<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Learning Hands-on and by Trial &amp; Error with Data Curation Profiles</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>D. Scott Brandt</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Purdue University Libraries West Lafayette</institution>
          ,
          <addr-line>IN</addr-line>
          <country country="US">US</country>
        </aff>
      </contrib-group>
      <abstract>
        <p>- The Data Curation Profiles Toolkit can be used in several ways to capture requirements for data sets, as articulated by researchers. As a flexible instrument it can facilitate dialog between librarians and researchers to discuss data concerns, current data workflow, and possible outcomes for the future. As a structured tool it can help identify areas of concern and need, to begin to making informed decisions about the data. Published Profiles offer insight into similarities and variations in data and data workflow, across multiple research areas or sub-disciplines. Designed as a tool for practitioners, it can help build knowledge and skill through application. Librarians who have completed Profiles have found the process to improve their comfort, increase their confidence, and build competencies in working with researchers. Mapping use of the Profiles Toolkit to the DigCurV Curriculum Framework can help identify strengths and weaknesses in what is currently a hands-on, trial and error selflearning approach.</p>
      </abstract>
      <kwd-group>
        <kwd>data curation</kwd>
        <kwd>competencies</kwd>
        <kwd>learning outcomes</kwd>
        <kwd>workshop</kwd>
        <kwd>self-learning</kwd>
        <kwd>Curriclum Framework</kwd>
        <kwd>Data Curation Profiles Toolkit</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>I. INTRODUCTION</title>
    </sec>
    <sec id="sec-2">
      <title>The Data Curation Profiles (DCP), and its associated</title>
      <p>
        Toolkit, is an outcome of a research investigation launched in
2007 to understand data sharing, particularly amongst single
or small groups of investigators [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ]. The instrument probes the
kinds of data researchers are working with, what they are
currently doing with it, and what they would like to do with it.
Librarians and other information professionals use the DCP
Toolkit to interview researchers, and the result can produce a
stand-alone publication describing a researcher’s intent and
needs related to a data set—the Data Curation Profiles
Directory currently has published 30 Profiles [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ]. In effect, the
DCP Toolkit supports the role of librarian or information
professional that finds herself or himself in the position of
mediator between the researcher and the archivist.
      </p>
    </sec>
    <sec id="sec-3">
      <title>There is no explicit curriculum for learning how to use the</title>
      <p>
        DCP Toolkit, which consists of a User Guide, Interviewer
Manual, Interviewee Worksheet and Profile Template. The US
Institute of Museum and Library Services, who funded the
original research project that produced the Toolkit, supported
a short project to teach librarians how to use it. Over 300
librarians attended a series of twelve workshops taught
between 2011-2012. The workshops were developed to guide
attendees in how to use the DCP Toolkit and what to expect
when doing so. The goal was for participants to gain comfort
using a tool, so that they could gain confidence in exploring
researcher needs and concerns, and later gain competence in
working with researchers on data management planning and
depositing data in a repository [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ].
      </p>
      <p>
        To this point training has consisted of a one-day session
focused on instruction in using the DCP Toolkit. It has
covered concepts of data curation as they relate to researchers
(identified in the original DCP research), how the components
of the Toolkit were designed to probe them, and detailed
information on working through two main sections, data
lifecycle and data sharing [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ]. Probing about the research data
lifecycle is seen as important for both researcher and librarian
because it uncovers what might otherwise be unrecognized:
that the research process can be seen as stages of data
collection and analysis that produce tangible products (data
sets) in addition to intellectual findings disseminated in
articles. Probing about data sharing also is seen as important
because it helps identify researcher concerns in doing so.
      </p>
    </sec>
    <sec id="sec-4">
      <title>An instructional design approach, ADDIE, was used in</title>
      <p>
        developing the workshop. It required identification of users’
needs (how to use the Toolkit), developing specific learning
outcomes (application of skills), designing learning objectives
to meet the outcomes through lecture, discussion, exercises
and additional resources, to scaffold learning (scenarios) [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ].
Assessing the workshop with the Framework would be helpful
in building and expanding training. Initially it is assumed that
training will continue to be generic in regard to digital curation
overall, but specific in focusing on the particular use of the
Toolkit. It is also hoped that the framework will help us
describe the value of the training in a meaningful way for a
broader range of people.
      </p>
      <p>II. CURRICULUM FRAMEWORK AS ASSESSMENT TOOL
The Framework is meant to be useful to those building
new training courses. Depending on the user’s aims, the
Framework can assist in providing a structure for a generic
training program for the role of digital curator, or it can
suggest which subjects should be covered in shorter, more
specialized courses addressing one particular area of
professional digital curation practice. The Framework may
also supply a common language to allow those building and
developing training to meaningfully describe the value of their
training offerings.</p>
      <p>Utilizing the Framework to assess the DCP Toolkit and
associated learning initially seems somewhat problematic.
