<!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>PerspectivesX: A Proposed Tool to Scaffold Collaborative Learning Activities within MOOCs</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Aneesha Bakharia</string-name>
          <email>a.bakharia@uq.edu.au</email>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>UQx</institution>
          ,
          <addr-line>ITaLI</addr-line>
          ,
          <institution>The University of Queensland</institution>
          ,
          <addr-line>St Lucia, Brisbane</addr-line>
          ,
          <country country="AU">Australia</country>
        </aff>
      </contrib-group>
      <pub-date>
        <year>2017</year>
      </pub-date>
      <fpage>59</fpage>
      <lpage>64</lpage>
      <abstract>
        <p>In this work-in-progress paper, we introduce the PerspectivesX tool which aims to scaffold collaborative learning activities within MOOCs. The PerspectivesX tool has been designed to promote learner knowledge construction and curation for a range of multi-perspective elaboration techniques (e.g., SWOT analysis and Six Thinking Hats). The PerspectivesX tool is designed to store learner submissions in a searchable knowledge base which is able to be persisted across course re-runs and promotes the use of natural language processing techniques to allow course moderators to provide scalable feedback. In this paper we outline the design principles that structured collaborative learning tools need to adhere to, design a prototype tool (PerspectivesX) and evaluate whether MOOC platform extension frameworks are able to support the implementation of the tool.</p>
      </abstract>
      <kwd-group>
        <kwd>computer supported collaborative learning</kwd>
        <kwd>massive open online courses</kwd>
        <kwd>edX Xblock</kwd>
        <kwd>learning tools interoperability</kwd>
        <kwd>knowledge construction</kwd>
        <kwd>critical thinking</kwd>
        <kwd>idea generation</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>
        The tool predominantly used in MOOCs to foster collaborative learning is the
discussion forum. Research has shown that learners that actively contribute to the
course forum, are more likely to complete the course and achieve higher grades [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ]. A
high percentage of learners however, don’t engage in a course discussion forum, with
recent estimates of forum participation being between 5-10% of participants [
        <xref ref-type="bibr" rid="ref7">7</xref>
        ]. There
currently exists a wide gap between the unstructured collaborative nature of forums and
other MOOC instructional content (i.e., videos, quizzes and social polls). Tools that are
able to scaffold collaborative learning activities are required.
      </p>
      <p>
        Computer Supported Collaborative Learning (CSCL) and the ideas behind providing
scripted learning activities has a long and rich research history. Unfortunately, the
theoretical underpinnings and practical manifestations of CSCL have all but been
forgotten in the era of the MOOC. In this paper, the PerspectivesX tool is introduced.
The PerspectivesX tool implements concepts from CSCL scripting; and the Knowledge
Community and Inquiry model (KCI) [
        <xref ref-type="bibr" rid="ref11">11</xref>
        ]. KCI uses Web 2.0 tools to add a layer of
collective knowledge building to scripted learning activities.
2
      </p>
    </sec>
    <sec id="sec-2">
      <title>The PerspectivesX Tool</title>
      <p>
        The PerspectivesX tool is able to scaffold a range of multi-perspective elaboration
activities. The tool is designed to promote active participation from learners that are
either not participating in a discussion forum or that are passive forum participants (i.e.,
only reading forum posts). PerspectivesX encourages learners to make a contribution
and also makes it easy for learners to explore, review and curate other learners
submissions. In a PerspectivesX activity, learners must think about a problem from an
assigned or selected perspective and actively contribute their ideas to a knowledge base
that is available to all course participants. Instructors can enable an optional curation
layer that requires learners to collate ideas from fellow learners in order to complete the
remaining perspectives of the activity. Curation is an important feature of the tool.
Curation is a 21st century digital literacy that is able to facilitate the development of
learner search and evaluation strategies as well as promote critical thinking, problem
solving, and participation in networked conversations [
        <xref ref-type="bibr" rid="ref10">10</xref>
        ].
