Online First

COAPI (Coalition of Open Access Policy Institutions) was formed against the recommendations proposed in two conferences held in the summer of 2011 from 22 North American institutions with existing faculty initiated open access (OA) policies. The University of Kansas (KU), which in June 2009 became the first U.S. public university to adopt OA policy regarding scholarly research in peer-reviewed journals, recently announced that it had spearheaded the formation of a 22-member COAPI. The objective of the COAPI is to bring together representatives from North American universities with established faculty OA policies and those in the process of developing such policies, to share information and experiences and to illuminate opportunities for moving faculty-led OA forward at member institutions and advocating for OA nationally and internationally. COAPI will offer a collection of best and evolving practices to act as a roadmap for inspiring, promoting and implementing open access polices at institutions without existing or effective OA policies. The main objective of this study is to provide the current status of self archiving policies for OARs of COAPI-members.

A survey conducted by the Coalition for Networked Information (CNI) found that research libraries mainly in North America have taken on a leadership role in both policy formulation and defining operational deployment roles for OARs at research universities. Basically, CNI is engaged in developing and managing networked information content; building technology, standards, and infrastructure; and policy and consultative activities. Almost all the COPAI member institutions belong to the CNI group. Lynch and Lippincott (2005) in their survey of the individual and consortia members of the CNI, concluded that 40% of respondent institutions have some type of OARs operating and that 88% of those who do not yet have a repository are planning to set up OARs. The survey also indicated that the number of repositories in non-research intensive universities was limited, while the majority of doctoral-granting institutions either had established or were planning to establish IRs. Another researcher (Westrienen & Lynch, 2005) surveyed IRs developed by CNI members abroad.

A team of The Primary Research Group (Primary Research Group, 2007) presented data from 56 OARs from eleven countries including the USA and Canada, and representations from Europe and Latin America, and early adopters of IR technology worldwide (Ware, 2004a). This report covers costs, budgets, software, personnel, OA policies, marketing, relations with faculty and other contributors of content, and many other issues relevant to those managing or designing an OAR including OARs of COPAI members.

Bailey et al. (2006) surveyed the 123 member libraries of the Association of Research Libraries (ARL) regarding their IR efforts and published findings in an ARL SPEC Kit. It is important to note that ARL includes 12 COPAI members. The survey found that eighty-seven libraries (71%), including the University of Oregon and the University of Kansas responded to the survey. Of those, 37 (43%) have an operational IR, 31 (35%) were planning for one by 2007 at the latest, and 19 (22%) had no immediate plans to develop an IR. This SPEC Kit includes documentation from respondents in the form of IR usage statistics, deposit policies, metadata policies, preservation policies etc.

In the MIRACLE project (MIRACLE - Making Institutional Repositories a Collaborative Learning Environment), Markey et al. (2007) reported that almost 85% of the research university respondents were either planning, pilot testing, or had already implemented IRs. Only 37% of the respondents were either planning, pilot testing, or had already implemented repositories. This project reported a brief discussion on the initial struggle of developing OARs in universities in North America. This project investigates the implementation of IRs in colleges and universities of the USA in order to identify models and best practices in the administration, technical infrastructure, and access to repository collections. Another group of authors (Markey et al., 2007; George, 2006) reported the development of IRs in the USA as a whole. Additionally, Rieh et al. (2007) reported that the majority of the research universities in the USA had implemented OARs for their own organization. Other researchers (Shreeves & Cragin, 2008) reported that only 40% of the institutions that offered doctoral degrees in the USA had an IR. On the basis of the above mentioned research studies and other related studies, the following key events (Table 1) are identified (http://www.sparc.arl.org/COAPI).

There are many OARs coming online and everyday one repository is being added to the OpenDOAR (Directory of Open Access Repositories) and ROAR (Registry of Open Access Repositories, 2015) databases. As per the OpenDOAR (2015) database, the total number of OARs is 2947 (August 2015) from all over the world whereas North America (Continent-wise) possesses 569 (19.3%) OARs and stands in 2^nd position after Europe and Asia which possesse 1300 (44.1%) and 593 (20.1%) OARs respectively.

