Exploring attitudes of learners with respect to different learning strategies and performances using statistical methods

Abstract

In this work the problem of identifying relationships between different learning strategies and learning outcomes is addressed. Classical statistical methods such as p values and chi square test, as well as Multiple Correspondence Analysis are employed; variables to be explained are performances of learners in Multiple Choice Tests (MCT) and Design Tests. It is shown that: the methods are able to detect differences with a different sensitivity; the methods are able to detect characteristics belonging to the metacognitive domain; specific strategies are effective to learn complex skills. Further applications are discussed, especially for what concerns cognitive and metacognitive changes happening in time.