Feedback Preferences of Students Learning in a Blended Environment: Worked Examples, Tutored and Untutored Problem-Solving

In contemporary technology-enhanced, learning platforms that combine the learning of new concepts with the practicing of newly learned skills, students are offered multiple feedback options. Typically, a problem-solving exercise allows the option to check the correctness of the answer, for calling hints that provide a partial help in the sequence of problem-solving steps, or calling a fully worked-out example. This opens new opportunities for research into student learning tactics and strategies, leaving the traditional context of lab-based research following experimental design principles behind, going into the research of revealed learning choices of students learning in authentic settings. In this empirical study, we apply multi-modal data consisting of logged trace data, self-report surveys and learning performance data, to investigate antecedents and consequences of learning tactics and strategies applied by students learning introductory mathematics and statistics. We do so by distinguishing different learning profiles, determined by the intensity of using the platform and the relative amounts of examples and hints called. These learning profiles are related to prior knowledge and learning dispositions, as antecedents, and course performance, as a consequence. One of our findings is that of ‘help abuse’: students who bypass the option to call for hints as concrete feedback in their problem-solving journey and instead opt for calling generic solutions of the problem: the worked examples. This help abuse is associated with prior knowledge and learning dispositions, but much less with course performance.