Cassidy Sugimoto - Open Access Mandates: Compliance by funders
Download as PPTX, PDF1 like827 views
This document analyzes compliance with open access mandates by major research funders. It finds that compliance levels vary significantly by funder, discipline, and whether access is provided through gold or green open access. Funders in the biomedical field have the highest compliance rates, while those in the social sciences and humanities have the lowest. Compliance has increased overall in recent years, but green open access is more common than gold. The authors recommend that funders focus on standardizing embargoes, improving self-archiving infrastructure, shifting scholarly cultures, and addressing affordability of article processing charges to further increase open access to published research.
Cassidy Sugimoto - Open Access Mandates: Compliance by funders
- 1. OPEN ACCESS MANDATES: COMPLIANCE BY FUNDERS Cassidy R. Sugimoto Vincent Larivière [email protected] [email protected] @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)
- 5. Routes to open access Non-subscription Subscription APC / subsidy Self-archiving Gold OA Green OA Toll access
- 6. OA miners
- 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
- 13. Sample of funder mandates
- 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%
- 24. Cassidy R. Sugimoto Vincent Larivière [email protected] [email protected] @csugimoto @lariviev Questions?
Editor's Notes
- #6: 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
- #8: 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
- #9: 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
- #11: 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
- #12: 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)
- #13: “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
- #14: 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
- #16: 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
- #17: 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
- #18: 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
- #19: 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
- #20: 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
- #21: 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
- #23: 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
- #24: 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)