|Teaching and Learning Forum 2003 [ Proceedings Contents ] |
Learning is a subtle activity, difficult to measure in an absolute way. A true evaluation of "deep" student learning using standard assessment mechanisms, like multiple choice tests, can be nearly impossible - the deep learning "signal" swamped by the "noise" of other test related processes, such as student test taking strategies. The problem of extracting the deep learning "pattern" in a set of learning data (if the pattern exists at all) is similar to the process of "data mining" found in the database industry, where statistical techniques are used to locate hidden patterns in large databases. Even a relatively short multiple choice test can provide a large enough set of data for data mining techniques to be an appropriate analysis method. The author's experience using data mining techniques to analyse the results of a multiple choice test, in order to evaluate a constructivist teaching methodology, will be presented.
|Please cite as: Bongiovanni, J. (2003). Looking for learning: A data mining experience. In Partners in Learning. Proceedings of the 12th Annual Teaching Learning Forum, 11-12 February 2003. Perth: Edith Cowan University. http://lsn.curtin.edu.au/tlf/tlf2003/abstracts/bongiovanni-abs.html|