Work

Knowledge tracing and cue contrast: Second language English grammar instruction

Year: 2013

Type: article

Abstract: This paper introduces a cognitive tutor designed for second language grammar instruction.The tutor adopted Corbett and Anderson's (1995) Bayesian knowledge tracing model and provided adaptive training on the English article system.We followed the Competition Model (MacWhinney, 1997) and understood the article system as a galaxy of cues determining article usage on the basis of form-function mapping.Cues are in competition during language acquisition; hence cue contrast is predicted to be an effective instructional method.Seventy-eight students were randomly assigned to four article training conditions (to learn 33 cues) and a control condition (to write essays).We found that article-training groups significantly outperformed the control group in an immediate posttest and a delayed posttest.Specifically, our result also suggested that there was a significant interaction between cue contrast and cue type (definite vs. indefinite).Cue contrast promoted more learning on the indefinite cues (more difficult for learners).Knowledge tracing did not demonstrate such an interactional effect with cue types.Instead, it boosted the instructional effect promoted by cue contrast.

Source: eScholarship (California Digital Library)

Authors: Helen Zhao, Kenneth R. Koedinger, John Kowalski

Institutions: Chinese University of Hong Kong, Carnegie Mellon University

Language: English

Topic: Intelligent Tutoring Systems and Adaptive Learning

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDG: Quality education

Open Access status: green

Funder: National Science Foundation

Award: SLC Center: Pittsburgh Science of Learning Center: Studying Robust Learning with Learning Experiments in Real Classrooms