MATH ONLINE TESTS: A RASCH ANALYSIS

The M.In.E.R.Va. project, which was presented at EDULEARN 2014 and 2016, was developed to help\r\nstudents of secondary schools to recover their deficiencies in Mathematics. As the experimentation\r\nhas produced good results, we extended the project to Calculus and Financial Mathematics courses in\r\nCatholic University in Milan. We will use Rasch Analysis and other statistical techniques (e.g. Classical\r\nTest Theory and Item Response Theory) to prove the “goodness” of the multiple-choice questions that\r\nwe used in the online exercises and to perform a differential analysis, dividing students in subgroups,\r\nin order to individuate the “disturbing” variables. We found that a few questions were problematic. In\r\nthis case, the common approach is to change the question. We preferred instead to change the\r\ndidactic approach to the subject of the question and to change the question only if the performance\r\nremains low. Since this trial showed quite good results, we can conclude that the platform can be\r\nsuccessfully used to point out early problematic subjects and to intervene immediately.