Use of IMMEX technology for the automated, rapid assessment of metacognition use in chemistry problem solving

CHED 367

Santiago Sandi-Urena, gsandiu@clemson.edu1, Melanie M. Cooper, cmelani@clemson.edu1, and Ron Stevens2. (1) Department of Chemistry, Clemson University, 201 Hunter Laboratories, Clemson, SC 29634, (2) The IMMEX Project, 5601 W. Slauson Ave., Suite 255, Culver City, CA 90230
The assessment of metacognition is intrinsically difficult because it is not an overt behavior. Traditional concurrent measurements like systematic observation or think aloud protocols are time consuming, prone to bias and subjectivity, inadequate for rapid identification, and the presence of an observer may result distracting for the participant. The IMMEX software has been recently introduced as a novel, automated way of accomplishing non-disruptive, rapid metacognition activity assessment. This web-based problem solving space uses an HTML tracking feature to create records of students' actions. Subsequent protocol-analysis through artificial neural networks (ANN) and Hidden Markov Models (HMM) modeling of a large number of performances enables the creation of a reduced number of metacognition descriptors to rapidly classify students. This talk describes the use of IMMEX to rapidly characterize students' use of metacognitive skills in chemistry problem solving thereby circumventing the drawbacks of self-reports and traditional concurrent instruments.