Impact of early activity on flipped classroom performance prediction: A case study for a first-year Engineering course

Abstract

This article addresses the problem of performance prediction in a flipped learning course on digital systems. The course is a part of a first-year Engineering degree offered at a public university. The extent to which failure at the end of the course can be anticipated is analyzed by examining the students' behavior related to academic activities. This prediction will be made by considering whether students perform different activities as instructed. The timeframe is also limited to the early stages of the course. Although anticipating results at early stages on the basis of behavioral aspects makes it possible to redress the behavior of at-risk students, at the same time, it also introduces limitations for potential models. As there are no assessments during the early period, no course activity grades will be available for modeling purposes. In this study, different data mining techniques will be used, which will offer better results. Results will be analyzed in terms of the performance of at-risk students and compared with previous research studies. Models will also provide information about factors of relevance for the academic success of the course.