive responses about including metadata content in the K-State IR. As shown in Figure 4, in Groups 1 and 2, faculty and librarians, over half of the respondents from each area preferred to have a more comprehensive list of search results in their area of research. The majority of surveyed university faculty from Group 1 indicated they were very or somewhat pleased about having metadata-only content if it provided more comprehensive search results and a complete picture of their digital scholarship (see Figure 5, respondent survey feedback responses).
jlsc-pub.org eP2220 | 19 The faculty in Group 1 also overwhelmingly preferred to have a complete record of their digital scholarship rather than a partial one or none at all ( Figure 6). Further studies will need to be conducted to gather a larger body of faculty responses and to further validate this finding, but this study indicated faculty preferences for comprehensive search results and publication representation, even with metadata-only records. These results provided support for the IR team to demonstrate overwhelmingly positive to neutral support from local stakeholders for this adjustment and gave evidence for IR managers to consider the inclusion of metadata records. Surveyed respondents were given opportunities for open comments, which provided insightful faculty feedback: • "My frustration is not with the library, but rather the scientific publishing industry and my colleagues who put their work in journals that try to sell reprints." • "I would need to access the full work in order to use/cite the research." • "I would be extremely pleased to know relevant work existed. Of course it would be inconvenient to hunt it down, but being frustrated about that would hardly make me frustrated to now have the information that it exists." • "I think if we are talking about having some information versus none, then some information always wins. But the same type of certainty also exists when comparing all versus some, having all of the information always wins.

NEXT STEPS
IR managers and institutions must decide for their individual institutions if they will add metadata-only records. As the project teams assess the outcomes of their pilot projects several lessons and next steps have become apparent.

Impact on Users
There is a risk of frustrating users when citations of works are the only option available due to copyright limits and unavailability of pre-or postprints. Researchers may hit paywalls when the library does not subscribe to the resource in which the content was published. It may be possible in many cases for researchers to access cited works, if the library has subscriptions to the journals in which they appear. Non-affiliates may have a more difficult time accessing cited works through the repositories, although depending on their affiliation they may have access to the content even if the university does not. When there is not a full-text version or direct link to the published research, links should take researchers to the publisher's page. K-State attempted to combat this risk by adding a note to each record warning users that they might not be able to access the full-text content.

Postprints
The lack of access to the full-text publications provides the opportunity for repository managers to engage faculty around the importance of retaining preprint or postprint versions of their published articles (or maybe more importantly around how to submit addendums to retain their rights and how to find journals that never take their copyrights to start).
jlsc-pub.org eP2220 | 21 The majority of journals only allow either a pre-or a postprint copy of an article to be archived in institutional repository, and each pilot project indicated that most faculty retain only the publisher's PDF. As demonstrated above, this creates a resource strain on library staff tasked with locating and ingesting the correct version.
PSU's pilot project reflected the importance of soliciting faculty for the corrected version. Shortly after the pilot project ended, PSU implemented new workflows for student workers to communicate with faculty and request the corrected version. This creates a more sustainable model that attempts to further increase the rate of deposit of full-text articles.
Prior to the launch of the pilot project at K-State, it was a standard practice to reach out to faculty and directly request postprint copies of published work. Decreases in staffing resources led to more efficient workflows for faculty work discovery, permissions checking, and IR ingest, but the practice of soliciting postprint versions of discovered work became unsustainable and was put on hold indefinitely. The project at K-State was partially meant to help make soliciting postprints sustainable by leveraging metadata-only IR records to give faculty more of a stake in how they are represented in their institution's IR.

Harvesting
Searching multiple academic databases provides the most comprehensive picture of the scholarly output happening at the institution, but handling multiple feeds (RSS and API) can add an additional layer of work for normalizing the data. This layer can be time consuming, as records must be compared for accuracy and checked for duplicates. With third parties controlling the data output, much of this process is currently a manual process due to the discrepancies in metadata displayed. As a result, PSU built a citation application (see Appendix B) that captures information from various databases' RSS feeds and APIs that feed into the repository by exporting citation metadata into an Excel spreadsheet. The citation application also confirms author affiliation and removes duplicate records. This tool allows for scalability beyond previous harvests and is much easier for library staff to implement than a more inclusive workflow.

