4 Survey Results: Executive Summary
Figure 1. Demand for data curation services by subject domain
The nascent nature of data curation services and treatments across the ARL institutional
landscape is evident in a number of results from this survey. Although the Office of Science and
Technology Policy memo on access to federally funded scientific data was released in 2013,11 library
technical and human infrastructure are just now reaching the point of accepting and curating data. Of
the 46 libraries that accept data, the majority (26 or 61%) have fewer than 50 data sets in their entire
collection. Ten libraries have between 51 and 200 data sets but only seven report having over 200 in their
repository. The growth of data deposits seems to be consistent, with 14 libraries receiving approximately
one new dataset a month, and three receiving more than 10 a month.
Describing data sets using standard metadata schemas is of significant importance for data
discovery, dissemination, and reuse. Yet, there are many schemas to choose from, including general,
discipline-specific, and institution specific. Survey respondents indicate six major metadata schemas are
in use: DublinCore, MODS, DDI, DataCite, Dataverse (which is based on a number of standards)12, and
MARC. A number of institutions also employ others, such as ISO19115, Geoblacklight, and VRACore4,
or custom metadata schemas. Additionally, many organizations use more than one schema for different
purposes, and some institutions reported they use up to four.
Many of the data repository services and technologies facilitate both self-deposit and mediated
deposit of data (22 of 46 responses or 48%). The majority of data repositories limit the size of file uploads
(35 or 78%) with an average reported at around 2.5 GB per file. Thirty-two of the responding libraries
(65%) also help researchers prepare their data for deposit to external repositories. As shown in the
graph below, the external data repositories they support most often are ICPSR, Figshare, and the Open
Science Framework.
Figure 1. Demand for data curation services by subject domain
The nascent nature of data curation services and treatments across the ARL institutional
landscape is evident in a number of results from this survey. Although the Office of Science and
Technology Policy memo on access to federally funded scientific data was released in 2013,11 library
technical and human infrastructure are just now reaching the point of accepting and curating data. Of
the 46 libraries that accept data, the majority (26 or 61%) have fewer than 50 data sets in their entire
collection. Ten libraries have between 51 and 200 data sets but only seven report having over 200 in their
repository. The growth of data deposits seems to be consistent, with 14 libraries receiving approximately
one new dataset a month, and three receiving more than 10 a month.
Describing data sets using standard metadata schemas is of significant importance for data
discovery, dissemination, and reuse. Yet, there are many schemas to choose from, including general,
discipline-specific, and institution specific. Survey respondents indicate six major metadata schemas are
in use: DublinCore, MODS, DDI, DataCite, Dataverse (which is based on a number of standards)12, and
MARC. A number of institutions also employ others, such as ISO19115, Geoblacklight, and VRACore4,
or custom metadata schemas. Additionally, many organizations use more than one schema for different
purposes, and some institutions reported they use up to four.
Many of the data repository services and technologies facilitate both self-deposit and mediated
deposit of data (22 of 46 responses or 48%). The majority of data repositories limit the size of file uploads
(35 or 78%) with an average reported at around 2.5 GB per file. Thirty-two of the responding libraries
(65%) also help researchers prepare their data for deposit to external repositories. As shown in the
graph below, the external data repositories they support most often are ICPSR, Figshare, and the Open
Science Framework.
