(This Opinion piece presents the opinions of the author. It does not necessarily reflect the views of D-Lib Magazine, its publisher, the Corporation for National Research Initiatives, or its sponsor.)
Policies for Repository Managers
This article is written mainly for repository managers who are at a loss as to what policies they (or their universities or research institutions) ought to deploy in order to ensure that most, if not all, of the institution's scholarly output is deposited in the institution's repository. In essence, there are only two pure policies:
This short article describes a third policy that provides a transitional path between the two.
The institutional mandate
The obvious and no-risk solution for ensuring deposit is for the institution to require researchers to deposit their publications in the institutional repository (IR). There is ample evidence both in pre-implementation surveys and in post-implementation outcomes that mandatory deposit is acceptable to over 95% of researchers. See for example two chapters by Swan and Sale in Jacobs .
One Australian university (Queensland University of Technology) that mandates deposit of its publications is leading the world in collecting 70% of its annual research output and the fraction is rising. This level of success is surprising, since the researcher's world is hemmed in with requirements to teach, to ask for student evaluations, to write and mark examinations, to supervise Ph.D. students, to publish research, to report to granting bodies, etc. Nevertheless, it may be difficult to convince some senior university executives that institutional mandates are required for significant deposition rates.
An institution-wide requirement to deposit in the IR is exactly what is needed. Once such a policy is in place the IR manager's approaches to researchers and heads of centers and all the plethora of feel-good activities actually work. People who are required to deposit their publications are grateful for advice. The occasional chase-up call is not resented. Just about everything that the university can put in place (for example, publicity for deposits, awards for the best author or paper, assistance with self-archiving, download statistics, etc.) will begin to work as it resonates with every academic in fulfilling his or her duty.
A mandatory deposit policy will approach a capture rate of 100% of current research publications, though it will take a couple of years to achieve that goal. Figures of 60-90% can be expected in a short time. See Sale  for some data on how mandates actually work.
Various voluntary deposit policies are not worth talking about for long. In the absence of mandates, every encouragement activity known to Man fails to convince more than a small fraction of researchers to invest the five minutes of time needed to deposit their publications, and the percentage does not grow with time. When you look at voluntary deposit policies closely, all the encouragement activities (awards to top authors, regular articles in the house magazine, great feedback, personal approaches, download statistics, seminars, explanation of the OA advantages, etc) fail to achieve significant participation. There are plenty of Australian examples of this, but it is a global experience as well. The reason is easy to grasp: these activities appeal to the already converted and practicing self-archivers, not to others who are skeptical non-participants.
Voluntary deposit policies are known to achieve no greater deposit rate of current research than 30% and more usually around 15%. At these levels the incoming participants and the drop-outs balance each other. The evidence for this can be produced and is absolutely clear [4, 5]. At such deposit rates, one wonders why it is worth bothering having a repository or undertaking the proselytizing activities, other than simply to have a repository in place (a yes/no tick).
It is also useless to look at growth rates of documents in the repository without taking their publication and deposit dates into account. The evidence shows that many 'converted' depositors busy themselves with mounting all their old papers. This is not to be discouraged, but it is not a significant performance indicator. The important performance indicator is: 'How much of your institution's annual research output appears in your repository by (say) six months after year end?'
So, when relying solely on voluntary deposit, many repository managers find themselves between a rock and a hard place. They can't convince the senior executives to bring in a mandate, and they know that voluntary deposition does not work. Fortunately there may be a transitional way ahead. I call it the patchwork mandate for reasons that will become obvious.
The Patchwork Mandate
What is the patchwork mandate? Simply this:
Let's look at this a bit more closely. There is solid evidence that departmental mandates work and much faster than university-wide mandates do. A year or so suffices to achieve a substantial acquisition rate of current research. This is because there are fewer people involved, and the researchers tend to trust their departmental leaders more. It is also easier to achieve conversion at the departmental level. Two documented examples are the Department of Electronics and Computer Science at the University of Southampton, and the School of Computing at the University of Tasmania .