      </p>
      <p>
        Example activities that the PerspectivesX tool is able to scaffold includes learner
submissions for design projects (i.e., knowledge construction), reflective journal entries
(i.e., critical thinking) and multi-perspective elaboration activities (i.e., idea
generation). The suggested approach will be able to support a range of idea generation
and multi-perspective activities such Strengths, Weakness, Opportunities and Threats
(SWOT) analysis, Six Thinking Hats [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ], Fishbowl [
        <xref ref-type="bibr" rid="ref8">8</xref>
        ] and SCAMPER [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ].
3
      </p>
    </sec>
    <sec id="sec-3">
      <title>Design Principles</title>
      <p>
        The design principles that underpin the PerspectivesX tool are outlined below:
- Support the design of structured knowledge construction, critical thinking
and multi-perspective elaboration activities. Instructors should be able to design
activities that are able to collate structured responses/submissions from learners.
The types of responses required by learners should be exible and allow learners to
submit multiple free text responses, media artifacts (e.g., images, infographics,
slides, videos, etc) and links to external resources (e.g., website links). Within
multi-perspective activities the instructor should be able to design activities that
allow the learner to select a perspective or be randomly assigned to perspective.
- Support opt-in and anonymous learner knowledge sharing. Learners should not
be forced to share their submissions with other course participants. Between 5-10%
of learners are active discussion forum participants in a MOOC while a larger
percentage of learners read forum posts (i.e. passive participation). Many learners
may not feel confident making their submissions available to other learners in a
non-anonymous environment. Submission should be mandatory in order to receive
a participation grade but the learner should be able to opt-out of sharing or choose
to be anonymous.
- Support instructor moderation. Course moderators need the ability to review and
curate useful learner contributions. Curated content will help learners to focus their
attention on relevant and important submissions [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ] from other learners. The
learner should be able to view moderator highlighted content in an accessible and
intuitive manner. This will give moderators the ability to use learner submitted
work as a starting point to trigger active participation in a discussion forum.
Support learner curation. The scripted collaborative activity should allow for the
inclusion of a learner curation sub-activity. As an illustrative example, the
collaborative activity might require the learner to submit a single section of a
SWOT activity (e.g., strengths) and then at a later stage, curate content from other
course participants for the other sections (e.g., weaknesses, opportunities and
threats).
      </p>
      <p>Support temporal independence. Both paced and self-paced MOOCs should be
able to include scaffolded collaborative learning activities. Learners should be able
to contribute to the activity at any time as well as review and curate the
submissions of other learners in a time independent manner. This is particularly
important for self-paced MOOCs where learners are able to commence a course at
any time and as a result would engage in collaborative learning activities at
different times. Discussion forums within self-paced MOOCs are also less active,
giving learners limited opportunities to either actively or passively participate in
collaborative learning activities.</p>
      <sec id="sec-3-1">
        <title>Support knowledge base growth across course re-runs. Learner contributions</title>
        <p>should collectively form a knowledge base which becomes available across course
re-runs o ered in a variety of delivery modes (i.e., paced and self-paced). Initial
course runs often have a higher number of enrolled learners and more discussion
forum activity as a result. Each MOOC re-run, begins with a refreshed discussion
forum which results in community knowledge between courses being lost.
Retaining student contributions will facilitate knowledge growth but also poses
information retrieval problems. The interface used to display learner contributions
will need to therefore include intuitive navigation, free text and tag based (i.e.,
folksonomy) search functionality.</p>
      </sec>
      <sec id="sec-3-2">
        <title>Facilitate the delivery of customised scalable feedback. While various Natural</title>
        <p>
          Language Processing (NLP) and Deep Learning algorithms exist, the ability to
accurately grade and provide feedback for free text student submissions within
MOOCs has not been realised. There are however, techniques that can be used
scale feedback provided by instructors, moderators and tutors. These techniques
rely on the similarity between learner submissions and are able to cluster similar
learner responses together. Topic modeling using the Latent Dirichlet Allocation
(LDA) [
          <xref ref-type="bibr" rid="ref1">1</xref>
          ] algorithm is a promising document clustering technique that can be used
to find common topics in student submissions. Instructors, moderators and tutors
can then view a summary of the topics that exist in learner submissions and
provide feedback. Various applications using clustering to provide feedback at
scale have been discussed by [
          <xref ref-type="bibr" rid="ref9">9</xref>
          ]. The topic modeling summary provides an
additional way for learners to gain an overview of other student submissions and
navigate the community constructed knowledge base.
