SPEC Kit 334: Research Data Management Services · 17
data). All but one of the libraries with archives use
persistent identifiers. Most common for IRs are the
Handle System (21, or 64%). Most of the data archives
use DOIs for datasets. Only eight archives use ARKs.
Several generate their own identifiers.
Finally, the survey asked about preservation ca-
pacities of archives, choosing standards most typical
for digital repositories. All five data archives provide
file integrity/fixity checking and multiple copy repli-
cation, as do 27 IRs (84%). Four data archives practice
geographic separation of backups, but this is less com-
mon for IRs (22, or 69%). Three data archives and half
of the IRs also provide format migration and conver-
sion for data files over time.
Overall responses about archive architecture re-
flect the different purposes of publication-oriented
IR platforms and archives that focus on the specific
needs of large and diverse research data collections. If
data archiving platforms become less resource inten-
sive for libraries to install and operate, future surveys
could gauge corresponding attitudes toward making
the library a center for data archiving. Here, the tra-
ditional cultural emphasis of libraries and archives
on long-term preservation and curation may support
libraries’ justification for taking on their operation. By
contrast, science domains and academic publishers
may emphasize shorter-term requirements of data
dissemination, yet may hesitate to invest in archiving
infrastructure. As interest grows in meeting public
funder requirements, and as research practices shift
toward data sharing for accelerated discovery and
collaboration, academic institutions may recognize
libraries as facilitators for research data. Survey re-
sponses suggest that infrastructure requirements are
significant, and implementation and adoption may be
slow. Data archiving by academic libraries, however,
is clearly an emerging field that future surveys and
case studies should follow.
It will be increasingly relevant to follow up on
libraries’ forays into data archiving, since 30 insti-
tutions indicated they plan to offer data archiving
within two years (Q56), but nearly a quarter of the
respondents included providing data archiving in
their top three challenges for RDM services (Q54),
most commenting on the difficulty in setting up in-
frastructure, from software to storage requirements.
RDM Service Staffing
The depth and range of RDM services that libraries
offer are, of course, directly proportional to staffing,
both in the number of positions and the amount of
time given to RDM activities when a position has other
responsibilities. Currently, the most prevalent organi-
zational structure for providing RDM services at the
53 responding libraries is a committee of staff from
departments within the library (27, or 51%)(Q32). Less
common organizational structures include a commit-
tee/group comprised of staff from across the univer-
sity, including the library (9, or 17%), a single position
within the library (8, or 15%), and a single department
within the library (6, or 11%).
At the libraries where RDM services are provided
by staff from different departments, no single depart-
ment dominates (Q33). About a quarter of the depart-
ments provide reference/liaison services, followed
by work in collections (19%), digital services (12%),
research/instruction (12%), and systems/IT (11%).
The range in RDMS position titles reported shows
that staff expertise is diverse and that no one type of
position dominates either (Q38).
50
38
18
17
13
12
11
11
10
9 9
Subject Librarian or
Liaison
Digital
Data Librarian
Metadata
Data Services
GIS or Geospa9al
Research Data
Cura9on
Repository
Soware or Systems
Data Management
Out of 231 positions, the most frequently reported title
is subject liaison/librarian (50 positions). One might
expect RDMS roles to be more prevalent for liaisons to
data). All but one of the libraries with archives use
persistent identifiers. Most common for IRs are the
Handle System (21, or 64%). Most of the data archives
use DOIs for datasets. Only eight archives use ARKs.
Several generate their own identifiers.
Finally, the survey asked about preservation ca-
pacities of archives, choosing standards most typical
for digital repositories. All five data archives provide
file integrity/fixity checking and multiple copy repli-
cation, as do 27 IRs (84%). Four data archives practice
geographic separation of backups, but this is less com-
mon for IRs (22, or 69%). Three data archives and half
of the IRs also provide format migration and conver-
sion for data files over time.
Overall responses about archive architecture re-
flect the different purposes of publication-oriented
IR platforms and archives that focus on the specific
needs of large and diverse research data collections. If
data archiving platforms become less resource inten-
sive for libraries to install and operate, future surveys
could gauge corresponding attitudes toward making
the library a center for data archiving. Here, the tra-
ditional cultural emphasis of libraries and archives
on long-term preservation and curation may support
libraries’ justification for taking on their operation. By
contrast, science domains and academic publishers
may emphasize shorter-term requirements of data
dissemination, yet may hesitate to invest in archiving
infrastructure. As interest grows in meeting public
funder requirements, and as research practices shift
toward data sharing for accelerated discovery and
collaboration, academic institutions may recognize
libraries as facilitators for research data. Survey re-
sponses suggest that infrastructure requirements are
significant, and implementation and adoption may be
slow. Data archiving by academic libraries, however,
is clearly an emerging field that future surveys and
case studies should follow.
It will be increasingly relevant to follow up on
libraries’ forays into data archiving, since 30 insti-
tutions indicated they plan to offer data archiving
within two years (Q56), but nearly a quarter of the
respondents included providing data archiving in
their top three challenges for RDM services (Q54),
most commenting on the difficulty in setting up in-
frastructure, from software to storage requirements.
RDM Service Staffing
The depth and range of RDM services that libraries
offer are, of course, directly proportional to staffing,
both in the number of positions and the amount of
time given to RDM activities when a position has other
responsibilities. Currently, the most prevalent organi-
zational structure for providing RDM services at the
53 responding libraries is a committee of staff from
departments within the library (27, or 51%)(Q32). Less
common organizational structures include a commit-
tee/group comprised of staff from across the univer-
sity, including the library (9, or 17%), a single position
within the library (8, or 15%), and a single department
within the library (6, or 11%).
At the libraries where RDM services are provided
by staff from different departments, no single depart-
ment dominates (Q33). About a quarter of the depart-
ments provide reference/liaison services, followed
by work in collections (19%), digital services (12%),
research/instruction (12%), and systems/IT (11%).
The range in RDMS position titles reported shows
that staff expertise is diverse and that no one type of
position dominates either (Q38).
50
38
18
17
13
12
11
11
10
9 9
Subject Librarian or
Liaison
Digital
Data Librarian
Metadata
Data Services
GIS or Geospa9al
Research Data
Cura9on
Repository
Soware or Systems
Data Management
Out of 231 positions, the most frequently reported title
is subject liaison/librarian (50 positions). One might
expect RDMS roles to be more prevalent for liaisons to
