How AI, OpenAI, and ChatGPT impact business and software.
Stevan Harnad: On Designing Green Open Access Self-Archiving Mandates for Universities and Research Funders
1. Why "green" open access self-archiving mandates must
come before "gold" open access publishing
or
“Don’t over-reach: grasp what’s within immediate reach
StevanHarnad
Chaire de recherche du Canada en sciences cognitives
Université de Québec à Montreal
&
Electronics and Computer Science
University of Southampton
FNRS Bruxelles 28 septembre 2011
2. What is OA?
• Free
• Immediate
• Permanent
• Online
• Access
FNRS Bruxelles 28 septembre 2011
3. Open access to what?
• 2.5 million articles
• Published yearly
• In 25,000 peer reviewed journals
• Across all scientific and scholarly disciplines
• In all countries and languages
• Don’t over-reach for open books, “open
knowledge,” “open information”
FNRS Bruxelles 28 septembre 2011
4. Why provide OA?
• Because it maximizes research uses and
impact
• Making it accessible to all users
• Not just those who can afford to subscribe to
the journal in which the article was published
FNRS Bruxelles 28 septembre 2011
5. How to provide OA?
1. By publishing in any journal at all, but self-
archiving a copy free for all on the web
2. By publishing in a journal that makes its
articles free for all on the web
FNRS Bruxelles 28 septembre 2011
6. What are "green" and " " OA?
• OA self-archiving is green
• OA publishing is gold OA
FNRS Bruxelles 28 septembre 2011
7. Why does green OA need to come
before gold OA?
1. Because green OA can be provided by the
research community and gold OA can only be
provided by the publisher community
2. Because green OA can be mandated by
research institutions and funders and gold
OA cannot
3. Because the money to pay for gold OA is still
locked up in journal subscriptions
4. Don’t over-reach for immediate gold OA
FNRS Bruxelles 28 septembre 2011
8. "Gratis" OA vs "Libre" OA
• Gratis OA means free online access
• Libre OA means free online access plus certain
further re-use rights
• Don’t over-reach for libre OA
FNRS Bruxelles 28 septembre 2011
9. Institutional repositories vs central repositories
Institutional vs funder mandates
• Deposit institutionally, harvest centrally
• Don’t over-reach for direct central deposit
FNRS Bruxelles 28 septembre 2011
11. Copyright, Embargoes
• Desirable but not necessary to retain
copyright
• Immediategreen OA self-archiving can be
mandated even if authors wish to comply with
publisher OA embargoes
• Don’t over-reach for copyright retention
FNRS Bruxelles 28 septembre 2011
12. The optimalgreen OA self-archiving
mandate
• Immediate Deposit/”Optional OA”
• Deposit is sole means of submitting research
for institutional performance review (“Liège
model”)
• “email eprint request” button for embargoed
deposits
• Don’t over-reach for immediate-OA mandates
FNRS Bruxelles 28 septembre 2011
13. Leveraged transition
1. Institutions and funders mandate ID/OA
2. Universal green OA
3. Subscriptions become unsustainable
4. Publishers downsize to peer review alone
5. Offload access-provision and archiving on
institutional repositories
6. Journals convert to gold OA
7. Gold (much lower) OA fees paid out of institutional
windfall subscription cancelation savings
FNRS Bruxelles 28 septembre 2011
14. No-fault peer review
• Instead of charging for publication
• Journals charge for each round go peer review
• “No fault”
• Rejected paper charges no longer wrapped
into accepted paper fees
• Journals not tempted to lower standards for
more revenue
FNRS Bruxelles 28 septembre 2011
15. Versions
• Author’s peer-reviewed final draft has the
fewest publisher restrictions on it
• Don’t over-reach for publisher’s version-of-
record
FNRS Bruxelles 28 septembre 2011
16. Why does green OA need to come
before gold OA?
1. Because green OA can be provided by the
research community and gold OA can only be
provided by the publisher community
2. Because green OA can be mandated by
research institutions and funders and gold
OA cannot
3. Because the money to pay for gold OA is still
locked up in journal subscriptions
4. Don’t over-reach for immediate gold OA
FNRS Bruxelles 28 septembre 2011
18. Mandated green OA levels (70%) vs self-selective (unmandated) green OA
levels (20%)
90%
80%
70%
60%
2002
50% 2003
2004
40%
2005
2006
30%
2007
20% 2008
2009
10%
0%
N= 63,518
FNRS Bruxelles 28 septembre 2011
19. Logistic regression
Set of 13 variables (plus one interaction) potentially influencing citation counts.
