Advice on filling your
repository
Arthur Sale, University of Tasmania, 26 March
2010
1. Preface
I was moved to produce this by Hugh Glaser’s remarks that
no-one was prepared to offer advice to beginners or people transitioning to a
properly mandated repository. This advice is not new. It has been said by many
others in part, and I have been preaching it in Australia and New Zealand for
at least five years. It is however firmly based on experience, and knowledge
of what works and what doesn’t in many universities, right around the
world.
2. Long-term target
The long term target of every repository manager must be
that depositing research publications should become an automatic part of an
academic’s work pattern, just like submitting their research for publication,
or setting their examinations. I take this as so obvious that I will not argue
it.
The only known route to achieving this is via what is called a “mandate”. As
an important transition to getting the long-term target to happen, and happen
worldwide, institutions (and research funders) have to require their employees
(grantees) to deposit their research publications in an OA repository. This is
unexceptionable and few complain about it. Alma Swan’s studies support this as
does actual on the ground evidence. Again, I will not argue this, except to
note that it is the only strategy known to fill repositories with more than
50% of the available research publications. This is well documented. See for
example http://eprints.utas.edu.au/388/.
Once depositing and making publications Open Access (OA) is universal (or even
as low as 50% of the world) the momentum of the technological revolution will
be unstoppable, and academics publishing quality papers will rush to deposit.
It has already happened in Physics and Computer Science in case you think this
is fanciful. Mandates will no longer be required except to deal with problem
cases or will exist as an historical leftover from the beginning of a
technological change. Those academics that don’t accept this will be the
cast-offs of that generation of scholars, just like previous scientific
revolutions such as that for plate tectonics.
3. Short term targets
Assuming that you do not have an enlightened senior
executive, each repository manager has to adopt a different strategy. If it
isn’t oriented towards gaining a mandate in the long term, you are wasting
your time, unless you believe in prayer and miracles. Voluntary persuasion has
consistently been shown to achieve around 15% (sometimes a little higher) of
available deposits. See http://eprints.utas.edu.au/264/.
Five years of experience has not changed that evidence. The only thing that
will change this in a voluntary environment is the universal adoption
of OA mentioned above which will drag in most.
4. Strategies
I suppose you are in this situation. You have not got a
mandate. What to do?
Firstly you should accept that the idea of one person (other than the
Vice-Chancellor or one of the Senior Executive issuing a mandate) changing the
work practices of a university is such a foolish notion a priori that
it can be put aside as a delusion. The theory of change says that change comes
either quickly through fiat (a mandate), or slowly through evolutionary
principles. So here are some suggestions to start slow but aim at
fast.
Patchwork Mandate
Do not try a scatter gun approach. It won’t work because
your effort is too thinly scattered. Soon backsliders come to balance the
converts. This is common centuries-old missionary experience.
Sit down and identify target departments in your university and target them.
(More missionary experience – go for the chiefs.) Each university is different
so you need to choose your own targets. Set in place measures that will
ensure that even when you move on to another target, you have a champion in
that target area who will be your surrogate to deal with backsliding. Support
the champion. Massage their ego and with information. Read the paper http://eprints.utas.edu.au/410/,
and apply known techniques of change management to the problem. Did I mention
that I was previously a Pro Vice-Chancellor before I retired?
Why do we do this?
Your first and main answer to an academic should be that
open access increases citations. Counts of publications are now passé, and
they were only ever the crudest form of metric for research impact. Citations
are now the fashion, and they are a less crude surrogate – if an article is
cited at least someone read it! As time goes on we shall see more complex
measures like the SJR and the set of Hirsch indexes coming to the fore in a
basket of metrics that attempt to measure the multi-dimensional aspects of
impact, different for every discipline. Work has to be online to be
assessed!
You should download Anne-Will Harzing’s Publish or
Perish tool http://www.harzing.com/pop.htm. Try
it out. Demo it to authors. Good for people with unique or rare names, like
me. Not so good for John Smiths. However it shows how online access (in this
case through Google Scholar) can be used to develop sophisticated metrics of
research impact. Has X published one excellent paper in their life twenty
years ago, or has X a record of good papers over that time? It
matters.
