Australia is a world leader in many aspects of open access. We have institutional repositories in all universities, funding mandates with the two main funding bodies, statements on or mandates for open access at a large number of institutions and a large research output available in many open access avenues. A summary of centrally supported initiatives in this area is here.
However we can do more. This blog outlines four impediments to the widespread uptake of open access in Australia: a lack of data about what Australian research is available, Copyright transfer agreements, the academic reward system and improved national discovery services. We suggest some solutions for each of these issues.
Issue 1 – Lack of data about what Australian research is available OA.
We collect good data in Australia there is good data about the amount of research being created and published annually. Equally, a considerable amount of Australian research is being made available to the wider community through deposit of research in institutional repositories, subject based repositories (PubMed Central, arXiv , SSRN and the like), through publication in open access journals.
However, this information is not compiled in a way to ascertain:
1. What percentage of current Australian research is available open access.
2. Where Australian research is being made available (institutional or subject-based repositories and open access journals).
3. The disciplinary spread of open access materials – an important indicator of areas needing attention.
Without this information it will be difficult to ascertain the level of impact the ARC and NHMRC policies are having on the availability of open access material from current Australian research. There are three actions that could help inform this area.
First it would be enormously helpful to know the percentage of Australian publications that are available open access.
There have been two definitive studies published on worldwide open access availability. Björk et al’s 2010 study concluded that 21% of research published in 2008 was openly accessible in 2009. Gargouri et al’s 2012 study found 24% of research was openly accessible.
But in these studies the method used to determine which work was available was to search for the items manually across several search platforms. This is clearly very time consuming. A study like this in Australia will require funding.
Second we need an easily accessible summary of the number of full text open access items in institutional repositories across the country. In an attempt to address this, the National Library of Australia aggregates research outputs from all Australian university repositories into Trove, and is working with the sector to improve discoverability and metrics around this collection. One challenge is that some repositories do not specify whether records have an open access full text item attached.
This issue was raised during a poll of repository managers in 2012. The poll found that as at June that year there were about 200,000 open access articles, theses and archive material (which includes images) in Australian university institutional repositories. Currently there is no automated way of obtaining an updated figure.
Third, a compliance monitoring tool needs to be developed to assist the ARC and NHMRC manage their open access policies. Currently all institutional repositories in Australia are implementing a standardised field in their repositories to indicate an item results from funding. But to date there is no indication of how this might be harvested and reported on.
Issue 2 – Copyright transfer agreements
As AOASG has already noted, there is a serious challenge keeping up with copyright agreements as they change. In reality, it is extremely difficult for an individual researcher to remain across all of the nuances of the copyright agreements. There have been studies to demonstrate that doing the copyright checking on behalf of the researcher increases deposits into repositories.
But the broader problem is actually two fold. First researchers often have little understanding of the copyright status of their published work. Many do not read the copyright transfer agreements they sign before publication. In addition, most researchers do not keep a copy of these legal documents. While there is currently some advice for researchers about copyright management, such as this page from the University of Sydney, generally awareness of copyright remains poor amongst the research community.
But before we start wagging our fingers at researchers, let’s consider the second, related issue. The copyright transfer agreements presented to researchers by publishers are often many pages long, written in small font and hard to understand. In addition these agreements are not consistent – they differ between publishers and often titles from the same publisher have different agreements.
Generally publishers ask researchers to assign exclusive copyright to them. But in most cases publishers only need the right of first publication of work, and normally do not need to control how research is used and distributed beyond this. There are options for researchers to negotiate different arrangements with their publishers, but the level of uptake of these in Australia is anecdotally very low.
It is highly unlikely there is any specific action that can force publishers to simplify their copyright transfer agreements. But there are a couple of actions the research community can make to improve the current situation.
It would help to have an Australian version of the SPARC Author Addendum tool which can be attached to copyright transfer agreements. This would need to be supported by a concerted education campaign about what rights researchers have, including training materials.
In addition the many researchers in Australia who work as editors for scholarly journals are in a good position to negotiate these arrangements with their publishers on behalf of their authors. An education campaign aimed at journal editors would assist them in this action.
Issue 3 – The academic reward system
The academic reward system supports the current publishing status quo. Widespread uptake of open access will continue to be a challenge while this is the case. A reliance on a numerical analysis of the number of articles published in journals with high Journal Impact Factors as a proxy for quality assessment is a narrow and limiting system.
