-
1.
OPEN ACCESS MANDATES:
COMPLIANCE BY FUNDERS
Cassidy R. Sugimoto
Vincent Larivière
csugimot@nsf.gov
vincent.lariviere@umontreal.ca
@csugimoto @lariviev
-
2.
Brazil: paving the way for OA
2005: several declarations of support OA
2007: bill presented to parliament proposing national policy for mandatory OA
2016: 97% of Brazilian journals are OA
2018: highest number of OA journals
one-third of Brazilian articles immediately available free to read
-
3.
Why OA as a success metric?
(Ashton University Library Services)
-
4.
Funding driven initiatives
-
5.
Routes to open access
Non-subscription
Subscription
APC / subsidy
Self-archiving
Gold OA
Green OA
Toll access
-
6.
OA miners
-
7.
Unpaywall
(Apr. 2018: 95M)
-
8.
Unpaywall analysis (Piwowar et al., 2018)
-
9.
Rise in mandates (ROARMAP, 2018)
-
10.
Effect of mandates (Gargouri et al., 2012)
0%
10%
20%
30%
40%
50%
60%
70%
80%
2002 2003 2004 2005 2006 2007 2008 2009
Mandated Institutions
PercentageofOA
2002 2003 2004 2005 2006 2007 2008 2009
Non-Mandated Institutions
-
11.
Mandating open access
Higher flexibility
Greater
heterogeneity
Higher restrictions
Use of embargos
Evaluate faculty
(e.g., University of Liège)
Withhold funding
(e.g., NIH)
DESCRIPTION LEVERS
-
12.
Characteristics of mandates
Required
Embargo
Repository
Dark/Ope
n
Copyright
Opt-out
-
13.
Sample of funder mandates
-
14.
Policy effective dates
-
15.
Calculating compliance
Document object identifier (DOI)
% COMPLIANCE
-
16.
Number of funded papers
Funder 2009 2010 2011 2012 2013 2014 2015 2016 2017 2009-2017
NSF 29,420 35,850 39,647 41,753 43,630 39,566 45,351 44,527 42,417 362,161
NIH 35,624 41,212 44,881 44,844 44,355 37,585 40,932 38,559 38,013 366,005
NSERC 9,037 10,527 11,693 12,250 12,422 11,203 13,229 12,927 12,868 106,156
EPSRC 4,577 5,512 5,966 6,304 6,623 5,824 7,163 7,682 8,231 57,882
CIHR 4,552 5,256 5,743 5,995 6,043 4,923 6,185 5,870 5,514 50,081
ERC 160 392 775 1,465 2,353 2,978 4,495 5,047 5,000 22,665
MRC 2,429 3,132 3,372 3,646 3,898 3,300 4,054 4,183 3,988 32,002
Wellcome trust 2,393 2,918 3,069 3,257 3,531 2,980 3,650 3,766 3,591 29,155
Gates 453 785 987 1,116 1,354 1,223 1,846 1,927 1,952 11,643
ESRC 336 452 566 631 713 634 1,544 1,764 1,683 8,323
SSHRC 194 269 284 355 337 349 1,318 1,375 1,415 5,896
-
17.
Share of OA by funder
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2010 2011 2012 2013 2014 2015 2016 2017
%ofOApapers
Publication year
United States
Gates
NIH
NSF
Others US
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2010 2011 2012 2013 2014 2015 2016 2017
%ofOApapers
Publication year
Canada
CIHR
NSERC
Others Can
SSHRC
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2010 2011 2012 2013 2014 2015 2016 2017
%ofOApapers
Publication year
United Kingdom
Wellcome
trust
MRC
EPSRC
ESRC
Others UK
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2010 2011 2012 2013 2014 2015 2016 2017
%ofOApapers
Publication year
EU
ERC
Other EU
papers
-
18.
Gold v. Green: N. America
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
SSHRC
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
NSERC
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
CIHR
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
NIH
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
NSF
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
Gates
Green Only Green and Gold
Green Only Green and Gold Gold Only
-
19.
Effect of international partners
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
NSERC
CIHR
SSHRC
EPSRC
Wellcome trust
BBSRC
NSF
ESRC
NIH
MRC
National
corresponding
author
Foreign
corresponding
author
-
20.
Gold v. Green: UK/Europe
Effect of
Finch
Report
(2012)?
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
ESRC
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
EPSRC
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
ERC
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
Wellcome Trust
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
MRC
Green Only Green and Gold
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2009 2011 2013 2015 2017
BBSRC
Green Only Green and Gold
Green Only Green and Gold Gold Only
-
21.
APCs and gold OA options
Springer: $2,093
Elsevier: $2,249
SciELO: $500
-
22.
Cultures of compliance
Funder
Biomedical
Research
Clinical
Medicine
Health
Mathematics
Earthand
Space
Psychology
Physics
Biology
Professional
Fields
Social
Sciences
Chemistry
Engineering
and
AllDisciplines
Wellcome trust 92% 84% 87% 96% 71% 80% 73% 88% 93% 74% 73% 79% 87%
NIH 93% 86% 79% 87% 73% 75% 84% 76% 74% 59% 81% 71% 87%
MRC 88% 75% 79% 87% 62% 62% 47% 83% 77% 73% 59% 50% 79%
Gates 89% 81% 83% 95% 50% 47% 51% 57% 28% 44% 52% 46% 79%
BBSRC 83% 71% 77% 90% 57% 44% 58% 68% 92% 52% 49% 52% 74%
ESRC 92% 76% 72% 70% 66% 60% 69% 60% 59% 63% 60% 56% 69%
ERC 80% 64% 59% 75% 82% 50% 75% 66% 46% 46% 36% 46% 67%
CIHR 71% 51% 52% 73% 43% 22% 36% 57% 47% 26% 25% 22% 56%
EPSRC 76% 64% 70% 78% 59% 54% 60% 68% 58% 62% 39% 49% 55%
NSF 76% 70% 52% 69% 54% 34% 48% 46% 35% 26% 24% 23% 47%
NSERC 57% 38% 42% 55% 31% 18% 40% 28% 14% 8% 10% 12% 30%
SSHRC 78% 35% 25% 40% 33% 17% 27% 36% 14% 16% 0% 17% 23%
All funded papers 85% 79% 73% 67% 57% 56% 56% 51% 42% 39% 35% 29% 66%
-
23.
