Computer-automated testing: An evaluation of student performance

Abstract

In the last few years, the need for an automated way to assess people has increased quickly because of the growing request from both private and public structures. Many Learning Management Systems (LMS) have been developed in order to automatize the learning and assessment process. In most of the cases these systems don't allow a quality content evaluation and an efficient ability estimation. In this paper we analyze the features of the Proportional Odd Model (POM), belonging to the Item Response Theory. The POM enables to translate an automatic test deliver in an efficient way in order to get an evaluation through either a summative or a formative way. The data have been collected in some undergraduate courses of Bologna University, by using test delivering and by authoring system developed in ASP and Java, respectively. © 2005 by International Federation for Information Processing.