Co-learners Bibliography II
Paper format:
Related work with colearning agents:
Similar in identifying expertise and behavior of colearner agent as important concerns, different in that they want the student to teach the companion. System name= LECOBA, application = boolean algebra. 2x2 experiment, strong v. weak expertise, free v. motivated interaction. Weak companions annoyed users, strong companions were annoyed by the colearner's unwillingness to accept their guidance.
Describes using colearners "for distubance." e.g. agent can propose erroneous solutions to test student self-confidence, create cognitive dissonance to motivate learing.
Study compares effect of having two pedagogical agents (Expert+motivator) vs. one agent (Mentor) in a learning exercise in an educational technology course. Participants with two agents had better recall of the material than those with one agent, and they found the material easier to learn. Not peer agents.
Argues that pedagogical agents (mentors) need 1) regulated intelligence 2) existence of a persona and 3) pedagogical control to be effective for teaching.
Most relevant part describes different approaches to collaborative learning:
Socio-constructivist: people learn more by interacting in pairs due to mediation of socio-cognitive conflict
Socio-cultural approach: social interaction induces individual cognitive change
Shared cognition approach: shared physical and social context affect the social meanings of interactions. Ideas are "jointly constructed"
see also Computer Supported Collaborative Learning Issues for Research Vive Kumar
Studies effect of colearner questions on participants estimation of own understanding. People felt that they were more confused if the colearner asked more questions. The study used simluated colearners, not colearning agents.
Experimental study indicating that students were better able to build a replica of a Lego model alone and with other students. Finds that collaboration more conductive to learning than individual work, and that collaborating with an expert makes you better able to generalize what you ahve learned. Focuses on expertise as key feature of collaborator.
Students working in cooperative triads performed better on post-tests and had better reasoning than did students in competitive triads and students learning alone.