First, training is focused on emulating use of the DCP Toolkit.
The workshop gives context (lecture), offers instruction
(reading), provides modeling (videos) and employs problem
solving (exercises). Basically, the learning objectives are
meant to facilitate self-learning through scenarios of how
others employed the DCP Toolkit, and providing expert
feedback by workshop leader during discussion. Second, users
of the DCP Toolkit are not learning about data curation per se
(i.e., digital preservation), they seek to learn what researchers
are doing with their research (i.e., data management). This is
because the DCP Toolkit supports the role of librarian or
information professional in the position of mediator between
the researcher and the archivist. However, the Framework can
still be helpful in assessing learning goals and outcomes.</p>
    </sec>
    <sec id="sec-5">
      <title>III. ASSESSING WITH THE FRAMEWORK</title>
      <p>The Framework looks first at the knowledge and principles
to be learned. The DCP Toolkit outcomes map especially well
with understanding research data management. That is, the
elements under the Framework map with the goal of
understanding the concepts and terminology employed in
research data collection and analysis, organizing and
describing or documenting it, and meeting funder or
institutional requirements. However, beyond a general data
lifecycle, it is hard to anticipate what someone will encounter
when meeting with a researcher—methodology differs
between physical, life and social science, not to mention the
humanities. What has been key, is to represent a data lifecycle
in a data table depicting various stages of collecting data,
processing or anonymizing it, analyzing it, and depositing and
publishing it. Classifying data stages can help clarify what will
or won’t or can’t be shared.</p>
    </sec>
    <sec id="sec-6">
      <title>Assessment with the Framework at this level raises the</title>
      <p>question of how much knowledge about digital curation a
librarian needs to interview a researcher to get “the story”
around his research data and workflow. Up to now, the goal of
using the DCP Toolkit is to gain insight and to gather many
Profiles so they can be studied to understand research data in
broader terms, in order to give greater context to digital
curation. However, to fulfill a mediating role and provide
specific services would demand more knowledge and
application of related principles. For instance, to make
suggestions about depositing a mediator would need to know
more about requirements related to file formats (i.e., which are
more appropriate for preservation).</p>
      <p>In assessing skills and competences, it can be argued in
this case that there are three perspectives: those related to the
researcher, to the librarian or information professional (as
mediator) and to the archivist. Subject knowledge in the
discipline relates to the researcher, and anyone who uses the
DCP Toolkit to interview should have some familiarity with
the discipline (i.e., background, or should review the
researchers work to become familiar with it). Skills for the
librarian or information professional include first those that
help in interviewing the researcher, which are enumerated in
the Framework: creativity (inquiring mind), professional
conduct (ethics regarding confidentiality, familiarity with
institutional policies), and communication (articulating and
clarifying needs). Skills for the librarian or information
professional also include those related to creating Profiles to
publish: creativity (an inquiring mind to pursue the interview),
personal qualities (able to engage in deep conversation),
selfmanagement (self-initiative) professional conduct (clear and
accurate reporting of information synthesized), research
management (project planning/delivery) and, obviously,
communication. Competence is required, even if these are
considered “soft skills;” although it is not clear how to teach
them. This is an area in particular for which the workshop
employs hands-on practice, and practicing through trial and
error is important.</p>
    </sec>
    <sec id="sec-7">
      <title>IV. AMBIGUITIES IN DESIGN</title>
      <p>Understanding how the DCP Toolkit fits with a digital
curation lifecycle depends mostly on where or how the
research lifecycle is seen as having overlap. As with skills and
competences, learning here may depend on three perspectives,
the researcher, the librarian or informational professional and
the archivist. Researchers may not see themselves directly
involved in digital curation. However, they conceptualize the
research project and associated data, and the intersection with
data management planning. Obviously they create or receive
data to analyze, they perform a kind of appraisal and selection
determining what they will share or publish of a given project,
and they allow access (e.g., usually to peers). Librarians and
information professionals need to understand how and where
the research lifecycle intersects the digital curation cycle, and
when or where to work with archivists and preservation.</p>
    </sec>
    <sec id="sec-8">
      <title>We have argued up to now that using the DCP Toolkit can</title>
      <p>
        be learned through hands-on application and trial and error.
