12 · SPEC Kit 300
management systems. For externally hosted titles,
they most often use OPAC records (94%), open
URL resolvers (79%), Web pages, (73%), or a third-
party title list or portal (67%). As one respondent
explained, they use “all the same channels as non-
OA resources.”
The survey asked for an approximate number
of OA titles linked to. Many respondents noted the
difficulty or impossibility of providing counts of
titles by specific categories. However, from those
who attempted an estimate, the three types of lo-
cally hosted resources with the highest median title
counts were government documents (15,050 titles),
theses and dissertations (493), and technical reports
(170). The externally hosted resources with the
highest median title counts were government doc-
uments (37,155), theses and dissertations (3,500),
journals (3,102), and legal documents (2,000).
Selection and Financial Support
Most of the responding libraries (84%) do not have
collection development policies that specifically
address criteria for selecting externally hosted OA
resources, though several libraries report that col-
lection development policies are undergoing revi-
sion and there are plans or desires to address OA
resources in their policies. Typically, the selection
criteria (79%) and the selectors (91%) for externally
hosted OA resources are the same as for other elec-
tronic resources. Some libraries provide access to
all titles in the Directory of Open Access Journals
(DOAJ) rather than making title-level decisions.
Because they are usually not supported by sub-
scription fees, many open access journals get their
funding through fees paid by authors or their insti-
tutions. These author fees are often paid by librar-
ies at the author’s institution. The majority of re-
sponding libraries (68%) provide financial support
for externally hosted OA journals, either directly or
through a consortium. Many contribute to BioMed
Central, Public Library of Science (PLoS), and the
Stanford Encyclopedia of Philosophy, among oth-
ers. Only a few (36%) provide financial support
for internally hosted OA journals. Although some
don’t provide direct financial support for these re-
sources, they do provide staff time and support for
processing.
Cataloging
For externally hosted OA resources, the most popu-
lar methods of obtaining catalog records are down-
loading records from OCLC (86%), creating origi-
nal, full records in the library (69%), and acquiring
records from another third party such as Serials
Solutions (56%). About a third also create original,
brief records or acquire them from other libraries
such as consortial partners. Catalog records for
locally hosted OA resources are most commonly
obtained by creating original, full records (87%),
creating original, brief records (56%), and down-
loading records from OCLC (42%). Other methods
of obtaining descriptive data for both locally and
externally hosted OA resources include the auto-
mated harvesting of XML metadata, records cre-
ated by the authors, brief records created by an
electronic resource management system, and re-
cords provided by the publishers. In all but three of
the responding libraries (95%), the staff who create
catalog records for OA resources are the same staff
who create records for other electronic resources.
Although two libraries report spending 100
to 120 hours per month cataloging OA resources,
most spend fewer than 20 hours per month. Others
don’t track the time. Several respondents pointed
out that cataloging of OA material was a very mi-
nor portion of cataloging work, perhaps less than
1% of total time spent on cataloging. Others em-
phasized that they did not treat the cataloging of
OA resources differently than non-OA resources—
catalogers simply integrate work with these mate-
rial into their usual workload. In many libraries,
both professional catalogers and support staff cre-
ate catalog records. Most of the libraries (89%) that
create original records for OA resources contribute
them to OCLC. Of the libraries that include links
to OA resources in the OPAC, 40% identify these
management systems. For externally hosted titles,
they most often use OPAC records (94%), open
URL resolvers (79%), Web pages, (73%), or a third-
party title list or portal (67%). As one respondent
explained, they use “all the same channels as non-
OA resources.”
The survey asked for an approximate number
of OA titles linked to. Many respondents noted the
difficulty or impossibility of providing counts of
titles by specific categories. However, from those
who attempted an estimate, the three types of lo-
cally hosted resources with the highest median title
counts were government documents (15,050 titles),
theses and dissertations (493), and technical reports
(170). The externally hosted resources with the
highest median title counts were government doc-
uments (37,155), theses and dissertations (3,500),
journals (3,102), and legal documents (2,000).
Selection and Financial Support
Most of the responding libraries (84%) do not have
collection development policies that specifically
address criteria for selecting externally hosted OA
resources, though several libraries report that col-
lection development policies are undergoing revi-
sion and there are plans or desires to address OA
resources in their policies. Typically, the selection
criteria (79%) and the selectors (91%) for externally
hosted OA resources are the same as for other elec-
tronic resources. Some libraries provide access to
all titles in the Directory of Open Access Journals
(DOAJ) rather than making title-level decisions.
Because they are usually not supported by sub-
scription fees, many open access journals get their
funding through fees paid by authors or their insti-
tutions. These author fees are often paid by librar-
ies at the author’s institution. The majority of re-
sponding libraries (68%) provide financial support
for externally hosted OA journals, either directly or
through a consortium. Many contribute to BioMed
Central, Public Library of Science (PLoS), and the
Stanford Encyclopedia of Philosophy, among oth-
ers. Only a few (36%) provide financial support
for internally hosted OA journals. Although some
don’t provide direct financial support for these re-
sources, they do provide staff time and support for
processing.
Cataloging
For externally hosted OA resources, the most popu-
lar methods of obtaining catalog records are down-
loading records from OCLC (86%), creating origi-
nal, full records in the library (69%), and acquiring
records from another third party such as Serials
Solutions (56%). About a third also create original,
brief records or acquire them from other libraries
such as consortial partners. Catalog records for
locally hosted OA resources are most commonly
obtained by creating original, full records (87%),
creating original, brief records (56%), and down-
loading records from OCLC (42%). Other methods
of obtaining descriptive data for both locally and
externally hosted OA resources include the auto-
mated harvesting of XML metadata, records cre-
ated by the authors, brief records created by an
electronic resource management system, and re-
cords provided by the publishers. In all but three of
the responding libraries (95%), the staff who create
catalog records for OA resources are the same staff
who create records for other electronic resources.
Although two libraries report spending 100
to 120 hours per month cataloging OA resources,
most spend fewer than 20 hours per month. Others
don’t track the time. Several respondents pointed
out that cataloging of OA material was a very mi-
nor portion of cataloging work, perhaps less than
1% of total time spent on cataloging. Others em-
phasized that they did not treat the cataloging of
OA resources differently than non-OA resources—
catalogers simply integrate work with these mate-
rial into their usual workload. In many libraries,
both professional catalogers and support staff cre-
ate catalog records. Most of the libraries (89%) that
create original records for OA resources contribute
them to OCLC. Of the libraries that include links
to OA resources in the OPAC, 40% identify these
