Location:
|
KFC Courts |
Date:
|
Tuesday, Aug.02 |
Time:
|
6:00PM - 6:45PM
|
Author:
|
Carolin N. Cardamone, MIT
617-324-2731, cnc@mit.edu
|
Co-Author(s):
|
Saif Rayyan, Daniel Seaton, Albert Wu, Dave Pritchard
|
Abstract:
|
Item Response Theory (IRT) algorithms are being developed to better assess student performance in our Integrated Learning Environment for Mechanics (ILEM; [1]). A student's skill, as determined by IRT, provides more information than the traditional student score because it takes into account universally calibrated problem difficulties. Importantly, it allows determination of skill on a universal scale independent of which questions the student answers. Our approaches seek to dynamically update student and class skill level in ILEM throughout the course based on their performance, rather than relying primarily on the gain from pre/post testing. We present results comparing IRT and pre/post gain analysis of the Mechanics Baseline Inventory Test, including discussion of item parameters for the 26 questions on the MBT exam.
|
Footnotes:
|
1R. Teodorescu, A. Pawl, S. Rayyan, A. Barrantes and D. E. Pritchard, Toward an Integrated Online Environment, 2010 Physics Education Research Conference Proceedings, edited by S. Rebello, M. Sabella and C. Singh
|
|
|