Technology-enhanced interactions & analysis

Georgios Fesakis, Angelique Dimitracopoulou, Aggelos Palaiodimos

Previous pilot work on blogging Fessakis et al 2008: graphical synopses of interaction raw data of group blogging helps group to increase level of awareness and regulate function leading to better collaboration by students and moderation by teachers

Q: To what extent do IA graphs increase self-regulation by group blog members, impact of different kinds of graphs?

• 147 students, 21 groups, 7 members each.
• 3 hours training
• manual of use on blog

Task: Design technology enhanced lesson plan

Phases

• socialization
• first design and peer review
• revision and peer review
• final project deliverables

Graphs

• first group
  • posts per member (per period)
  • comments per member
  • SNA
• second group
  • all graphs on group base, not user base
  • comments-posts bubble graph per group (compare groups at a glance)

Result

• graphs of first category produce more posts and comments
• intragroup stronger than inter-group
• why?
  • increase competition
  • gives students feeling that teacher is monitoring participation

Further direction

• automatic production of graphs
• smooth integration in learning environment

Social Annotating in the Online Margins- Re(designing) an Annotation Tool drawing on Unintended Ways University Students and Faculty Chose to Use It Annelies Raes, Tammy Schellens, Bram De Wever, Ellen Vanderhoven (Ghent University, Belgium)

When students are searching for evidence on the web, they use ineffective strategies or don't use strategies at all. What skills do they need, how to scaffold?

Conceptual framework of Brand-Gruwel 2005. Information problem solving while using internet.

Scaffolding

• teacher/human tutor
• software tools
• distributed scaffolding
• …
• technology-enhanced scaffolds (embedded hints and question prompts, static, fixed, faded through project)
• teacher-enhanced scaffolds (cues and prompts given by teacher or human tutor, dynamic and interactive)

RQs

• WISE project, but also web search etc
• impact of scaffolding conditions on students learning science
  • domain-specific knowledge
  • regulation of cognition (Schraw & Dennison 1994)
• do different scaffolding conditions interact with students' characteristics, gender and students' prior knowledge

Pre-test/post-test

Condition with multiple scaffolds significantly outperformed other conditions.

Boys perform worse with teacher + technology, but girls better if teacher is involved. Main impact on students with low prior knowledge.

Eva Bures, Andrew Feenberg, Cindy Xin, Philip Abrami (Bishop's University, Canada)

Can highlight text, and write little notes on the side.

Design research: changing features to meet their emergent uses.

Twelve grad/undergrad classes in Canada, ed and philosophy.

Expected to use it as a private tool - highlight something, just like you do when reading on paper.

Feature hacked into Moodle.

The students wanted to see each others annotations. First big change: default public. Nobody makes them private.

Many of the benefits from students were in looking at other people's annotations - seeing what other students, or what instructor thought.

Functions:

• retrieval online (like “bookmarking” passages)
• reflection
• reply
  • unexpected - students would reply to comments, starting small discussions on the side
  • kinds
    • socio-emotive (encouraging, etc)
    • meta-cognitive (help progress through task, coordination)
    • constructing knowledge through content dialogue (questioning directly)
• instructor feedback
  • also unexpected, really love it

A review tool became a social tool.

Controversial changes:

• reply to annotation button was added
  • didn't use that, kept adding normal annotations
• length of annotations set to 250 words (increase) - very aggravating, need more

Johnny Yuen

Reflective judgment model (King & Kitchener 1994, 2004)

• students reasoning styles in response to ill-structured problems
  • 7 levels
  • pre-reflective → quasi-reflective → reflective

Epistemological authenticity framework (Chinn & Brewer, 2001; Chinn & Malhotra 2002)

Use Knowledge Forum

Data

• survey
• coded final essays
• contributions to KF (argumentative or questioning markers)

Quality of students engagements discourse-markers in socio-constructive online discourse are somewhat related to students personal reflective judgment.

Susan Bull, Ravi Vatrapu (U Birmingham, Copenhagen Business School)

• HCI
  • intersubjective epistemology
    • socio-technological affordances (Suthers 2006, Vatrapu 2010)
    • technological intersubjectivity (Vatrapu 2009)
    • group cognition (Stahl 2006)
  • HCI & CSCL
    • awareness, articulation and appropriation (CSCW)
    • learner-centered design (Quintana, Shin, Norris & Soloway 2006)
      • user-centered design
      • usability, sociability, learnability (Vatrapu 2008)
• Open Learner Models (OLM)
  • domain model, learner model as subset of domain model?
  • constructed based on activities of learner
  • in intelligent tutoring systems, learner model is not accessible to user
  • open: available to learner
    • representation (learner knowledge, beliefs, understanding, skills)
    • data (problem-solving tasks, specific questions, open-ended interaction)
    • techniques (constraint-based, Bayesian networks, fuzzy, simple weighted algorithms)
  • adaption to interaction and personalization of learner
  • OLM is a human-understandable externalisation of underlying learner model
  • representations to
    • learners
    • teachers
    • parents
  • kinds of representations
    • graphs
    • text
    • concept maps
    • etc
  • aims
    • promoting metacognitive skills
    • navigation
    • etc
• OLM & CSCL
  • three approaches
    • individual learner models, available to peers
    • group model comprising data from individual team members
    • combined group model which is available to group members
  • optional sharing of OLM
    • seek suitable collaborators
    • or seek more capable others
  • even when not shared
    • they talk, spontaneously discuss their learner models
  • potential benefits
    • individual perception
      • socio-technical affordances
        • meaning making opportunities and action-taking psosibilities
      • computer support
        • artifact-centered discourse - models become artifacts for discourse
      • collaborative learning
        • empowers student to be in control of own zone of proximal development
        • open possibility for inter-group collaborative learning
• open research questions
  • roles
  • differences
  • access to data

Interesting tension between what is open and what is private, public, ad-hoc sharing etc (annotation and OLM).

• also consider longevity - data that is there forever
• bad annotators might be the most vocal ones
• you might be unwilling to write publicly critical annotations to peers

Critical question: lot's of process variables, but did they learn? However you define learning. (Very skeptical to learner models based on domain models).

Penstein Rose: How do process data link to summative measures?