The Toolkit includes interview questions and a worksheet,
along with suggestions for how to ask questions, what to focus
on, and how to stay on track. By following general instructions
on how to probe data needs and concerns, a practitioner can
demonstrate success at a performing a set of associated tasks,
which evidences learning. As there are no “right” answers,
practitioners can (must) use trial and error to work through the
interview process, collecting information and distilling
information into a Profile. Trial and error is important because
“[W]hen the problem space is too large to explore completely,
a learning agent must have the ability to guess about new
situations based on experience with similar situations” [
        <xref ref-type="bibr" rid="ref6">6</xref>
        ]. In
previous workshops, several teaching methods were used:
lecture, small group exercise, manual, one-to-one training, and
workshops. The Framework helps understand how training will
likely require using an online approach that integrates
webinars, readings and videos and allowing learners to move at
their own pace. Without funding for more workshops, it will be
critical to use the Framework as a guide to create a tutorial in
which there will be no expert leading the lessons.
      </p>
      <p>V. CONCLUSION</p>
      <p>
        There is a paradox that is highlighted by using the
Framework in assessing this situation. The knowledge to be
gained is about the researcher’s data (workflow, etc.), yet the
skill is about learning to explore and understand the needs
before being able to attend to them. As the gatekeeper, some
researchers hold onto their data for reasons associated with
lack of time to do more and lack of understanding how to
organize and disseminate it. For the researcher to relinquish
the gatekeeper goal, librarians must understand the
researchers’ perspective, context, situation and needs. They
must, in effect, learn to use a tool to learn about the researcher
data, and then learn what can be done with it [
        <xref ref-type="bibr" rid="ref7">7</xref>
        ].
      </p>
      <p>Trying to map the use of the DCP Toolkit to the
Framework reveals some weaknesses and strengths in relying
on the Toolkit itself to facilitate “self-training,” and a
traditional training approach. The Framework is meant to be
useful to those building new training courses, but can give
insight into assessing for changes or using different
approaches, which will be a next step for us. While the
Framework doesn’t assist in providing a structure for a generic
training program, it has helped identify the need to clarify
perspectives, clarify which subjects should be covered, and
where more specialized coursework addressing professional
digital curation practice would be helpful. With further
application, the Framework will help articulate better the value
of learning how to use, and use, the DCP Toolkit.</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          [1]
          <string-name>
            <given-names>Data</given-names>
            <surname>Curation Profiles Toolkit</surname>
          </string-name>
          . http://datacurationprofiles.org
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          [2]
          <string-name>
            <given-names>Data</given-names>
            <surname>Curation Profiles Directory</surname>
          </string-name>
          . http://docs.lib.purdue.edu/dcp/
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          <source>[3] [Data Curation Profiles Symposium, September</source>
          <volume>24</volume>
          ,
          <year>2012</year>
          , Purdue University, West Lafayette IN http://docs.lib.purdue.edu/dcpsymposium/
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          [4]
          <string-name>
            <surname>Carlson</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          “
          <article-title>Demystifying the data interview: developing a foundation for reference librarians to talk with researchers about their data”</article-title>
          <source>Reference Services Review</source>
          <volume>40</volume>
          (
          <issue>1</issue>
          ):
          <year>2012</year>
          .
          <fpage>7</fpage>
          -
          <lpage>23</lpage>
          . http://docs.lib.purdue.edu/lib_research/153/
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          [5]
          <string-name>
            <surname>Brandt</surname>
            ,
            <given-names>D. S.</given-names>
          </string-name>
          <string-name>
            <surname>Teaching Technology</surname>
          </string-name>
          :
          <article-title>A How-To-Do-It Manual for Librarians. How-To-Do-It Manuals for Librarians</article-title>
          .
          <string-name>
            <surname>Neal-Schuman</surname>
            <given-names>Publishers</given-names>
          </string-name>
          , Inc., 100 Varick St., New York, NY,
          <year>2002</year>
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          [6]
          <string-name>
            <surname>Lin</surname>
            ,
            <given-names>L. J.</given-names>
          </string-name>
          <article-title>Self-improving reactive agents based on reinforcement learning, planning and teaching</article-title>
          .
          <source>Machine learning</source>
          ,
          <volume>8</volume>
          (
          <issue>3-4</issue>
          ),
          <year>1992</year>
          ,
          <fpage>293</fpage>
          -
          <lpage>321</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          [7]
          <string-name>
            <surname>Brandt</surname>
            ,
            <given-names>D. S. “</given-names>
          </string-name>
          <article-title>Disambiguating the role of data lifecycle gatekeeper</article-title>
          ,
          <source>” Workshop on Research Data Lifecycle Management, July 18-20</source>
          ,
          <year>2011</year>
          , Princeton University, Princeton NJ.
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>