dc.creator | Segedy, James René | |
dc.date.accessioned | 2020-08-22T17:03:32Z | |
dc.date.available | 2014-06-09 | |
dc.date.issued | 2014-06-09 | |
dc.identifier.uri | https://etd.library.vanderbilt.edu/etd-06062014-120717 | |
dc.identifier.uri | http://hdl.handle.net/1803/12486 | |
dc.description.abstract | Open-ended computer-based learning environments (OELEs) challenge learners to independently solve complex problems. These environments provide powerful opportunities for learners to develop and utilize strategies for self-regulated learning and problem-solving. However, novice learners often struggle in such open-ended environments, therefore, the extent of their effectiveness depends on the capabilities of their computer-based scaffolding agents: software agents embedded within the system that provide adaptive support to struggling learners. To be effective, these agents require systematic methods for effectively modeling and scaffolding (i.e., supporting) learners so that they can provide help that is targeted to addressing weaknesses in their problem-solving capabilities.
The research presented in this dissertation has focused on expanding the repertoire of scaffolding agents in OELEs along two fronts. First, an approach to modeling learners called coherence graph analysis (CGA) has been developed. The CGA approach models learners in terms of: (i) the quality of their problem solutions; (ii) their skillfulness in solving open-ended problems; and (iii) the coherence between the actions they perform as part of their problem-solving tasks. Second, a three-stage approach to scaffolding students has been developed and evaluated through classroom studies. This scaffolding strategy actively helps students by: (i) offering to answer their questions; (ii) diagnosing their skill deficiencies; and (iii) requiring them to develop problem-solving skills through guided practice. These approaches were evaluated in a study with two instructional units in four 6th grade classrooms. The results demonstrated the utility of the CGA approach in predicting learners’ performance and learning. Exploratory clustering analyses were employed to explore student behavior. The analyses identified a set of distinct and persistent behavioral profiles among the students. Despite significant changes in students’ behaviors, the set of behavioral profiles identified by the clustering analyses were similar for both instructional units. The analyses also revealed a productive strategy shift: of the 98 students who took part in the study, 60 of them exhibited improved problem-solving behaviors during the second instructional unit. Analyses also provided suggestive evidence for the value of the three-stage scaffolding strategy in helping students learn how to succeed at complex open-ended problem-solving tasks. | |
dc.format.mimetype | application/pdf | |
dc.subject | Scaffolding | |
dc.subject | Open-Ended Learning Environment | |
dc.subject | Learner Modeling | |
dc.subject | Adaptive Support | |
dc.subject | Learning Analytics | |
dc.title | Adaptive Scaffolds in Open-Ended Computer-Based Learning Environments | |
dc.type | dissertation | |
dc.contributor.committeeMember | Dr. Julie Adams | |
dc.contributor.committeeMember | Dr. Robert Bodenheimer | |
dc.contributor.committeeMember | Dr. Doug Fisher | |
dc.contributor.committeeMember | Dr. Doug Clark | |
dc.type.material | text | |
thesis.degree.name | PHD | |
thesis.degree.level | dissertation | |
thesis.degree.discipline | Computer Science | |
thesis.degree.grantor | Vanderbilt University | |
local.embargo.terms | 2014-06-09 | |
local.embargo.lift | 2014-06-09 | |
dc.contributor.committeeChair | Dr. Gautam Biswas | |