What is a departmental mandate? A decision by the Head of Department (or a Research Director or a democratic staff meeting) that all peer-reviewed articles in the department must be deposited in the IR as a postprint, at the time of acceptance. ROARMAP contains a draft policy  you can adapt. Its effect is immediate, and most members of the department comply without objection. The 'enforcement' of the policy (if any is needed) is in the hands of the responsible person of the department, and all that person needs to do is to watch what researchers in the department claim they have published and ask: "Have you archived that yet?" That is enough no punitive action is required.
How do you achieve departmental mandates? Well, what you don't do is try a scatter-gun approach across the entire institution. Not only does such an approach waste effort, but it alienates people with whom you will want to participate later. You first analyze all your institution's departments and research centers. You decide which senior people within them might be amenable to persuasion. A high research profile is a good indicator for participation, as is a discipline where online access is already widespread. Another pointer for success is an area where a funding body mandate will affect many people. You know your institution better than any outsider, so choose your own criteria.
Next, you concentrate on the leader of one of the selected departments and possibly people around him or her to firstly deposit their own current research. Show them what they can get out of depositing their research (for example, obtaining download statistics), and then persuade them that their whole department should deposit as well. Give the department leaders the words to use to persuade their colleagues to participate. Suggest implementation. Provide support. Run seminars. Provide monthly deposit data summaries. But all of this activity should be strictly targeted at the selected department. Once you have a mandate from that department, keep up your support and especially continue providing feedback. Publicize the department's successes across the institution, and move on to the next target. Of course, you might tackle a few targets at the same time, but not too many. Successful departmental mandates are what you are after, not individuals.
You will end up with an odd collection of mandated departments, with the rest of the institution's departments still depositing voluntarily. Hence the term patchwork mandate, like a calico or tortoiseshell cat. You won't achieve 100% deposit rates, but you may escape from the 15% - 30% ceiling of voluntary deposit.
Two of the key features of the patchwork mandate are that (a) you are convincing the very people you need to support you in instituting a university-wide mandate, and (b) you are demonstrating that mandates work without difficulty.
When you as repository manager have (say) 40-50% of the departments (or the same proportion of research output, which ever occurs first) with departmental mandates, go back and argue with your institution's senior executives. Take some of your converted heads of departments with you they are key resources and are likely to be involved in the senior echelons of power. If the senior executives still don't agree to bring in an institutional mandate, point out that without an institutional mandate, you will have to tackle the remaining, more difficult departments one by one to try to convince each to mandate deposit at the department level, which will result in greater expense and delay of benefits to the university as a whole.
I am convinced that the patchwork mandate strategy described in this article will work in most cases. It is being trialed in Australia, and although it won't achieve 100% deposit of content into the institutional repository instantly, it is a clear way to work towards that goal. You can even explain the patchwork mandate approach to your senior executives, and they probably won't stop you from trying it. They may even encourage you in your efforts.
Just remember that voluntary persuasion of individuals is known not to work beyond a pitiful participation level. Self-archiving needs to be made part of the routine academic duty, and this requires a policy endorsement of mandatory deposit by someone.
 Jacobs, N (Ed), Open Access: Key Strategic, Technical and Economic Aspects, Chandos Publishing, Oxford, 2006. ISBN 1843342030. The two chapters mentioned are also available at <http://eprints.ecs.soton.ac.uk/12428/> (Swan) and <http://eprints.utas.edu.au/257/> (Sale).
 ROARMAP. Signup at <http://www.eprints.org/openaccess/policysignup/>. Sample policy page at <http://www.eprints.org/openaccess/policysignup/sign.php>.
 Sale, A. The acquisition of open access research articles. First Monday, 11(9), October 2006. <http://www.firstmonday.org/issues/issue11_10/sale/index.html>.
 Sale, A. Comparison of content policies for institutional repositories in Australia. First Monday, 11(4), April 2006. <http://firstmonday.org/issues/issue11_4/sale/index.html>.
Copyright © 2007 Arthur Sale