The United States (country-wise) possesses 470 (15.9%) OARs and stands in 1^st position in the global repository movement. Another two North American countries, Canada and Mexico possess 72 and 27 OARs respectively. Many repositories with a mandate do not have a corresponding entry on the OpenDOAR or ROAR databases. Some repositories that have an open access mandate did not register their names in ROARMAP (Registry of Open Access Repositories Mandatory Archiving Policies) and their URL (Uniform Resource Locator) is not working properly. Some of these are pilot projects and some are now becoming full OAR services, and it is at this stage that they need to start formulating policy documentation, or at least formalizing their policy decisions.

There are no sources that provide a numerical count or cumulative growth of all OARs in North America at any one point in time. In order to examine the development of open access (OA) mandate policies, this study consulted ROARMAP, a site created and maintained by the University of Southampton in England, as an online location for policy registration. The study consulted with other important sources like OpenDOAR and ROAR databases to improve the accuracy level. Other sources consulted included SHERPA/RoMEO (Securing a Hybrid Environment for Research, Preservation and Access/ Rights MEtadata for Open archiving) Publisher copyright policies & self-archiving policy list; SHERPA/JULIET’s research funders’ open access policy list; SHERPA/FACT (Funders’ & Authors’ Compliance Tool).

The United States is the main contributor in this open movement and North America is witnessing not only the growth in the number of open access repositories (OARs) and journals, but also the development of other related expressions and initiatives that have emerged in the domain of science and technology institutions.

As per the ROARMAP (2015) database, a total of 724 (August, 2015) policy proposals and implementations have been registered and documented worldwide. All the mandates are listed under five (5) broad categories such as - Funder (79), Funder and research organization (54), Multiple research organizations (08), Research organizations (e.g. university or research institution) (512), and Sub-unit of research organizations (e.g. department, faculty, or school) (71). North America possesses 157 (21.6%) mandates (the United States 127 & Canada 29) both from institutional and funding agencies. The SHERPA/JULIET (2015) list/database of funder OA policies, archiving mandates, and guidelines, includes 14 from the United States and 15 research funder policies in Canada, making Canada second only to the UK in quantity. Nearly all of these Canadian funders are in the biomedical and/or health disciplines. From the SHERPA/RoMEO (2015) database (list of OA publishers copy right policy) it is noted that 578 publishers are from North America (66 from Canada & 505 from United States, and seven (7) publishers are from Mexico.

Much has been written to date about OA self archiving policy implementations because OA self archiving policy implementation is a difficult task. There are therefore many thousands of universities, research institutes, and research funding agencies across the world that have not yet implemented an OA self archiving policy. Policy pioneers have faced considerable challenges in meeting their own aims and achieving recognized success (Armbruster, 2011). Laundry lists of policy issues are given by Ware (2004b), Barton and Waters (2004), Rieh et al. (2008), and Shearer (2005). In a survey for OpenDOAR in early 2006, Peter Millington discovered that about two thirds of OARs registered did not have publicly stated policies (Millington, 2006). It is important that the necessary policies and procedures are in place to ensure that the responsibilities for content, management and maintenance of the repositories are accurately specified. Therefore, it seems appropriate at this time to ensure that policy documents are formulated correctly (Asamoah-Hassan, 2010), but the situation is not satisfactory. Some policies may be listed under a different name or entered incorrectly.

This analytical study is limited to only OARs of COAPI-members countries and for the analysis and interpretation of policy documentations, three databases, OpenDOAR, ROAR and ROARMAP have been consulted. Annex I provides basic information regarding all 22 repositories, but for the policy analysis, only 18 (81.8%) repositories (annex II) having OA self archiving policies (at least one policy mentioned in table 2) have been considered. Finally their policies have been discussed in relation to those issues mentioned in table 2. The policy documentations of selected OARs (annex II) have been discussed below under section eight (8) with broad headings that correspond to the evaluation criteria. Policies can be separated out into the following areas:

Several key findings have been identified after analyzing selected repositories along with their policy documentations (annex I & II). As stated, many repositories that possess a good number of OA resources do not have any self archiving policy. It has been observed that many issues related to the OA self archiving policies are either not properly configured or not included at all. Generally repositories of elite institutions have more detailed policy documentations than those OARs that are at implementation stages or in their infancy.