There are many issues with the Journal Impact Factor. It also causes challenges for open access is it retains emphasis on a small number of specific journals which are in, the vast majority, subscription based. Yet there is evidence to show that open access & subscription journals of the same age have the same impact, indicating that it is time to look at other methods of assessing quality.
Currently the markers used to assess promotion do not differ much from those used for grant allocation. However, the contribution made by researchers to their academic community reaches far beyond simply their publication output. This includes editing journals and the peer review of papers. As there is currently no quantification of this work, the extent of the problem is unknown, although concerns about work overload have been expressed by the academic community. There are serious implications for the sustainability of scholarly publication in terms of human capital.
We need to move to assessment based on article level metrics rather than the organ of publication. It would be helpful if assessments such as ERA and funding allocation were to embrace new, existing alternative metrics. Examples include: Impact Story, Plum Analytics, PLOS ALM, Altmetrics and Google Scholar.
Institutions could consider recognising the hidden work academics undertake supporting the publication process in their promotion rounds. Recognition of peer review and editing roles as well as those researchers who are also publishing journals by running an open access journal using OJS or the like would add value to these activities and make the scholarly publication system more sustainable.
Issue 4 – Improved national discovery services
This last issue is in some ways, related to the first – knowing more about where the research we are producing is ending up. But it has a broader remit, for example incorporating data as a research outcome. Currently researchers can register their data with Research Data Australia which lists over 87,000 collections from over 23,000 contributing research teams.
We need to move beyond simply collecting research, and start working on ways to link data as research outcomes to reports on research publications.
During 2004 and 2008 the Australian Partnership for Sustainable Repositories (APSR) provided assistance and support for the repository community and developed technical solutions relating to interoperability and other repository issues.
APSR was supported by Systemic Infrastructure Initiative (SII) funding [note original post said NCRIS funding – thanks to David Groenewegen for pointing out the error. Amended 16 August]. When this ended, repository manager support was taken over by CAIRSS, financed in 2009-2010 by remainder money from another SII funded project, Australian Research Repositories Online to the World (ARROW) . The university library community through CAUL continued to support this project in 2011-2012 and the work has now been folded into the responsibility of another CAUL committee.
But the work APSR did developing country-wide technical solutions has not continued. Currently repositories around the country are being developed and maintained in isolation from one another.
An investment in current institutional repositories to increase functionality and interoperability will assist compliance with mandates (both Australian and international) and usability into the future. It will also enable a resolution of the metadata issue for country-wide harvesting by Trove.
We suggest revisiting support for country-wide technical development of solutions to common problems facing repositories throughout Australia. An example of a project that could be undertaken is the Funders and Authors Compliance Tool developed in the UK – SHERPA/FACT. This assists researchers to comply with open access mandates.
Dr Danny Kingsley
Australian Open Access Support Group
4 thoughts on “Four issues restricting widespread green OA in Australia”
Reblogged this on Open Access India.
Reblogged this on mary anne kennan.
Thanks for this post – it’s good to see the reflection on repository activity carried on in the midst of the seemingly idle Northern Hemisphere summer 🙂
These comments are based on the work carried out by the Jisc-funded UK RepositoryNet+ (aka RepNet) project at EDINA National Data Centre in Edinburgh, http://repositorynet.ac.uk/.
On issue 1: “Lack of data about what [Australian] research is available OA”
A.- The (much-critiziced) RCUK Open Access policy for funding article processing charges in the UK will require reporting from institutions in a year time since it started running Apr 1st. This has prompted a wide institutional effort aimed to track the institutional research output at submission time – since it’s payments what institutional research information management officers are now concerned about. This is both a very significant change of culture and key for deeper engagement with researchers. If the institutional service is able to persuade researchers to become involved with the research output reporting practice at a far earlier stage, we will have the opportunity from now on to both:
A.1) Progress in changing the strange procedure by which the producers of the research output metadata (the institutions) are forced to buy them back from the international databases in order to keep track of their own outputs
A.2) Implement a set of tools that will enable to track compliance for specific Open Access policies (incidentally beyond RCUK’s, covering diverse institutional and funder mandates, with special attention on the European Commission’s one for FP7 and Horizon2020-funded project outputs)
B.- It is now possible to automatically tell whether a given item filed in an institutional repository does have an open access full-text object available through text-mining mechanisms. In order enable this process, institutional repositories must first properly expose their contents to external aggregation crawlers.
