Modeling the Acquisition of Skills in Motivational Interviewing With MITutor

• In Event: Learning Analytics Tools to Improve Instruction

Abstract

This paper examines the rate of skill acquisition in the initial stages of learning in the domain of counseling to improve adherence to therapeutic protocols. The focus is on Additive Factors Models (AFM), a data mining method applied to log trace data in MITutor. MITutor is an intelligent tutoring system that relies on natural language processing and data-mined models for the purposes of ongoing diagnosis, calibration, and fading of instruction. The data consists of 691 responses made by 21 students to virtual agents in several instructional conditions and modules. The AFM predictions warrant an overall decrease in over-practiced skills in as much as 36% of practice opportunities, leading to significant improvements with regards to instructional efficiency.

Authors

• Kent Ellsworth

• Eric G. Poitras, Dalhousie University

• Zac Imel, The University of Utah

• Derek Caperton

• Grin Lord

• Jake Van Epps, The University of Utah

• Michael Tanana, The University of Utah

• David Atkins, University of Washington