A total of nine (9) key policies (annex II) have been identified, which are common to almost all repositories, and only 18 (81.8%) repositories have OA self archiving policy (at least one policy mentioned in table 2) documentations. The key findings are: all items are not available as full-text (s.l.no. 3); all repositories having data access policy impose access restriction; around seven (7) repositories do not provide OAI-PMH base URL and thereby do not expose metadata for harvesting by service providers; non-availability of access and usage statistics is a common problem of almost all repositories (except s.l.no. 4, 9, & 17); a total of four (4) repositories have not mentioned software type; altogether three (3) repositories have not provided a total number of objects uploaded; a total of three (3) repositories have not mentioned content types; repositories are not in concert with others who have established a form of repository network at the national/international level; almost all the repositories are multidisciplinary in nature and cover subjects mainly related to Arts & Humanities; only one (1) repository (s.l.no. 18) has a customized user interface and it is available in Spanish & Portuguese other than the default English language; and four (4) repositories hold special items including administrative documents (s.l.no. 10), notebooks and sheet music (s.l.no. 01), links to other resources (s.l.no. 18), musical scores, posters, etc. (s.l.no. 19).

Several technical as well as non-technical issues have not been discussed in policy documentations of the OARs mentioned in annex II. The following recommendations may be considered at the time of devising an OA self archiving policy for building up institute-specific OAR systems.

• ∙ Most of the OARs don’t have well stated policy documentations regarding different issues (annex II). It is well established that the success of any repository system depends on proper policy documentations, so, OA self archiving policies based on global recommendations and existing best practice guidelines are key to the success of any OAR system;
• ∙ In archiving policy, it is not clear whether OARs follow mandatory/optional policy for inclusion of content. This study proposes to follow a mandatory archiving policy because only mandatory policy can be effective for encouraging wider contribution of contents to the OARs (Roy, 2015). Repositories should mention both archiving timing, i.e. when to archive, and archiving format, i.e. in which format it (content) is to be archived;
• ∙ In content and collection policy, a repository may adopt control vocabulary or any appropriate subject taxonomy in organizing and managing collections to fulfill the subject approach of the users (Roy, 2014);
• ∙ In copyrights and licensing policy, repositories should mention who holds copyrights of the items deposited and what the retention period should be for the objects deposited to OARs;
• ∙ In data access policy, repositories should mention whether they provide full text access or metadata only, even if it is not clear whether repositories follow different access policy for different types of users depending upon the types of documents;
• ∙ An embargo/retention policy has been neglected by most of the repositories. This study respects any embargo requested by the authors or journal publishers or any funders. No item (full text) should be accessible during this period and items should be submitted and archived only after the retention period expires;
• ∙ In metadata policy, repositories should use any standard indexing tools and should support extraction of metadata from different OAI-PMH compliant OARs. A repository may follow domain-specific metadata schemas in order to support, ingest, manage, and use of non-textual data in their collections;
• ∙ In submission policy, workflow issues have not been discussed. Repositories should mention different stages of workflows and different roles played by E-persons at different workflow stages;
• ∙ To be compatible with a global system and to achieve technical interoperability, OARs should apply open standards, open technologies, and open source software (OSS) as much as possible.

The OAR movement as Green Path (also known as self-archiving) to Open Access was initiated by leading universities in the USA like MIT. In fact most of the COPAI members belong to the USA although the base membership of COPAI covers the entire North America. This study shows that OA self archiving policies, even in these forerunner organizations, require improvements in different aspects like quality control procedures, version control, multilingual issues, user interface, etc. Some of these issues related to OA self archiving like archiving policy, content policy, collection policy, withdrawal policy, etc. are well covered by the COPAI members and some of the important issues like submission workflows, interoperability standard, preservation techniques, etc. are not included at all in the policy documents of these organizations. This study shows clearly the strength and weakness of the policies of COPAI members and argues that successful deployment of OARs requires that various OA self archiving policy issues should be clearly defined and followed by COPAI members. An in-depth analysis, as given in annexure II, shows that existing practices of the OARs of COPAI members do not provide an integrated OA self archiving policy framework. This is possibly due to the fact that different countries may have different approaches towards developing OAR strategies and policies depending upon varying local conditions. For example, funder mandates differ from country to country. Even within the same country these mandates differ from organization to organization on several levels. COPAI members, as leading players of OA self archiving in the world, need to develop a core set of OA self archiving policies with an option to add country specific and institution specific policies into the core part of the policy framework. The core policy framework is required to support by COPAI members as prominent players of the domain for promoting OA self archiving on a global scale. This paper recommends a set of policy elements (included in annexure II) as part of the core policy framework for COPAI members.