Discovery and Indexing
Repositories are designed to facilitate the efficient dissemination of content to a variety of locations including discovery layers, ILSs, OAI-PMH aggregators, and emerging tools such as Unpaywall and the Open Access Button. The inclusion of IR metadata records into these disseminations potentially muddies the waters of the otherwise clean data Journal of Librarianship and Scholarly Communication streams to tools that expect full-text items. In these cases, the universities anticipate modifying these harvests to ignore metadata records. This also provides the opportunity to discuss and explore other avenues to showcase institutional research outputs, such as a CRIS platform. Efforts should be made to reduce any confusion or discovery issues where possible.
The effects on searching within library systems and any other platforms where IR records are indexed were considered during these pilots. The pilot teams determined that it would be important to have clear identifying information on metadata-only records to use in filtering records and aiding in user navigation of IR records. Working with the integrated library discovery services teams, IR staff will exclude IR records that are metadata-only or metadata with links, which should eliminate the problem of users going in circles from the discovery layer to the IR to the discovery layer. To indicate metadata-only records, PSU added "Citation" as a document type in their metadata, and K-State included the addition of the fields "Metadata" and "Publisher URL" within each record. The new field "Metadata" included text to let the user know that they have accessed a metadata record. The "Publisher URL" field gave the user a hyperlink to the original publication, which may or may not be open access.
"It would be helpful if a statement was added that the publisher did not allow the content to be posted. Then I would be frustrated with the publisher but not K-State Libraries." This and other statements by faculty show us that their frustrations may not lie with the library or this service. Through statements added to the IR records and education on campus, IR managers can work to further reduce frustrations on campus and help our authors understand how to retain certain rights to their content.

CONCLUSION
Institutional repositories were developed to make academic institutions' intellectual capital accessible to the world and simultaneously provide a value-added service for enhancing scholarly communication. Thus, the IR metadata pilot project teams recommend that their university IRs move forward with adding metadata-only records and metadata records with links to publications. The workflow and process used to ingest metadata-only records will be continually refined as needed, including addressing issues discussed within this article.
After extensive conversations and meetings, PSU university administration and library administration agreed with the recommendations of the pilot project team and approved transitioning to adding metadata-only records across the entire repository. The pilot team felt that the need for additional staffing, as mentioned above, needed to be addressed in order to fulfill the true mission of a comprehensive reflection of the breadth of works created at PSU. With the implementation of the citation tracking tool (Harvesting section above), the pilot team began jlsc-pub.org eP2220 | 23 implementing new workflows and started to add metadata-only records to the repository. The K-State library administration agreed with recommendations of their local pilot team and approved the addition of metadata records to the IR. Administration and the pilot team felt strongly that several issues and features, many found within Next Steps section above, needed to be adjusted and/or reviewed before proceeding with ingesting metadata records into the IR. After administrative approval, librarians began reviewing and preparing the live IR site and interconnected systems to properly handle metadata records. Unfortunately K-State has experienced several staff attritions and extensive damage to their library due to fire, both of which have hindered implementation efforts.
It is important to consider return on investment when deciding whether or not to implement a new process. While adding metadata records into the IR will increase demand on IR resources, this increase would yield several benefits that may justify the allocation of these resources, including but not limited to • Reducing disciplinary silos across campus • Increasing diversity of represented content • Improving university recognition through centralized, visible content

• Enabling comprehensive representation of university scholarship and research
If metadata records, however, are not added, potential problems include • Risk of frustrating users unable to find any record of a university-affiliated publication • Missing scholarship in the institutional record • Added difficulties for researchers in connecting an article to specific faculty • IRs limited to full-text versions of scholarship The pilot teams feel that metadata-only content does not change the purpose and scope of the IR-to disseminate the research at the university-and does move the IR in the direction of documenting as much scholarly output as possible. These pilots showed an incredible average of 60% increase in representation and also demonstrated that expanding the definition of the repository programs permitted library staff to include more faculty records. K-State also demonstrated a 58% increase in included departments. These results satisfy the pilot goal to demonstrate that adding metadata records creates a more inclusive repository that provides reliable information about the research output produced by faculty affiliated with the institution.