4
        </p>
      </sec>
    </sec>
    <sec id="sec-4">
      <title>Design Prototype</title>
      <p>
        In this section, screen mockups for a prototype that adheres to all of design principles
listed in the previous section are presented. Most tools that support pedagogical
scripting of CSCL employ a visual flowchart metaphor [
        <xref ref-type="bibr" rid="ref6">6</xref>
        ]. The flowchart metaphor
allows the designer to sequence key stages in the activity and specify whether an
individual or group will contribute to the activity. The flowchart metaphor provides a
high level overview of the activity, but the instructor is still required to configure each
stage of the activity. We take a declarative approach to the configuration of the activity,
which both simplifies and reduces the steps required to use the tool. The declarative
approach is encapsulated in a simple user interface that allows the instructor to
configure the activity.
      </p>
      <p>The activity creation interface (see Figure 1), allows instructors to choose a template
and specify the activity configuration settings. The instructor can specify how learners
contribute to the perspectives in an activity (i.e., the learner contributions section).
Options are provided for the instructor to allow learners to choose a perspective,
contribute to all perspectives, or be randomly assigned a perspective. The instructor is
able to enable a curation stage and configure the knowledge base.</p>
      <p>
        Central to the design of the PerspectivesX tool, is a structured template that
instructors are able to create. It is envisaged that the tool will include standard
templates for common activities such as Six Thinking Hats [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ], SCAMPER [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ] and
Fish Bowl [
        <xref ref-type="bibr" rid="ref8">8</xref>
        ]. Instructors will also be able to create custom templates. As an example,
a template can be created for a SWOT activity using a multi-perspective fieldset to
include each text contribution field that is required (i.e., Strengths, Weaknesses,
Opportunities and Threats). The interface to create a template is shown in Figure 2a.
      </p>
      <p>An example learner submission user interface is displayed in Figure 2b. The fields
that a learner is required to complete is dependent upon the settings the instructor has
selected. In Figure 3, the learner has to select a perspective, enter their contribution and
decide whether their contribution will be shared with other students. A knowledge base
is displayed after a learner submits their perspective. The learner is able to see their
contribution as well as view other student contributions that have similar or opposing
views. Content curated by a moderator will be included.
5</p>
    </sec>
    <sec id="sec-5">
      <title>Implementation Considerations: LTI Tool vs XBlock</title>
      <p>The PerspectivesX tool can either be implemented using the Learning Tools
Interoperability (LTI) specification or as an XBlock for the Open edX platform.
LTI tools can be built in any programming language, have their own user interface and
are able to run on their own server. LTI tools are also able to integrate with a range of
Learning Management Systems that implement the LTI specification. XBlocks are
extensions for the Open edX platform, must be built in the Python programming
language and adhere to the Open edX user interface standard. Both the LTI and XBlock
implementation options are comparable in terms of creating a user interface for the
instructor and learner. As LTI’s have the flexibility of being installed on a separate
server, key features for the knowledge base will be easier to implement and scale.
These features include the persistence of knowledge base data across course-runs and
content indexing for search. Implementing PerspectivesX as an LTI would provide
more flexibility to readily integrate with advance NLP and Deep Learning algorithms.
6</p>
    </sec>
    <sec id="sec-6">
      <title>Conclusion and Future Directions</title>
      <p>The PerspectivesX tool will be developed as an open source LTI tool from the
design mockups proposed in this paper. Future research will focus on the evaluation of
the PerspectivesX tool and extending the design principles to support synchronous
collaborative activities.</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          1.
          <string-name>
            <surname>Blei</surname>
            ,
            <given-names>D.M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Ng</surname>
            ,
            <given-names>A.Y.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Jordan</surname>
            ,
            <given-names>M.I.</given-names>
          </string-name>
          :
          <article-title>Latent dirichlet allocation</article-title>
          .
          <source>Journal of machine Learning research 3(Jan)</source>
          ,
          <fpage>993</fpage>
          -
          <lpage>1022</lpage>
          (
          <year>2003</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          2.
