Jisc provides several services to support open access discovery and monitoring in the UK. However, a large proportion of open access material is still difficult to find due to issues with messy, lossy, expensive, and unreliable discovery. Improving the use of persistent identifiers across repositories, publishers, and funders could help create a more sustainable and comprehensive system for discovering open access content globally. This would better support finding, reusing, and measuring the impact of publicly available scholarship.
Transaction Management in Database Management System
Jisc Vision for Improving OA Discovery
1. Jisc OA Discovery
AVision for OA Discovery
Submission Acceptance Publication Use
SHERPA
JULIET
SHERPA
RoMEO
SHERPA
REF Beta
SHERPA
Fact
Monitor
UK
Jisc
collections
OpenDOAR
Research
publication
lifecycle
Jisc services
Report on
compliance
Deposit in
repository
Manage
costs
Check
compliance
Select
Journal
Maximise
impact
Record
impact
Publications
Router
Report
Monitor
local
CORE
IRUS-UK
RIOXX
Sarah Fahmy, Scholarly Communications Services Manager
8/1/2018
2. Introduction
»OA discovery is rapidly evolving area
»Jisc is looking at ways in which it can improve OA
discovery ‘on the ground’:
› Potential improvements to current Jisc services (CORE)
› Whether/ how we might engage in the future with
existing OA Discovery tools
3. A vision for OA discovery
»However, given that we do already have a thriving market
place and given Jisc’s wider remit, potentially our role
should be:
› to address the underlying infrastructure that all
discovery services rest on
› … and so lower the common barriers to better OA
discovery
› … so what does this mean?
4. ‘The Universe of articles’
Credit: Robert Sullivan/ Public Domain licence: http://bit.ly/2DreAZ7
The entire body of research literature is currently estimated at over 100 million
publications (Khabsa & Giles, 2014) …
…with an estimated 10% year on year increase in the annual number of these outputs
(Bornmann & Mutz, 2014)
5. Proportion of OA
» The State of OA (Piwowar &
Priem,2017) study looked across of
100,000 articles, to investigate OA in
three populations:
› All journal articles assigned a
Crossref DOI
› Recent journal articles indexed in
Web of Science
› Articles viewed by users of
Unpaywall
Findings:
• Of 67mill articles Unpaywall searches
across, 28% of the scholarly literature is
OA (19M in total)
• The most recent year analysed (2015)
also has the highest percentage of OA
(45%), which has almost certainly
increased by 2018
So…OA is a significant success,
and is now becoming the
mainstream in many areas of
research
7. » In many ways, we’re well served in terms of OA discovery services
› Great deal of very agile innovation (Unpaywall, OAButton, Kopernio)
› More established public infrastructure (OpenAIRE, CORE)
› Well-used library discovery services (Primo, Summon, Ebsco, OCLC)
› Long-standing commercial services (Web of Science, Scopus, Google
Scholar).
BUT..
Despite the objective of OA being to enable material to be more easily
found, re-used and impactful (and the rationale for the considerable
investment in it)….
OA Discovery Landscape
8. OA Discovery Landscape
THERE IS STILL A LARGE
PROPORTION OF OA
MATERIALTHAT CAN’T
BE EASILY FOUND
(it’s the old adage..)
CC BY-NC-ND 2.0, Credit: Andy Sim: https://bit.ly/2GWcB2w
11. » Messy: the user experience is poor. Scholars and others
interested in using public scholarship miss out on material that
should be available to them.
» Lossy: none of the discovery services is, or could possibly be,
comprehensive.
» Expensive: each discovery service has to spend time and
money repeating and maintaining the exploration and
harvesting work done by all the others
» Unreliable: the same search of the same content can deliver
different results day to day
What do we mean by that?
12. How can we improve this situation?
Improved use of
persistent identifiers
13. What might those interventions be?
Everyone to contribute to, and benefit from, the PID commons.
Examples:
1. Repositories register PIDs for content they hold. Where relevant, link to
journal DOIs, DataCite DOIs, etc
2. Publishers register a DOI at acceptance, using Crossref guidance on how to
do this when no article is yet available
3. Funders take ownership of the development of a global research project
identifier registry, to ensure it is community-governed
4. Everyone require ORCIDs and use the ORCID API…
14. The Outcome?
A sustainable and significant improvement in the extent
to which OA content globally can be found, re-used and
make a difference..
15. PIDVision
15
» Mapping the Persistent Identifier landscape
› PIDVision
https://orcid.org/blog/2018/06/07/mapping-pid-
landscape
› Nature article
https://www.nature.com/articles/d41586-018-05456-
8
› PID campaign – launching soon
Contact community@orcid.org if you’d like to get involved.
Using Crossref will give you a pretty good estimate of this number (excepting articles that don’t have a DOI, but these aren’t a huge number). But if you want to drill down by country or institution, Crossref currently doesn’t have enough affiliations or ORCIDs to do this with any accuracy.
For the denominator (number of articles): using Crossref will give you a pretty good estimate of this number (excepting articles that don’t have a DOI, but these aren’t a huge number). But if you want to drill down by country or institution, Crossref currently doesn’t have enough affiliations or ORCIDs to do this with any accuracy.
For the numerator (number of OA articles): Licence information is being filled in spottily, so the best you can do is get a lower bound for the number of OA articles (those that have an open licence in Crossref), with no way to accurately estimate the upper bound.
1. Completeness:
All of these services are attempting to find OA scholarly material from a variety of places in which it might reside, each of which has its strengths and weaknesses as sources, making the job of those services rather difficult.
Reliant on accuracy of the metadata- for example, we know licence information is being applied spottily so it’s hard to determine it OA status
2. Business models:
A range of commercial, public and other business models co-exist for discovery services. Up to now the market has been the decider, but long-term sustainability for these services is a key issue
Need to stress very strongly that this is purely illustrative, in no way meant to be an accurate description of the landscape
We propose that a huge amount could be achieved through a programme of interventions that led to persistent identifiers being used better.
“Persistent identifiers (DOIs, ORCIDs, ISBNs, etc) are the signposts in the research landscape and act as coordinates on the research map. They both tell us where something is located, and also act as signposts, guiding us to information sources and helping us to discover connections between people, ideas, organisations, funding, employment, publications, activities, and more.”
Where they are used well, then OA discovery is relatively easy - anyone can build a service that uses the Crossref API. However, PIDs are not always used well, and often are not used at all. This is for a variety of reasons, but most of which are tractable via a combination of technical, cultural and organisational interventions within different stakeholder groups.