Toward Computer-Based Support of Meta-Cognitive Skills: a Computational Framework to Coach Self-Explanation
We present a computational framework designed to improve learning from examples
by supporting self-explanation – the process of clarifying and making more complete to oneself
the solution of an example. The framework is innovative in two ways. First, it represents the
first attempt to provide computer support to example studying instead of problem solving.
Second, it explicitly coaches a domain-general, meta-cognitive skill that many studies in
cognitive science have shown to greatly improve learning. The framework includes solutions to
three main problems: (1) to design an interface that effectively monitors and supports selfexplanation;
(2) to devise a student model that allows the assessment of example understanding
from reading and self-explanation actions; (3) to effectively elicit further self-explanation that
improves student's example understanding. In this paper, we describe how these solutions have
been implemented in a computer tutor that coaches self-explanation within Andes, a tutoring
system for Newtonian physics. We also present the results of a formal study to evaluate the
usability and effectiveness of the system. Finally, we discuss some hypotheses to explain the
obtained results, based on the analysis of the data collected during the study.