Recommendations
2
3
1
Embargos
Infrastructure
Culture
4 APCs
-
24.
Cassidy R. Sugimoto
Vincent Larivière
csugimot@nsf.gov
vincent.lariviere@umontreal.ca
@csugimoto @lariviev
Questions?
Self-archiving of papers in an institutional or disciplinary repository or on an researcher’s website
Submitted version
Original proofs
Corrected proofs
Final version accepted by journal
Unpaywall is a browser extension that tracks the OA status of papers, based on paper information obtained from Crossref.
Aggregates links to OA papers DOAJ, PubMed Central, as well as thousands of journal websites and repositories
Excludes versions available through Sci-Hub or on social networking websites, such as Academia or ResearchGate
Also provides whether scholarly papers are available on a publishers’ website (Gold OA) or in a repository (Green OA)
Publishers’ version was preferred to the repository version, following Harnad’s distinction
Green OA papers are thus papers for which no gold OA version exists.
As of April 18th 2018, Unpaywall contained OA status for 95,842,233 DOIs from scholarly documents; making this information available through an API as well as a raw data file
Given the diversity of sources on which OA version of papers can be found (both legally and illegally), the Unpaywall algorithm’s accuracy errs more on the precision side than recall
Manual analysis of a sample of papers in Piwowar et al. (2018) has shown that, while an actual OA version could be found for 96.6% of papers which were considered OA, and OA version could be found for 12.3% of papers which were considered as closed. Hence, results presented here can be considered as a minimum proportion of papers available in OA.
Unpaywall records from the raw data file were matched with WoS papers published between 2008 and 2017, for a total number of journals articles analyzed of 12,495,074.
Unpaywall provides a lower bound of OA—there are more papers available through OA than what is in the tool.
But those are less easy to find, and have less permanence (such as on researchers’ websites)
Suggests that centralized repositories—such as arXiv, etc—might be more efficient than decentralized deposit (i.e. on researchers’ websites, etc.)
In some cases, some of the papers might have been funded a while ago, at a moment when
OA coverage of green is lower – priority given to gold OA.
Limited to Web of Science papers
Very little research on the topic, let alone on funders’ mandates
Gargouri et al (2012): Post Finch report work, aimed at assessing the following claim in the Finch report
"The [Green OA] policies of neither research funders nor universities themselves have yet had a major effect in ensuring that researchers make their publications accessible in institutional repositories…"
Vincent-Lamarre et al. (2016): what are the most efficient mandates?
Those for which deposit is mandatory
Researchers can opt out
Institutions keep the rights
Institutional mandates
Generally less restrictive
Many forms
University of Liège (evaluation of faculty)
Funders mandates
Considered to be more restrictive (contractual condition)
Many have embargo (typically 1 year)
Science as a public good (that is publicly funded)
“Obligation” for researchers to make their paper open access
Encouragement vs obligation
Embargo vs immediate deposit
Institutional / disciplinary repository vs any repository vs gold
Dark vs open deposit
Right retention vs no right retention
Opt out policy
How compliant are funded reseachers to funders’ mandates?
How does compliance vary across time, discipline and funder?
What is the relative importance of Gold and Green OA?
11 OA mandates are analyzed:
Canada (CIHR, NSERC, SSHRC)
United States (NIH, NSF, Bill and Melinda Gates Foundation)
United Kingdom (EPSRC, ESRC, MRC, and Wellcome Trust)
Europe (European Research Council)
Some mandates are relatively old, others more recent, and cover a large spectrum of disciplines
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
In-house version of the WoS, based on the XML files of the Science Citation Index Expanded (SCIE), the Social Science Citation Index (SSCI) and the Arts and Humanities Citation Index (AHCI)
Use of funding acknowledgements and of country of corresponding authors to assess whether a paper falls under the mandate of research council:
NSF papers: both have acknowledge funding from NSF and have a US-based corresponding author
Cleaning of funders’ names (NSF, U.S. NSF, National Science Foundation, National Science Foundation under the XXX grant, etc.)
Digital Object Identifiers (DOIs) are used to match Web of Science records with the OA status of paper from Unpaywall, thus restricted the analysis to journal articles that have unique DOIs between 2009 and 2017
Compliance = medicine > natural sciences > social sciences
Importance of infrastructure: high levels of deposit for medical disciplines may be due to the existence of dedicated repositories (i.e. PubMed Central)
Global effect of embargos
They seem to have become longer
Funders that allow them have much lower deposit rate
Decline of share of OA of both NIH and CIHR, which is difficult to explain
Effect of embargos? Measurement issue?
Canadian mandates have never been enforced, nor publicized. Researchers are not even aware that they exist.
Effects of providing dedicated funds for APCs : increase of Gold OA
Finch report and growth of gold in the UK
High cost (half a billion in APCs for the 11 funders at 1,500$ per paper)