The discriminative learning gain: a two-parameter quantification of the difference in learning success between courses
© 2018, © 2018 Engineers Australia. Pre- and post-tests are often used in Engineering Education Research to assess teaching and learning. Sometimes, it is reasonable or even necessary to use a different pre- than post-test. In that case, it is difficult to analyse the data with traditional methods such as average normalised gain or normalised change scores. We propose to use the so-called discriminative learning gain (DLG) to analyse such data. This two-parameter statistic describes the post-test performance of model students, taking into account varying pre-test performances. Thus, it describes not only the learning of the average student, but also the discriminative effect of the instruction with respect to initial performance. Teachers and researchers can use the DLG as a tool to effectively compare course performances. Using confidence bounds, the difference in learning success among courses can be quantified and easily visualised. Consequently, informed conclusions can be drawn if courses differ in effectiveness. This article demonstrates how to apply and interpret the DLG. Limitations of the method and the application to the special case of identical pre- and post-tests are discussed.
More Funding Information
This work was supported by the German Federal Ministry of Education and Research (BMBF) [grant number 01PL11047]. Any opinions expressed here are those of the authors.