          <string-name>
            <surname>Boyd</surname>
            ,
            <given-names>D.</given-names>
          </string-name>
          :
          <article-title>Streams of content, limited attention: The flow of information through social media</article-title>
          .
          <source>Educause Review</source>
          <volume>45</volume>
          (
          <issue>5</issue>
          ),
          <volume>26</volume>
          (
          <year>2010</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          3.
          <string-name>
            <surname>Corrin</surname>
          </string-name>
          , L.,
          <string-name>
            <surname>de Barba</surname>
            ,
            <given-names>P.G.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Bakharia</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          :
          <article-title>Using learning analytics to explore help-seeking learner profiles in moocs</article-title>
          .
          <source>In: Proceedings of the Seventh International Learning Analytics &amp; Knowledge Conference</source>
          . pp.
          <fpage>424</fpage>
          -
          <lpage>428</lpage>
          . ACM (
          <year>2017</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          4.
          <string-name>
            <surname>De Bono</surname>
            ,
            <given-names>E.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Pandolfo</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          :
          <article-title>Six thinking hats</article-title>
          , vol.
          <volume>192</volume>
          . Back Bay Books New York (
          <year>1999</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          5.
          <string-name>
            <surname>Eberle</surname>
            ,
            <given-names>B.</given-names>
          </string-name>
          :
          <article-title>Scamper on: Games for imagination development</article-title>
          . Prufrock Press Inc. (
          <year>1996</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          6.
          <string-name>
            <surname>Faucon</surname>
            ,
            <given-names>L.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Haklev</surname>
            ,
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Hadzilacos</surname>
            ,
            <given-names>T.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Dillenbourg</surname>
            ,
            <given-names>P.</given-names>
          </string-name>
          :
          <article-title>Demo of orchestration graph engine: Enabling rich social pedagogical scenarios in moocs (</article-title>
          <year>2017</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          7.
          <string-name>
            <surname>Hill</surname>
            ,
            <given-names>P.</given-names>
          </string-name>
          :
          <article-title>Mooc discussion forums: barrier to engagement? (</article-title>
          <year>2013</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          8.
          <string-name>
            <surname>Miller</surname>
            ,
            <given-names>R.L.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Benz</surname>
            ,
            <given-names>J.J.</given-names>
          </string-name>
          :
          <article-title>Techniques for encouraging peer collaboration: Online threaded discussion or shbowl interaction</article-title>
          .
          <source>Journal of Instructional Psychology</source>
          <volume>35</volume>
          (
          <issue>1</issue>
          ),
          <fpage>87</fpage>
          -
          <lpage>94</lpage>
          (
          <year>2008</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref9">
        <mixed-citation>
          9.
          <string-name>
            <surname>Mohler</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Mihalcea</surname>
          </string-name>
          , R.:
          <article-title>Text-to-text semantic similarity for automatic short answer grading</article-title>
          .
          <source>In: Proceedings of the 12th Conference of the European Chapter of the Association for Computational Linguistics</source>
          . pp.
          <fpage>567</fpage>
          -
          <lpage>575</lpage>
          . Association for Computational Linguistics (
          <year>2009</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref10">
        <mixed-citation>
          10.
          <string-name>
            <given-names>O</given-names>
            <surname>'Connell</surname>
          </string-name>
          ,
          <string-name>
            <surname>J.</surname>
          </string-name>
          :
          <article-title>Content curation in libraries: is it the new black? Collected Magazine (6</article-title>
          ),
          <fpage>4</fpage>
          -
          <lpage>6</lpage>
          (
          <year>2012</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref11">
        <mixed-citation>
          11.
          <string-name>
            <surname>Slotta</surname>
            ,
            <given-names>J.D.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Naja</surname>
          </string-name>
          , H.:
          <article-title>Supporting Collaborative Knowledge Construction with Web 2</article-title>
          .0 Technologies, pp.
          <fpage>93</fpage>
          -
          <lpage>112</lpage>
          . Springer New York, New York, NY (
          <year>2013</year>
          )
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>