Use of knowledge mapping in computer-based assessment
- Knowledge mapping vs concept mapping. Concept mapping has been trademarked in the US.
- Learning (each box has its own instructional strategy and assessment - sort what you need to teach into these five boxes)
- content understanding
- collaboration/teamwork
- problem solving
- communication
- self-regulation
- Today talk is about content understanding and problem solving
- O'Neil's Problem Solving Model
- Problem Solving
- Domain specific PBS strategies
- Domain independent PSS
- Content Understanding
- Self regulation
- Metacognition
- Motivation
- Knowledge maps
- node-link representation (node = concepts, links = relationships)
- has typically been used for instructional purposes, we're using it for assessment purposes
- substitute for an essay test (rather than an MC test)
- Can be scored automatically
- examples
- “sunlight → leads to → surface warming”
- example of knowledge map about 1930's depression
- software compares student-generated map with “perfect map”?
- ask more than one expert to represent the area, map student map against all three experts, take average, use that number
- Computer gaming (problem solving task)
- measures
- knowledge mapper
- retention similar to Mayer
- self-regulation
- Research questions
- what are the scoring methods for KMs
- what is reliability and validity
- scoring approaches (matrix)
- referent free vs referent based
- configural vs semantic
- choice: compares semantic structure of student work with expert work (referent based and semantic)
- can train people to rate maps very reliably
- tedious, labor-intensive, time-consuming (but only way to do it with unconstrained tasks)
- Conclusion
- feasible, reliable, sensitive to knowledge differences and instructional effects
- task format influences reliability
- automated scoring tractable with predefined concepts and links
(http://cresst.org, UCLA, honeil@usc.edu)