Augmenting Electronic Environments for Leadership
Abstract
Adapting instructional content to match the background knowledge of the student has been a long-standing goal of student modeling and tutoring (Bruner, 1966; Burton & Brown, 1979). To this end, cognitive scientists have developed student models that rely on manual knowledge bases relevant to the particular instructional task at hand. Normally, these include domain content, instructional content, models of student misconceptions, and more. While tailoring instruction to the learner has been shown to be effective (Anderson, 2002), current approaches are difficult to implement because of the enormous amount of skilled professional effort required. Ideally, the system should automatically select the most appropriate content for the student based on a minimal amount of student data. This educational desideratum has been coined the "Goldilocks Principle" --- providing lessons, texts, probes, etc. that are neither too advanced nor too elementary, but just right---in the "zone of proximal development (ZPD)" (Palincsar & Brown, 1984; Vygotsky, 1978), known to be important for promoting efficient learning. In an early study, Wolfe et al. (1998) used Latent Semantic Analysis (LSA) to automatically select the next text passage for students to read and achieved one-sigma learning augmentation effects (Bloom, 1976). In the study reported here, we examined the Goldilocks Principle in the context of an LSA - enhanced online discussion environment, where contributions were automatically selected to be most similar in meaning to learners' notes or contributions. We found that these contributions were almost always of somewhat higher judged quality. We suggest that this implementation of the Goldilocks Principle is a consequence of how LSA represents consensual knowledge, and provides an automatic way for selecting the next best piece of material for a student to learn, making this an important contribution to tutoring.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 2004
- Accession Number
- ADA428351
Entities
People
- Joseph Psotka
- Karen Lochbaum
- Ken Robinson
- Lynn Streeter
- Thomas Landauer
Organizations
- U.S. Army Research Institute for the Behavioral and Social Sciences