On issue 2: “Copyright transfer agreements”
Besides the sound proposals mentioned in the post, the solution here may again very much lie in the hands of the funders – whose commitment should intensively be sought in this regard. Much discussion was held with repository managers at RepNet workshops on the acute need to simplify the way copyright conditions for self-archiving by researchers are featured on databases such as SHERPA RoMEO, but this requires publishers to accept a simplified range of options from which they should select their own – something most of them are unlikely to accept. Large research funders like RCUK and Wellcome Trust have funded the SHERPA FACT tool, http://www.sherpa.ac.uk/fact/, which is a step in the right direction, and could force some further progress if they were able to persuade publishers that a far simpler model is required for dealing with the copyright information in order to enhance compliance rates for their Open Access mandates.
On issue 3: “The academic reward system”
While completely agreeing that article-level metrics are key for moving onto a more reasonable research assessment mechanism, it should also be acknowledged that altmetrics-related initiatives are currently in their infancy. Furthermore, they are very fragmented and often run by commercially-oriented stakeholders, which may add further complexity to the picture. The emphasis here should probably lie in providing an adequate dissemination to inspiring best practices, especially those related to institutions actually committing to take those indicators into account for research assessment purposes, something yet to be seen.
ORCID could also be of much help for raising the profile of research activities beyond the sheer publication count, including coverage of research data, patents, refereeing and other journal-related activities. But it will always be up to the institutions to decide whether they want to pay any attention to these additional activities for evaluation (see Stephen Curry’s enlightening post on the matter, http://occamstypewriter.org/scurry/2013/05/16/impact-factors-declared-unfit-for-duty/).
On issue 4: “Improved national discovery services”
Initiatives for coordinating repository service development and implementation at national scale do make total sense when there is a clear idea for the kind of coverage to be attempted. The RepNet project has played such a role in the UK, contributing to the implementation of very useful –and internationally re-usable– repository services such as the Repository Junction Broker (RJB) for automated SWORD-mediated content delivery into institutional repositories or the IRUS-UK project for collecting cross-repository COUNTER-compliant usage statistics. Such wide-scope approach –also taken by supranational projects such as OpenAIRE in Europe or LAReferencia in Latin America– could yield very useful services to the international repository community.
ONLY THING NEEDED FOR GLOBAL OA IS EFFECTIVE GREEN OA MANDATES
There’s nothing wrong with OA growth in Australia that the adoption of the Queensland University Green OA self-archiving mandate model by all Australian universities and research funding councils would not fix.
(1) The problem is not with knowing what’s OA in Australia. (Well configured repository software plus ROAR will tell you that and Google will find it.) A mandate compliance monitoring mechanism, however, is indeed needed. But the ones to monitor compliance are authors’ own institutions, by requiring deposit immediately upon acceptance for publication, time-stamped within days or weeks of the date of the acceptance letter, for all published articles. Immediate-deposit should be a condition for research funding and performance evaluation, as it is at University of Liege, and as it has been proposed for UK funded research by HEFCE for REF 2020.
(2) It’s always good to agree on fair copyright agreements, but trying to convince publishers to agree to those should on no account be holding up the mandating and provision of Green OA. And for journals that embargo OA, there’s always the immediate-deposit mandate and the repository’s eprint-request Button to provide immediate Almost-OA with one click from the requester and one click from the author.
(3) The “academic reward system” is certainly not holding up OA. OA increases research uptake and impact, including citations. And the notion that OA needs some sort of preferential treatment for Gold OA journals, rather than just weighting them based on their track-record for quality, like all other journals, is and has always been complete nonsense, ever since it began to be mooted over a decade ago. The way to provide OA is to publish in the highest standard journal possible for one’s work, and then self-archive the refereed final draft. To pay to publish in a Gold OA journal just because it is OA (rather than because of its quality standards) is to pay for Fools Gold. (There is no OA problem for unrefereed or unpublished work; nor is getting academic credit for such work an OA problem.)
(4) Discovery tools can always be improved, but they are already pretty powerful. They will not discover OA content that is not there. Hence the only thing that is really needed for OA is effective Green OA mandates, along with effective monitoring of compliance, in order to get it up there, out in the OApen, to be “discovered.”