SLIDE 1
Teacher-sourcing semantic information in a Physics blended-learning environment
Elad Yacobson Asaf Bar Yosef Eliran Hen Giora Alexandron
Teacher-sourcing semantic information in a Physics blended-learning - - PowerPoint PPT Presentation
Teacher-sourcing semantic information in a Physics blended-learning - - PowerPoint PPT Presentation
Teacher-sourcing semantic information in a Physics blended-learning environment Elad Yacobson Asaf Bar Yosef Eliran Hen Giora Alexandron Overall rationale Generating Semantic metadata
SLIDE 2
SLIDE 3
לחץלקבלתהצעותלפעילויות המשך
SLIDE 4
Overall rationale
SLIDE 5
Generating Semantic metadata
- Conservative approach: Expert tagging
- Limitations:
- Doesn’t Scale
- Content and context are dynamic
SLIDE 6
Our approach – Crowdsourcing
- Crowdsourcing from teachers and learners (teacher/learner-sourcing)
- Can (potentially) address all the limitations just mentioned
- Challenges:
- Accuracy
- Motivation
- Coverage [how do we cover all the aspects that we need]
- Design research: How to optimize accuracy, incentives, and coverage
SLIDE 7
Example of one unit – Magnetism
Developing the Level 4 taxonomy + tagging the content took a few months !!!
SLIDE 8
Pilot Experiment: Teacher-sourcing
A question on Magnetism
4 Possible tags
Tag Yes/No
SLIDE 9
RESULTS - Experiment #1
8 teachers, 8 items
SLIDE 10
RESULTS - Experiment #2
7 teachers, 2 items
SLIDE 11
Pilot study – Lessons Learned
- High agreement between experts and teachers (Cohen Kappa 0.56)
- Teachers were able to tag questions without full knowledge of the
taxonomy from which the tags were taken
- Teachers needed to solve the questions in order to tag them
- Teachers’ motivation affected accuracy of tagging
- Contribution to teachers’ Professional Development?
- Next step: a pop-up survey to collect metadata @scale
SLIDE 12
Research Team
Elad Yacobson Asaf Bar Yosef Eliran Hen Giora Alexandron