Performance
Using performance management to promote your repository is
a double edged sword, so be careful. Some academics might hate you, and then
where are you? Better if a Head of Department asks for help. However, get your
repository able to deliver a research record summary so that a high-performing
academic can attach it to his or her report for performance management. The
Head of Department will get the message...
Forget about using your repository for promotion cases, etc
unless you have senior management support. It comes with the mandate
territory, not as a precursor.
Things that do not cut any ice
Don’t bother to explain that the university administration
would like to know what research is being carried out. 99% of academics would
say “Stuff that!” Don’t try to explain how when everyone goes OA, then your
researchers will have free access to the world’s literature. They don’t care –
they have never paid for access anyway.
An especial word of warning. Do not publish lists of individual academics who
are most downloaded this month, or similar. Do not try to award prizes to
individual academics. Most academics absolutely hate this, and you’ve lost
them for years. The exceptions are (surprise) those few at the top. My
experience is that the same people are at the top rather consistently, so what
are you doing? Pandering to their ego, that’s what. And maybe you get pats on
the back for your perspicacity. As a matter of interest, the same object has
been consistently top of the UTas downloads for at least three years!
Psychiatry is popular it seems. It probably has impact above its citation
count.
Other things that are useful
Most of the following tips are very useful with individual
academics, but responses to each one varies widely. They are worth doing
centrally in your repository. Especially they are the answer to the people who
already put their papers on their websites (so they are OA) and think that a
repository isn’t any better than their website. Disabuse them, because they
aren’t converts, they are dyed-in-the-wool OA providers. They won’t backslide
once you’ve convinced them.
- Put effort into making
sure that Google indexes your repository really well. Many academic crow
about being top of the list in Google on their search term. You can’t guess
that search term – it is the topic of their research that other researchers
are likely to use.
- Make sure that your
authors can read good statistics on how often their paper is downloaded.
Some academics follow these avidly and deduce from the patterns that someone
just cited their work and try to find out who. See http://eprints.utas.edu.au/es/index.php?action=show_detail_eprint;id=410;
or indeed any paper in the UTas repository that takes your fancy.
- Provide an easy service
so that authors can put a link to an up-to-date list of publications on
their personal website. For example http://eprints.utas.edu.au/view/authors/Sale,_AHJ.html.
- Make sure that each
author can download his own paper (even if restricted) wherever they are in
the world. Very active researchers value this a lot, because it is like
carrying a no-weight library of all your publications with you when you
travel internationally. Makes collaborative research a lot
easier.
- Offer Departmental
seminars on how to use OA information to best advantage. Use Google Scholar
shamelessly (doesn’t really work without OA). Talk about Harzing’s
Publish or Perish and demo it. Talk lots about metrics, citations,
and Scimago. Throw in citations as a way of searching forward in time from a
significant paper. Especially target postgraduate (PhD) candidates. They
influence more indolent supervisors.
5. Debunking some myths and Summary
This advice is meant to be useful to repository managers
and librarians who are new to this Open Access enterprise. However there are a
few extra myths that need to be squashed (not for academic
eyes).
- The metadata entered by
authors is in general as good as or better than that entered by librarians.
Do not try to do better. You are wasting money.
- Since almost all
searches are done by Google, full text analysis rules. Keywords and Phrases
are of such little consequence that it is a pity to waste a librarian’s time
on them when they could be doing something useful.
- Copyright issues can be
vastly overdone. If your mandate or your advice calls for the Accepted
Manuscript (aka postprint) as it should, just ignore checking copyright and
let it go up. If a publisher complains (they almost never do nor have any
grounds to), make it restricted without arguing. Do not accept the Version
of Record (aka publisher’s “pdf”) unless it is guaranteed open access. In
case it isn’t tell the author that the submission has been rejected because
they violated copyright. They’ll learn. They’re good at learning. They won’t
learn unless you teach them.
So to summarize: mandates are the only thing that works
now. Persuasion is a weak reed, but worth doing to prepare the ground for a
mandate, but be very selective in who you seek to persuade. Use change
management tools to assist you. The persuasive techniques continue to work
once you’ve got a mandate. Keep them going.
Arthur Sale
University of Tasmania
26 March 2010