FNRS Bruxelles 28 septembre 2011
20. Exp(B)-1 values of logistic regression (N = 63,518)
3 Age
JIF
Auth_N
Ref_N
Page_N
2
Sci
USA
OA
Age*OA
1 M
CERN
South
Minho
Queens
0
N=24,893 & 22,281 22,281& 5,796 22,281 & 4,798 22,281 & 2,834
zero/lo lo/med-lo lo/med-hi lo/hi
0 vs 1-4 cites 1-4 vs 5-9 cites 1-4 vs 10-19 cites 1-4 vs 20+ cites
-1
FNRS Bruxelles 28 septembre 2011
21. Cumulative number of citations per article for Mandated and Self-
SelectedOAvsNOA articles (articles published in 2002)
25
20
Cumulative count of citations
15
OM
ØM
10 OS
ØS
5
0
2002 2003 2004 2005 2006 2007 2008 2009
OM N = 897
ØM N = 493
OS N = 1,098
ØS N = 3,269
FNRS Bruxelles 28 septembre 2011
22. Separate proportions for OAvsNOA articles at each citation count
50%
45%
N = 63,518 N(OA) = 44,497 N(NOA)= 19,021
40%
35%
30%
25% NOA
20% OA
15%
10%
5%
0%
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 >30
FNRS Bruxelles 28 septembre 2011
23. Average cumulative number of citations per article as a function of article
age for articles published in 1998-2009
FNRS Bruxelles 28 septembre 2011
24. Average cumulative number of citations per article as a function of article
age by field for articles published in 2000
FNRS Bruxelles 28 septembre 2011
25. Separate proportions for OA vsNOA articles at each citation count
FNRS Bruxelles 28 septembre 2011
27. OA citation advantage for mandated vsunmandatedOA
4 institutions (N = 63,518)
0.5
2002
0.4
2003
2004
0.3
2005
Citation ratios
2006
0.2
2007
2008
0.1
2009
0.0
O/Ø O S/Ø O M/Ø O S/Ø S O M/Ø S O M/Ø M O M/O S
FNRS Bruxelles 28 septembre 2011
28. ROAR &ROARMAP
Registry of Open Access Repositories
Registry of Open Access Repository Archiving Mandates
• AGE : age of institution’s deposit mandate (in months)
• STRENGTH: strength of institution’s deposit mandate
• deposits : total number of deposits in institutional repository
• rate : rate of deposit (number of days per year with 10-99 deposits)
• RANK : Webometrics rank of institution
FNRS Bruxelles 28 septembre 2011
29. Scatter plot matrix :
pairwise scatter plots of all variables
FNRS Bruxelles 28 septembre 2011
30. Leveraged transition
1. Institutions and funders mandate ID/OA
2. Universal green OA
3. Subscriptions become unsustainable
4. Publishers downsize to peer review alone
5. Offload access-provision and archiving on
institutional repositories
6. Journals convert to gold OA
7. Gold (much lower) OA fees paid out of institutional
windfall subscription cancelation savings
FNRS Bruxelles 28 septembre 2011
31. Grasp what is within immediate reach
Don’t keep over-reaching and getting next to
nothing for yet another decade…
90%
80%
70%
60% 2002
50% 2003
2004
40%
2005
30%
2006
20%
2007
10%
2008
0%
2009
N= 63,518
FNRS Bruxelles 28 septembre 2011
- In order to display the OA advantage in a different way and show more graphically how it is distributed across the citation ranges, we calculated the proportions of OA articles at each citation count and compared them to those for NOA articles.- This Figure shows that as the citation count increases, the fraction of the total OA articles at each citation count overtakes the corresponding fraction of the total NOA articles.
In order to explore the evolution of OA advantage over time and by field, we plot the average cumulative number of citations per article as a function of article age, by field, for articles published in 2000.These figures show for each field, the average cumulative number of citations per article for OA and non-OA articles. For fields, as an article gets older, and its cumulative citations grow, its OA advantage grows too. However, this advantage differ in magnitude depending on the field; it is relatively high for Chemistery, ClininicalMedecine, Physics and Biomedical Research and is relatively weak in Mathematics, Biology, Social Science and Health.
These figures show that as the citation count increases, the fraction of the total OA articles at each citation count overtakes the corresponding fraction of the total NOA articles. The overtaking starts from 2 or 3 citations for Engineering, Mathematics and Social Science; from 4 or 5 citations and more for Biology; from 8 citations and more for Clinical Medecine; from 9 or 10 citations and more for Physics, Psychology, Chemistery and Earth Science; and from 11 or 12 citations and more for Biomedical research. However, for Health, the overtake is irregular for different levels of citation counts